1 INSIGHTS INTO THE MECHANISMS OF SINGLE WALL CARBON NANOTUBE (SWCNT ) SPECIATION AND IMPLICATIONS FOR ORGANISM SW C NT INTERACTIONS By JUSTIN G EORGE CLAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLO RIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2 2014 Justin G eorge Clar
3 To my family
4 ACKNOWLEDGMENTS I would like to thank my acad emic advisor and committee chairman Dr. J ean Claude J. Bonzongo for his support and guidance over the last several years. Special appreciation goes to my co chair, Dr. Kirk J. Ziegler for his endless support, treating me like one of his own graduate studen ts. I would also like to thank my other committee members Drs. Trevor Boyer and Andy Zimmerman for their input and assistance during my time at UF. I also would like to acknowledge the support from the National Science Foundation, and Dr. Doug J. Levey for being a fantastic mentor. I would also like to thank all my colleagues and undergraduate students I had the opportunity to work with I know I learned more from you all than you ever learned from me. Most importantly, I would like to thank my father Kevin, mother Margaret, and brother Brendon. Their never ending support made me believe anything was possible and it will never be forgotten.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION: RESEARCH OVERVIEW AND DISSERTATION ORGANIZATION ................................ ................................ ................................ .... 14 1.1 Problem Statement ................................ ................................ ........................... 14 1.2 Organization of this Dissertation ................................ ................................ ....... 16 2 SIN GLE WALL CARBON NANOTUBES: PROCCESSING AND TOXICITY .......... 18 2.1 Characteristics and Structure ................................ ................................ ............ 18 2.2 SWCNT Synthesis ................................ ................................ ............................ 19 2.3 SWCNT Stabilization ................................ ................................ ........................ 20 2.4 SWCNT Characterization ................................ ................................ .................. 21 2.5 Post Suspension Processing ................................ ................................ ............ 24 2.6 SWCNT Toxicity ................................ ................................ ................................ 26 2.6.1 Cytotoxicity ................................ ................................ .............................. 27 2.6.2 Aquatic Organisms ................................ ................................ .................. 28 2.6.3 Human Cell Lines ................................ ................................ .................... 30 2.7 Remaining Questions and Research Goals ................................ ...................... 31 3 INTERACTIVE FORCES BETWEEN SODIUM DODECLY SULFATE SUSPENDED SINGLE WALL CARBON NANOTUBES AND AGAROSE GELS ... 37 3.1 Background ................................ ................................ ................................ ....... 37 3.2 Methods ................................ ................................ ................................ ............ 39 3.2.1 Materials ................................ ................................ ................................ .. 39 3.2.2 Aqueous SWCNT Suspensions ................................ ............................... 40 3.2.3 Equilibrium Adsorption ................................ ................................ ............. 40 3.2.4 Nonequilibrium Adsorption ................................ ................................ ...... 41 3.2.5 SWCNT Characterization ................................ ................................ ........ 41 3.3. Result and Discussion ................................ ................................ ..................... 42 3.3.1 Physical and Chemical Structure of Agarose Gels ................................ .. 42 3.3.2 Retention of SDS SWCNTs on Agarose Gels ................................ ........ 44 3.3.3 Probing the Interaction of SWCNTs with Agarose ................................ ... 45 188.8.131.52 Adsorption predominantly through van der Waals interactions ...... 47
6 184.108.40.206 Adsorption predominantly through ionic interactions ...................... 48 220.127.116.11 Adsorption predominantly throug h hydrophobic interactions .......... 49 18.104.22.168 Adsorption predominantly through interactions ........................ 50 3.3.4 Nature of Adsorption between SWCNTs and Agarose ............................ 51 22.214.171.124 Role of ion dipole interactions in selectiv e adsorption ................... 52 126.96.36.199 Role of SDS in separation selectivity ................................ ............. 54 188.8.131.52 Adsorption isotherm behavior ................................ ........................ 57 3.4 Closing Remarks ................................ ................................ ............................... 59 4 EVALUATION OF CRITICAL PARAMETERS IN THE SEPARATION OF SINGLE WALL CARBON NANOTUBES THROUGH SELECTIVE ADSORPTION ONTO HYDROGELS ................................ ................................ ..... 71 4.1 Background ................................ ................................ ................................ ....... 71 4.2 Methods ................................ ................................ ................................ ............ 74 4.2.1 Materials ................................ ................................ ................................ .. 74 4.2.2 Aqueous Suspension Preparation and Concentration ............................. 74 4.2.3 Equilibrium Adsorption Isotherms ................................ ............................ 75 4.2.4 Column Separations ................................ ................................ ................ 76 4.2.5 SWCNT Characterization ................................ ................................ ........ 76 4.3 Results and Discussion ................................ ................................ ..................... 77 4.3.1 Hydrogel Characteristics used in SWCNT Separations ........................... 78 4.3.2 Typical Elution Profile of SWCNTs during Selective Adsorption onto Hydrog els ................................ ................................ ................................ ...... 79 4.3.3 SWCNT Separations for Various Dextran Gels ................................ ....... 79 4.3.4 SWCNT Separations for Various Agarose Gels ................................ ...... 83 4.3.5 Lifetime of Hydrogels from Repetitive Separations ................................ .. 86 4.3.6 Effect of Altering the SWCNT Loading ................................ .................... 88 4.3.7 Selecting a Hydrogel for SWCNT Separations ................................ ........ 89 4.4 Closing Remarks ................................ ................................ ............................... 90 5 EFFECTS OF THE ELECTRONIC CHA RACTER OF SINGLE WALL CARBON NANOTUBES ON THE BIOLOGICAL RESPONSE OF A FRESHWATER GREEN ALGAE ................................ ................................ ................................ .... 104 5.1 Background ................................ ................................ ................................ ..... 104 5.2 Methods ................................ ................................ ................................ .......... 106 5.2.1 Materials ................................ ................................ ................................ 106 5.2.2 Aqueous Suspension Preparation ................................ ......................... 107 5.2.3 Surfactant Exchange on SWCNT Surfaces ................................ ........... 107 5.2.4 SWCNT Separation into m SWCNTs and s SWCNTs .......................... 108 5.2.5 SWCNT Characteriza tion ................................ ................................ ...... 109 5.2.6 Algal Growth in SWCNT Containing Culture Media ............................... 110 5.3 Results and Discussion ................................ ................................ ................... 111 5.3.1 Initial Screening of Surfactant Toxicity ................................ ................... 111 5.3.2 Biological Response of Test Organism Exposed to as Prepared SWCNTs ................................ ................................ ................................ ..... 112
7 5.3.3 Separation of SDS SWCNTs to Produce m SWCNT and s SWCNT Fractions ................................ ................................ ................................ ..... 112 5.3.4 Validation of Surfactant Exchange Method ................................ ............ 114 5.3.5 Characterization of Separated m SWCNT and s SWCNT Fractions ..... 115 5.3.6 Biological Responses of P. subcapitata to Type Separated SWCNT Fractions ................................ ................................ ................................ ..... 116 5.3.7 Mitigation of s SWCNT Toxicity ................................ ............................. 119 5.3.8 Hydrogel Based Separation Technique as Model for Predicting the Affinity of NPs for Spec ific Bio molecules ................................ .................... 120 5.4 Closing Remarks ................................ ................................ ............................. 121 6 CONCLUSIONS ................................ ................................ ................................ ... 130 6 .1 Summary of Findings ................................ ................................ ...................... 130 6.2 Engineered Nanoparticle Toxicity Management ................................ ............. 133 6.3 Future Directions ................................ ................................ ............................. 134 APPENDIX A SUPPORTING INFORMATION FOR CHAPTER 3 ................................ .............. 138 B SUPPORTING INFORMATION FOR CHAPTER 4 ................................ .............. 145 C SUPPO RTING INFORMATION FOR CHAPTER 5 ................................ .............. 150 LIST OF REFERENCES ................................ ................................ ............................. 152 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 163
8 LIST OF TA BLES Table page 3 1 Characteristics of gel media used in this work ................................ .................... 61 4 1 Parameters used for all elution studies. ................................ .............................. 92 4 2 Physical parameters of hydrogels used in this study. ................................ ......... 93 4 3 Comparison of hydrogel attributes in the separation of 1 wt% SDS SWCN Ts thro ugh selective adsorption ................................ ................................ ............. 103
9 LIST OF FIGURES Figure page 2 1 2D graphite sheet used to define SW CNT vector and classification ................... 34 2 2 Typical Raman Spectra of unpurifie d SWCNTs ................................ .................. 35 2 3 Typical fluorescence spectra of a 1 wt % SDS SWCNT suspension following high shear mixing, sonica t ion, and ultracentrifugation ......................... 36 3 1 Physical an d chemical structure of agarose ................................ ...................... 62 3 2 Separation of SWCN Ts through selective adsorpti on ................................ ........ 63 3 3 Comparison of retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 4 FF and 6 FF at different electrolyte concentrations. .............. 64 3 4 Retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 6 FF and Sepharose 6 FF functionalized with ionic groups, sp and Q that contain negative and positive charges, respectively ................................ ........... 65 3 5 Retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 4 FF and Sepharose 4 FF functionali zed with octyl and butyl groups .................... 66 3 6 Retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 6 FF and Sepharose 6 FF functionalized with phenyl grou ps at low and high substitution ................................ ................................ ................................ ......... 67 3 7 Relationship between ligand density and the mass of SD S SWCNTs retained by different gel media in nonequilibrium (column) studies. ................... 68 3 8 Mechanism of interaction and selectivity during agarose gel based separations of SDS SWCNTs ................................ ................................ ............ 69 3 9 Retention of surfactants used in agarose ge l separations on Sepharose 6 FF .. 70 4 1 Physical and chemical structure of the hydrogels used in this s tudy .................. 94 4 2 Separation of 1 wt % SDS SWCNT suspension with a stati onary phase of Sephacryl 200 HR ................................ ................................ ............................. 95 4 3 Equilibrium adsorption iso therms for 1 wt % SDS SWCNT suspen sion on dextran based hydrogels ................................ ................................ .................... 96 4 4 Retention behavior of 1 wt % SDS SWCNT suspension using dextran based hydrogels. ................................ ................................ ................................ ........... 97 4 5 Elution curves of 1 wt % SDS SWCNT suspe nsions with different hydrogels .... 98
10 4 6 Retention behavior of 1 wt % SDS SWCNT suspensi on using 4 % agarose based gels ................................ ................................ ................................ ......... 99 4 7 effluent collected in both P1 and P2 fractions for columns with a sta tionary phase of Sepharose 6 FF ................................ ................................ ................. 100 4 8 Changes to retention behavior of 1 wt % SDS SWCNT suspension during three consecutive sepa rations for different hydrogels ................................ ...... 101 4 9 The effect of SWCNT loading concentration on the retention behavior of 1 wt % SDS SWCNT suspension using Sephacryl 200 HR as the stationary phase ................................ ................................ ................................ ................ 102 5 1 Effect of increasing concentrations of surfactants used in the aqueous dispersions of SWCNTs on the growth of P. subcapitata in a standar d 96 hour chronic algal assay ................................ ................................ ................... 123 5 2 Effect of increasing concentrations of non separated SWCNTs (i.e. mixture) on the growth of P. subcapitata in a standard 96 hour chronic algal assay as measured by chlorophyll a ................................ ................................ ............... 124 5 3 Spectra of separated SDS suspended SWCNT s using hydrogel packed columns ................................ ................................ ................................ ............ 124 5 4 Evaluation of the efficiency of Amicon Ultrafiltration method in surfactant exchange f rom SDS to SC on SWCNT surfaces ................................ .............. 125 5 5 Effect of su spension preparation on the growth of P. subcapitata in a standa rd 96 hour chronic algal assay. ................................ .............................. 126 5 6 Characterization of separated SWCNT fractions used in dose response studies ................................ ................................ ................................ .............. 127 5 7 Effect of increasing concentrations of separated SWCNTs (m SWCNTs and s SWCNTs) on the growth of P. subcapitata in a stand ard 96 hour chronic algal assay ................................ ................................ ................................ ........ 128 5 8 Effect of increasing concentrations of SC at a fixed concentration of s SWCNTs on the growth of P. subcapitata in a standar d 96 hour chronic algal assay ................................ ................................ ................................ ................ 129 6 1 Experime ntal approach used in this study ................................ ........................ 136 6 2 Flow chart illustrating the research paradigm for future studies evaluating the toxicity of nanoparticles and assessing risks of engineerin g and using nanopar ticles ................................ ................................ ................................ .... 137
11 A 1 Characterization o f suspensions used in Chapter 3 ................................ ......... 139 A 2 Leng ths of SWCNTs used in Chapter 3 ................................ ............................ 140 A 3 Equilibrium adsorption isotherms for SWCNTs in 1 wt % SDS with various agarose gels ................................ ................................ ................................ ..... 141 A 4 Separation us ing functionalized Agarose Gels ................................ ................ 142 A 5 Sepharose separ ation using functionalized gels ................................ ............... 143 A 6 Functionalized Phenyl Series ................................ ................................ ........... 144 B 1 Characterization of SDS SWCNT suspens ions used in this study ................... 145 B 2 ex = 662 nm) of SWCNT fractions collected using S ephacryl 400 HR as an adso rbent ................................ ......... 145 B 3 Retention behavior of 1 wt % SDS SWCNT suspension using 6 % agarose based hydrogels at di fferent levels of cross linking ................................ ......... 147 B 4 ex = 662 nm) of SWCNT fractions collected usi ng Sepharose 6B as an adsorbent ................................ ............... 148 B 5 Fluorescence spectra of SWCNT fractions collected in P1 using Sephacryl 200 HR as an adsorbent with increasing initial SDS SWCNT concentration .... 149 C 1 Characterization of initial suspensions used in separation and of SDS SWCNTs for use in toxicological analysis ................................ ........................ 150 C 2 Aliquot volumes of SWCNT suspensions used in this study ............................ 150 C 3 Effect of separated SWCNTs (m SWCNTs and s SWCNTs) on the growth of P. subcapitata in a standard 96 hour chronic algal assay using SWCN T purchased from NanoIntegris ................................ ................................ ........... 151
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Ful fillment of the Requirements for the Degree of Doctor of Philosophy INSIGHTS INTO THE MECHANISMS OF SINGLE WALL CARBON NANOTUBE (SWCNT ) SPECIATION AND IMPLICATIONS FOR ORGANISM SW C NT INTERACTIONS By Justin G. Clar May 2014 Chair: Jean Claude J. Bonzong o Co Chair: Kirk Jeremy Ziegler Major: Environmental Engineering Science s Nanotechnology is one of the fastest growing sectors of the economy, both in the United States and globally. The develop ment of novel electronic devices and consumer products that take advantage of the unique properties of engineered nanomaterials will inevitably lead to their release in the environment from intentional disposal, or non point single wall carbon nanotube (SWCNT). Published data on the potential for SWCNTs to have negative biological effects on aquatic organisms has been conflicting, likely due to the heterogeneous nature of SWCNT powders. The goal of this dissertation was to elucidate potential differences in biological response of test organisms based on the electronic character of a variety of SWCNT types. A mechanistic study was completed to better understand the forces that drive the separation of SWCNT species by electron ic type through selective adsorption on hydrogels. It was determined that ion dipole interactions between surfactant stabilized SWCNTs and measured dipoles on the hydrogel surface are responsible for nanotube retention in these systems. Insights from the mechanistic study were used to
13 understand how slight changes to the separation protocol, specifically the hydrogel used in separation and initial concentration of SWCNT suspensions can be altered to optimize the selectivity, reproducibility, and throughput of the separation. Optimized process settings were then used to obtain SW CNT fractions of sufficient purity for use in toxicological analysis. Using a simple bench scale bioassays, it was determined that the semiconducting SWCNT species was much more toxic than their metallic SWCNT counterparts. Therefore, the toxicity observ ed in studies using as produced SWCNT powders is driven by the presence of the semiconducting species in the sample. Furthermore, the results highlight the importance of understanding the potential toxicity induced by dispersants commonly used in SWCNT pr ocess as a driver for toxicity response, as opposed to that of the SWCNTs themselves. Finally, these results are a foundation for future work toward the development of simple bench scale test with carbohydrate containing hydrogels to be used in as an init ial screening approach and predictor of SWCNT toxicity when released into aquatic systems.
14 CHAPTER 1 INTRODUCTION : RESEARCH OVERVIEW AND DISSERTATION ORGANIZATION 1.1 Problem Statement Nanomaterials (NMs) are operationally defined as any material w ith at least one dimension of 100 nm or less. These NMs vary greatly in size, shape, composition, and include both organic and inorganic particles. NMs can be further divided into three subcategories (i) natural; (ii) inci dental; and (iii) engineered. 1 Naturally occurring NMs are produced in the environment through traditional earth processes and may be millions of years old, e.g. nano diamonds. 2 Natural NMs have been found in sediments along the Cretace ous Tertiary boundary in Italy 3 and ice cores in Greenland. 4 Incidental NMs are those produced as byproducts of industrial processes or applications not designed to manufacture NMs. Finally, engineered NMs, referred to as nanoparticles (NPs) herein, are produced through a defined synthesis or production process for a specific use or purpose. Over the past two decades, the research and development of NPs has become a major growth industry. The promises and possibilities of NP research include but are not limited to, lighter/stronger building materials, increased data storage capabi lities, biomedical treatment applications, environmental and agricultural application, and manufacturing of much smaller and more powerful electronic devices. As researchers continue to expand their knowledge base, as well as develop and refine current pro duction methods, new uses for NPs are being on Emerging Nanot 00 c onsumer products available in 30 different countries that utilize nanotechnology 5 During 2009, the revenue generated from t he sale of these nanotechnology based products was $1,545 million dollars, and
15 is expected to rise to $5, 335 million dollars by 2015. 6 The U.S. National Science Foundation (NSF) estimates that about 2 million workers w ill be needed to support the growth of the nanotechnology industry. 7 As NP production and inclusion in consumer products continue to rise, the ris k for these new materials to enter and cycle through the environment also rises. 8 Potential pathways include intentional release through wastewater disposal or land filling, or nonpoint release throu ghout the products life cycle, leading to a new class of manmade nanosized environmental toxins. 9 While the greatest portion of money spent on nanotechnology research has been used for the development of new manuf acturing techniques and applications of NPs, money is now being spent on research focusing on the environmental fate, transport, and toxicological implications of NPs. One major obstacle is the lack of analytical technology to accurately identify and quant ify NPs in complex matrices characteristic of most environmental and biological samples. An improved understanding of NP fate and transport will likely lead to regulations regarding their release into the environment. Although the effects of these potenti ally toxic NPs on both the environment and human health remain largely unknown, they are yet to be regulated at the federal level in the U.S. and in Europe. 10 A great deal of research into NP fate and transport still needs to be done in order for regulatory agencies such as the U.S. Environmental Protection Agency (EPA) to establish guidelines for their use and release into the environment. This task becomes exceedingly difficult due to the wide variety of NPs currently produced, and their potential to cycle through the environment in different ways depending on synthesis and stabilization techniques.
16 It is now well established that all NPs have t he potential to enter the environment. However, in this dissertation, a special emphasis is placed on single wall carbon nanotubes (SWCNTs) due to their anticipated widespread production and use in industrial settings. Additionally, the use of SWCNTs in in dustrial applications requires the use of surfactants to obtain ideal dispersions and increase the stability of produced suspensions. The ultimate goal of this study is to develop an understanding of the fate, transport, and environmental impacts of SWCNT s in natural aquatic systems based on a comprehensive investigation of the different interaction forces between SWCNTs and their surrounding environments. 1.2 Organization of this Dissertation In addition to this introductory chapter (Chapter 1), this dis sertation includes 5 chapters which are organized as follows. The current state of knowledge based on literature review of SWCNT synthesis, stabilization, characterization, processing, and potential toxicity is summarized in Chapter 2. This chapter high lights knowledge gaps in the current understanding of how SWCNTs impact the environment, with particular emphasis on the role of SWCNT purification and separation processes on toxicological evaluations. Chapter 3 details the results of a mechanistic study on the separation of Sodium Dodecyl Sulfate (SDS) suspended SWCNTs into distinct fractions by their electronic character using agarose gels. The results presented in this chapter have been published in the Journal of the American Chemical Society 11 Chapter 4 is an extension of the work presented in Chapter 3. Here, laboratory studies were conducted to evaluate how the purity, reproducibility, and throughput of the
17 separation process is affected by a ltering common separation parameters including packing material used in separation, initial sample concentration, and column lifetime. Chapter 5 explores the biological implications of these newly developed separation techniques as well as how initial pro cessing parameters of SWCNT suspension preparation including surfactant stabilization and exchange affect biological responses. Finally, general conclusions summarizing the key finding of this study, as well as future research avenues are presented in Chap ter 6.
18 CHAPTER 2 SI NGLE WALL CARBON NAN OTUBES: PROCCESSING AND TOXICITY 2.1 Characteristics and Structure In general, SWCNTs are part of the fulleren e family, a third allotropic for m of carbon. Many theoretical and practical applications for SWCNT S have been proposed due to their incredible physical properties. 12 The simplest way to envision a SWCNT is the roll ing up of a thin graphite sheet that is only one atom thick throughout. Furthermore, the angle at which this graphite sheet is rolled gives the resulting SWCNT certain characteristics which can be described by the unit vectors ( n,m ). These angles or chir alities are of three classes: armchair, zig zag, and chiral. Using this system, all nanotubes with an m=0 designation are zig zag, while nanotubes where n=m are armchair. When the difference between the unit vectors is a multiple of three, the SWCNT is c onsidered metallic with zero band gap, while all other SWCNTs are semi conducting with a finite band gap. 13 Figu re 2 1 shows the various angles and resulting SWCNT configurations. For example, in order to produce a metallic SWCNT with a st ructure ( 6,3 ) the graphite sheet would be rolled in such an angle that the atom labeled (0,0) would be superimposed on the ( 6, 3 ) atom. Regardless of their electronic configuration, SWCNTs share similar properties. Typically, SWCNTs can range in diameter from roughly 0.4 2.5 nm, while lengths are much more variable, ranging from the nanometer range to several micrometers. 14 17 Ideally, all SWCNTs contain only carbon atoms. SWCNTs have unique size related characteristics such as strength, elasticity, adsorption capacity, volume to surface area ratios, and controllable conductivity which mak e them attractive for industrial applications and incorporation into several commercial products. The low size/strength
19 sporting equipment, as well as increase fuel efficienc y of automobiles and aircraft. Both theoretical calculations and bench s tudies have shown that SWCNTs have a tensile strength roughly 100 times stronger than common steal. As such, SWCNTs have been added to a variety of industrial produced materials inclu ding cement, 18 and carbon fiber 19 t o increase their mechanical strength. Also, the electronic properties of SWCNTs have potential for use as selective probes, sensors, energy storage, and field emission devices. 12 Furthermore, continued research on conductive SWCNT may lead to the development of nanoscale electronic devices previously thought impossible due to the failure of conventional metal wires to conduct current at such small sizes. 20 2.2 SWCNT S ynthesis While the synthesis of SWCNTs is still a continually evolving process, any synthesis method must include i) a source of carbon, ii) a heat source, and iii ) the presence of a metal catalyst. 21 Unfortunately, metal catalysts are a necessity for the synthesis and production of SWCNTs. These can include iron (Fe) nickel (Ni) cobalt (Co) molybdenum (Mo) and a range of other metals. 22 While essential to their production, catalytic metals are usually abundant in unpurified SWCNTs 22 Currently, the three most widely used SWCNT synthesis methods are Arc Discharge, Laser Ablation, and Chemical Vapor Deposition (CVD) Arc Discharge technique s use two graphite rods each functioning as bo th electrodes and carbon source SWCNTs are produced through the evaporation of the graphite anode. A higher percentage of SWCNTs are produced through varying the location of the metallic catalysts within the system. 23 During Laser Ablation, the carbon source is vaporized by a focused laser in the presence of a high temperature inert gas. The res ulting SWCNTs
20 can be tuned to desired diameters by varying the gas flow rate, temperature, and the type of metallic catalysts 24 27 CVD process es rely on the heating of metallic substrate at temperatures between 900 1200 0 C, as hydrocarbon vapor is introduced. 21,28 While the synthesis mechanisms are debated in the CVD process, it is generally agreed that the catalytic decomposing of the precursor gas is followed by saturation of carbon within the metal catalyst resulti ng in CNT growth. 29 As the interest in SWCNT use in commercial applications has increased, resear chers have recently begun to study alternative production methods. One such method is HiPco a type of CVD process that can be scaled up to the industrial level producing a large amount of usable SWCNTs. 30 The HiPco process has become the most favored synthesis technique as it p roduces an exceedingly high percentage of SWCNTs (over 70%) with extremely tunable diameters based on the pressure of the carbon monoxide feedstock. 31 Other groups have begun to synthesize SWCN Ts using a flame heat source. This technique theoretically could produce large amount of SWCNTs at a fraction of the cost once refined. 32 2.3 SWCNT Stabilization At the nanoscale level, fundamental interactions are governed by electrostatic and van der Waals forces. After entering a charged environment, hydrophobic SWCNTs tend to quickly bundle or aggregate to achieve the most energetically stable confo rmation 33 This characteristic is problematic as the majority of applications require very well, if not individually dispersed SWCNTs. In order to ov ercome these attractive forces in aqueous environments, researchers have used two differe nt techniques to improve SWCNT dispersion and stability. The first technique utilizes a variety of chemical reactions in an attempt to increase the number of hydrophil ic groups on the surface of the nanotubes, thereby increasing their solubility 34 36 However, subsequent
21 studies have demonstrated that these chemical manipulations may alter many of the SWCNT intrinsic propertie s 37,38 The second stabilization option adopts a physical approach whereby the nanotubes are separated using stabilizing agents or surfactants. Surfactant molecules are chosen for their amphiphillic qualities, th e combination of both hydrophobic and hydrophilic groups within the same molecule. 33 The hydro phobic ends of these molecules intera ct with the surface of the nanotubes, wh ile the hydrophilic zones interact with the polar environment, shielding the nanotube from the overall aqueous envi ronment and each other. These sorption reactions do not chemically alter the surface of the nanotubes preserving their beneficial characteristics. A wide range of surfactants have been used in attempt s to stabilize SWCNTs including organic polymers 39,40 various ionic solvents such as sodium dodecyl sulfate (SD S) 41,42 and sodium dodec ylbenzene sulfonate ( SDBS ) 43 and organic solvents 44 A recent comparative study using both UV Vis and f luorescence spectroscopy determine d that the ionic s urfactant sodium deoxycholate ( DOC ) and organic solvent carboxymethyl cellulose (CBMC 250k) created the most stable dispersions of SWCNTs among the many tested. 45 The continued development of stabilization techniques will undoubtedly increase the number and different physical forms in which SWCNTs would potentially enter natural aquatic systems. It is therefore critical to n ote that SWCNTs may cycle through the environment differently based on specific surfactant induced modifications and the resulting stability of obtained suspensions. 2.4 SWCNT Characterization While both the methods of synthesis and stabilization of SWCNTs are important factors to consider when predicting their potential adverse environmental effects, the characterization of the both raw SWCNTs, and stabilized suspensions is equally
22 important. The ability to adequately characterize pristine SWCNTs, as well as to t rack structural changes as they move through natural systems is critical to any study, as structural changes can be linked to possible environmental responses. Several optical techniques have been used to characterize both raw SWCNTs, and aqueous s uspensions. The first such technique utilized by many researchers is Raman Spectroscopy. Raman Spectroscopy is an effective tool to characterize both individual and bundled SWCNTs. Figure 2 3 shows a typical Raman Spectra for unpurified SWCNTs purchased fr om an industrial supplier. 46 T he spectra can be broken down into three distinct sections, the Radial Breathing Mode (RBM), the D band, and the Tangential Modes or G Bands. Each region of the spectra can be used to extract important structural and electronic information on the SWCNTs be ing studied The RBM can be found at wavelengths between 120 and 250 cm 1 ( ) and can be related to the diameter of the particular SWCNTs ( dt ) through E quation 1 1; in which constants A and B vary based on the substrate used during analysis, as well as the aggregation state of the SWCNTs. 47 Parameter values can vary from 234 248 for A, and 0 10 for B. (RBM) = (1 1) Constant values become much more influential for diameters significantly smaller than 1 nm and larger than 2 nm. The larges t resulting peaks in the Raman s pectra of SWCNTs can be found in the t angential modes or G Bands. The G Bands are found in the range of 1550 to 1600 cm 1 and can include up to 6 different individually resolved peaks. The most intense peak in this series is labeled G + while the next intense peak is
23 labeled G In general, the G band has a very broad line shape in the presence of metallic SWCNTs. Samples containing a high proportion of semiconducting tubes will result in a small G band, but a very sharp G + band. A metallic like G band in the presence of only semiconducting SWCNTs can help identify the presence of metallic impurities remaining in the system. 47 Finally, the smallest of the three re gions identified is the D band, found in a range of 1300 to 1400 cm 1 D Band peaks should have an intensity no larger than 1/100 of the G Band. In several cases, large D Band signatures have been used to detect large amounts of amorphous carbon in the sam ple. 46 Using information gathe red from each region of a typical SWCNT Raman spectrum can help characterize the studied sample. Another technique commonly used to characterize SWCNT suspensions in aqueous media is photoluminescence (PL) and photoluminescence excitation (PLE). et al found that individually suspended SWCNTs in an aqueous surfactant solution have very large fluoresce intensities in the 8 00 1600 nm range. Furthermore, the same batches of SWCNTs have a similar signal, only varying in individual peak intensity as t he surfactant solution is modified. 48 Subsequent studies have found that each individual peak in th ese PL spectra is representative of a particular SWCNT structure (n,m) present in the suspension. 49 Figure 2 3 sh ows a spectrum of well dispersed SWCNTs in a 1 wt % SDS solution. Both PL and PLE spectra have been used as a tool to determine the stability and dispersion of SWCNT suspensions. A loss of intensity coupled with a broadening of individual peaks demonstra tes the formations of aggregates within the suspension. Suspensions are commonly rerun and compared wi th original optical spectra to detect changes in dispersion through time. It is also
24 important to recognize that during SWCNT suspension preparation a significant amount of bulk SWCNTs are not sufficiently stabilized and are discarded after centrifugation. As a result, the final concentration of SWCNTs in solution must be calculated after suspension preparation. This calculation can be easily done by fo 41 Using an extinction coefficient of 0.043 and the UV Vis absorbance reading at 763 nm SWCNT con centrations within the suspension can be calculated. Typi cally concentrations can range 15 30 ppm. While optical methods are very common during SWCNT experimentations, visual methods are also widely used, particularly S canning Electron Microscopy (SEM) an d T ransmission Electron Microcopy (TEM) imaging. Both SEM and TEM images are used extensively by toxicologist s to detect obvious structure changes to SWCNT surfaces after interaction with a test organism. 50,51 2 .5 Post Suspension Processing While the potential for the development of novel products that take advantage of the unique properties of SWCNTs is nearly limitless, widespread production of these products has been lacking due to the heterogeneous natu re of as produced SWCNT powders, containing SWCNT s pecies of different electronic type, diameter, and length. Advancements in the HiPco synthesis process have allowed for production of SWCNTs with v e ry narrow diameter distributions. 14,15 However, there remains great variability in the length of the produced SWCNTs ranging from a few hundred nm to several hundred m A s a result, inclusion of as produced SWCNTs into novel electronic devices is challenging T hus, time consuming, and costly post synthesis processing steps are
25 needed to produce SWCNT fractions of discrete length and diameter. One of the most effective met hod s for length separations of as produced SWCNTs has been Size Exclusion Chromatography (SEC) using porous silica beads. 52 Other methods for SWCNT length separation include field flow fractionation, 53 ultracentrifugation, 54 and selective precipitation. 55 Another obstacle for the development of novel electronics using SWCNT s is the inclusion of both m and s SWCNT types in as produce d batches. Manufactures desire very pure fractions of m SWNTs to conduct electricity, and s SWCNTs with well defined band gaps for use as switches. A variety of methods have been developed to se parate SWCNT suspensions by electronic type, each with variability i n purity, efficiency, and cost. These methods include Density Gradient Ultracentrifugation (DGU), 56,57 gel electr ophoresis, 58,59 selective oxidation, 60,61 or selective wrapping with a variety of molecules. 62 64 While each of these methods are effective at separating m and s SWCNT fractions, they are all limit ed on an industrial scale due to equipment needs and associated cost. T he most promising method for the large scale separation of SWCNT by electric type has been the selective adsorption on hydrogels pioneered by Kataura and coworkers. 65 67 In this method, SWCNT stabilized with SDS are passed through colum ns packed with either dextran or aga r ose. After injection, m SWCNT elute isocraticly with a 1 wt % SDS solution. s SWCNTs are only eluted from the system once th e eluent is changed to high concentrations of SDS (> 3 wt %) or a bile salt such as sodium c holate (S C ) or sodium deoxycholate ( DOC ) While this method is highly effective, little is currently understood about the driving forces of separation and selectivi ty. Advancements to this method have begun to resul t in diameter separations
26 using co surfactant solutions, 68 and single chirality fractions of s SWCNTs 69 through the use of temperature 70 or pH modifications 71 2.6 SWCNT Toxicity After examining the physicochemical ch anges possible from the release of SWCNTs into the environment the next step in a risk assessment framework is establishing the potential toxicity these particles may cause through their transport and interaction with organisms. While there has been a gr eat deal of work done on the potential toxicity of NPs in the environment, there is still confusion as to their ability to cause serious threat s to living organisms 72 76 For instance, based on experimental data a vailable in the literature, o ne set of studies might point to the severe acute and chronic toxicit ies of a particular NP while similar stud ies by different research group s do not lead to the same conclusions Assessing the potential toxicity of NPs in the environment becomes even more difficult when the large variety of surface modifications and chemic al surfactants used to increase NP s uspension in aqueous media are taken into consideration Although one can point to several reasons for the above mentione d contradictory results it is likely that the lack of standardized methodologies for these emerging contaminants plays a significant role Furthermore, the variability in the character of as produced SWCNT materials (i.e, stabilizing agents, length, diame ter, electronic type, etc ) adds complexity to any toxicological assessment. With no agreement amongst toxicologist s on how to test the potential consequences of NP, different research groups are left to determine which traditional test best fits their needs as well as what initial SWCNT material to use in their studies I n some cases, i t is possible that some of the traditional toxicity methods might not be adequate for NMs
27 The following is an attempt to summarize relevant info rmation that is currently available in peer reviewed literature on the toxicity of SWCNT s 2.6.1 Cytotoxicity When considering the potential impact of NPs on organisms and ecosystem functions one of the first risk assessment tools is toxicity screening with both model test organisms and mammalian cell lines. The literature on the cyto toxicity of SWCNTs is abundant, and varies in both its results and reported mechanis ms of toxicity. Most studies agree that SWCNTs have antimicrobial properties, with the active mechanism of toxicity being the production of reactive oxygen species (ROS.) 77 However, other rese arch suggests that while high purity SWCNTs are antimicrobial the mechanism of toxicity is direct contact of SWCNTs with th e cell membrane resulting in puncture and destruction 78 The diameter of SWC NTs used also has an effect on the antimicrobial properties 79 The cylindrical shape of SWCNTs may also contribute to their antimicrobial activity as they were found to be the more toxic to commonly found bacteria in wastewater streams when compared to other fullerene structures 80 Other studies indicate d that while physical pa rameters influence d antimicrobial activity, the drive r of the toxic response was actually the electronic nature of the SWCNTs. Test using E. Coli as a test organism have found m SWNCTs to be much more toxic than their s SWCNT counterparts. 81 It appears that SWCNTs will have a strong antimi crobial effect after entering the natural environment. At the same time, th e antimicrobial property of SWCNTs may be harnessed to develop antiseptic surfaces. Several studies have also been conducted with a variety of mammalian cell lines to evaluate the potential toxicity of SWCNTs with particular interest in development of biomedical application s and drug delivery systems. For example, studies using differ ent
28 SWCNTs and preparation techniques demonstrated significant inflammation in mice lungs. 82 84 Other studies utilizing cell lines from rat livers found that SWCNT induced no toxic response unless they were stabilized with a known toxic surfactant, as surfactant identity played a major role in the observed bio logical response s 85 Similar results using mice macrophages have also been published, demonstrating that SWCNT dispers ed with n on toxic surfactant do not cause a negative biological response 86 The authors did state that while the surfactants had no effect on the toxicity results, there were significant differences to the o ptical properties of the dispersed SWNCTs base on surfactant used in stabilization. 86 While surfactant stabilizers play a role in potential toxicity other studies suggest the unique shape of SWNCTs may allow for cell membrane puncture and increased toxicity. In a comparative study using the P12 Cell li ne derived from male rats, acid washed SWCNTs showed increase d toxicity over graphene counterparts, although both produced significant ROS in dose response studies. 87 A gain, this brief review of cytotoxic effects of SWCNTs demonstrates the need for a well defined procedure when studying the biological impacts of NPs as the level of purification and stabilization techniques p lay s a major role in their potential toxicity. 2.6.2 Aquatic Organisms It has been documented that the aquatic environment will function as a primary sink for SWCNTs due to the large amount of research conducted on making them well suspended in aqueous s olutions for industrial applications. S tudies using the freshwater green algae P seudokirchneriella s ubcapitata as a test organism have found that water chemistry parameters greatly affect the biological response of this organism
29 when exposed to SWCNTs. Yo un et al. demonstrated that toxic doses of surfactant dispersed SWCNTs c ould be mitigated with either a high concentration of gum a rabic (GA) or by increasing the concentrations of natural organic m atter (NOM) in the culture media 88 A more recent study has shown that ionic strength and pH do play a major role in SWCNT mobility in natural waters and found that the toxicity of SWCNT is driven by their photo activity and subsequent ROS generation and not toxic metal impurities as suggested by others 89 The water flea Daphnia magna is also a common test organism used to evaluate the potential toxicity of SWCNT released in aquatic environments Roberts et al found acute toxicity of SWCNT only when concentrations were much higher than those that would enter the natural environment 90 while Kim et al showed that a combination of these SWCNTs with soluble copper commonly present in the water magnified their toxicity. 91 The effect of SWCNTs on D magna is also affected by the stabilization method used in SWCNT suspension preparation, with significant differences found between suspensions prepared with non covalen t stabilizer (i.e. surfactant) and functionalized sidewalls. 92 Both rainbow trout and zebra fish have been used in a variety of studies to assess the impact of SWCNT on aquatic vertebrates These studies ha ve resulted in a wide range of responses based on the cha racteristics of the initial SWCNT suspensions and life cycle of the test organism. Recent research has shown that direct injection of SWCNT suspensions into adult zebra fish activates the organism s immune response, but does not result in immobility or m ortality. 93 However, previous studies have shown that other SWCNT suspensions have delayed hatching of zebra f ish embryos, even
30 when 90% of metallic impurities were removed. 94 Studies using rainbow trout as test organism are just as c onflicting. Early research suggested that SWCNTs act as a potential respiratory toxin to rainbow trout, but also raised other major questions regarding the SWCNT s potential neurotoxicity with long term effects on the cell cycle 95 However, other studies have shown that at the anticipated environmentally relevant concentrations (0.25 mg/L) of SDS SWCNTs show no i nhibition of adult rainbow trout over a 10 day exposure period. The result of each of these studies further highlight how a lack of standardization in the toxicity evaluation process for SWCNTs causes confusion about t heir potential harm if introduced to natural environmental compartments. 2.6.3 Human Cell Lines While significant work has been done attempting to quantify the risk of SWCNT release into the environmental, the potential effects to human health are a l s o of concern. Risk assessment of human exposure to SWCNTs has been completed on a variety of cell lines, particularly human lung cells as initial fears linked SWCNT structure with asbestos, raising concern for exposure of laboratory personal during SW CNT synthesis Chromosome damage at low SWCNT doses has been reported on human epithelial cells, 96 while other studies using A549 human lungs cells resulted in little to no acute toxicity, needing concentrations of nearly 800 record an adverse biological response 97 Other re searchers have found that surface oxidized SWCNTs from acid washing are much more toxic to epithelial cells than suspensions made from the same material without acidic purification. 98 Most troubling, some researchers have suggested that negative effects attributed to direct SWCNT exposure
31 may be incorrect. The observed response may be a byproduct of the SWCNTs inducing changes in the cell culture me dium misrepresenting their toxicity 99 Once again, the results of these studies h ave been inconsistent as there is no standard proce dure for toxicity analysis. A variety of other human cells lines have been used in the risk assessment of SWCNTs. Increases in ROS have been seen in human intestinal cells at doses over 100 ppm, much hi gher than would be expected in any real world scenario. 100 However, toxicity has been reported in human astrocytoma cells at concentrations as low as 0.5 ppm, but this effect may have been caused by the use of surfactants including SDS and SDBS, with known cytotoxic characteristics 101 Other studies using human keratinocytes have shown that acid purified SWCNTs devoid of metallic impurities are create much less toxicity than the same SWCNTs that have not been purified. 102 2.7 Remaining Questions and Research Goals It is clear that the po tential harm of SWCNT release to environmental systems i s controversial. A summary of result s curren tly available in the peer reviewed literature shows a lack of consensus p rimarily due to the absence of unified and standardized toxicity assays specific to these emerging NP contaminates. Furthermore, studies that h ave reported a toxic response to test organisms exposed to SWCNTs lack consistency in the reported mechanisms of toxicity citing the production of ROS, membrane puncture, and DNA damage, to name a few In addition to the lack of an adequate methodology for toxicity testing, t he variability see n in published data on the toxicity of SWCNTs may also be explained by the heterogeneous nature of the initial SWCNTs suspensions used in these different studies. These observation s suggest that in some cases, true cross study comparisons are not possible as all SWCNT suspension do not
32 have the same properties. For instance several SWCNT suspensions prepared using the same starting material may contain individual SWCNTs with variability in length, diameter, electronic character (i.e. m or s SWCNTs) and su rfactant structure along the sidewalls. While a great deal of work has been completed attempting to elucidate the effects of SWCNT length and diameter on biological response, very little attention has been given to how the electronic nature of the nano tube s (i.e m or s SWNCTs) effects biological response s This gap in knowledge persists as methods to separate the m SWNCT and s SWCNT species from starting material has not been fully developed and made available to all research groups Initial separation processes including DGU require extensive equipment costs that may be prohibitive to most research efforts. Fortunately, a new chromatographic method of SWCNT type separation was developed by Kataura and coworkers that significantly decreased cost and inc reased throughput. 65,66 Initially, the purity and concentration of these resulting suspensions has not been ideal for use in toxicological studies. In this dissertati on, a two step approach is used to fill the above outlined knowledge gaps and to fundamentally understand how different SWCNTs forms such as m and s SWCNT species may impact SWCNT organism interactions. First, a process developed by Kataura and coworker s was used to achieve a fundamental mechanistic understanding of the driving forces at play during the separation of m and s SWCNTs from parent materials. Insights gained from this mechanistic study were then applied to optimize the separation process to produce highly purified m and s SWCNT fractions; ideal for the determination of potential toxicity. Finally, the produced purified fractions were used in a series of bioassays to
33 determine differences in biological responses using model test organisms in dose expos ur e stu d ies. Obtained results provide insight into the potential effects SWCNT s might have if release d to natural aquatic systems. The determination of d ifferences in biological response s of organisms exposed to either m SWCNTs or s SWCNTs, if any, will be instrumental in the development of guidelines for the disposal of products containing SWCNTs. Accordingly, this research could be seen as a step towards an environmentally conscious design and regulations to guide discharge of SWCNTs as comme rcial products containing them enter the market place and reach the end of their life cycle
34 Figure 2 1. 2D graphite sheet used to define SWCNT vector and classification Three classed of SWCNT are shown including zig zag (m=0), armchair (n=m) and chairal (all others). Additionally, when the difference of n and m is a multiple of three, the SWCNT is metallic with no band gap
35 Figure 2 2 Typical Raman Spectra of unpurified SWCNTs. The spectra can be broken down into three distinct sections, the Radial Breathing Mode (RBM), the D band, and the Tangential Modes or G Bands. Each region of the spectra can be used to extract important structural and electronic information on the SWCNTs being studie d
36 Figure 2 3 Typical fluorescence spectra of a 1 wt % SDS SWCNT suspension following high shear mixing, sonication, and ultracentrifugation. Spectra are recorded using excitation wavelengths of A) 662 nm and B) 784 nm. A B
37 CHAPTER 3 INTERACTIV E FORCES BETWEEN SODIU M DODECLY SULFATE SUSPENDED SINGLE WALL CARBON NANOTUBE S AND AGAROSE GELS 3.1 Background Since the discovery of single walled carbon nanotubes (SWCNTs), researchers have envisioned many applications that take advantage of their ast ounding physical properties. 12 Conceptually, SWCNTs are a single atom thick sheet of graphene that is rolled into a s eamless cylinder. The angle of this roll is de fi ned by the unit vectors ( n m ), which gives rise to SWCNTs with speci fi c properties governed by the crystalline structure. When the di ff erence between the values of n and m are divisible by 3, the SWCNTs are metallic (m); otherwise, the SWCNTs are semiconducting (s) with a de fi ned band gap. 13 Theoretically, m SWCNTs s hould make up a third of all possible nanotubes. Currently, all SWCNT synthetic approaches produce a variety of SWCNT ( n m ) types that limit their use in many applications. 104 Although considerable progress has been made in controlling the diameters and types of SWCNTs produced 14,15 a variety of post synthesis separations are still required to produce nanotubes of speci fi c length, diameter, and electronic type (i.e., purel y m or s SWCNTs). 16,17 Of particular interest over the last several years has been the development of a simple and scalable technique for the separation of SWCNTs. Several methods have been used to separate SWCNTs by chirality or by electronic type, including density gradient ultracentrifugation 56,57 gel electrophoresis, 58,59 selective oxidation 60,61 and selective wrapping with DNA 62,105 polymers 63,106,107 and amines. 64,108 While e ach of Reprinted with permission from The Journal of American Chemical Society, Vol. 135, Issue 47, pp. 17653 18006, 2013 Copyright 2013 American Chemical Society
38 these techniques are capable of separating the m and s SWCNTs with varying levels of success, selective adsorption on agarose or dextran gels, which was pioneered by Kataura and co workers, 65,66 is currently one of the most promising methods for large scale, high throughput separations. While the use of agarose gel columns has been e ff ective in separating m and s SWCNTs, little is understood about the mechanism. Our prior study 109 proposed that the mechanism was due to selective retention of s SWCNTs, which was later confi rmed by the work of Tvrdy et al 69 As highlighted in our previous study, the selective adsorption was controlled by the packing of sodium dodecyl sulfate (SDS) molecules arou nd SWCNTs. 109 This mechanism implies inherent differences in surfactant structure around m and s SWCNTs in the suspension. Indeed, other researchers have observed different buoyancies for m and s SWCNTs, 57 which suggests differences in surfactant coverage for each SWCNT type. Molecular dynamics simulations have also shown that different surfactant structu res are formed around specific (n, m) types. 42,110 In this paper, we aim to understand the nature of intermolecular forces that yield selective adsorption of SDS SWCNTs onto agarose gel. The heterogeneous interface of the SDS SWCNT complexes (i.e., the coexistence of hydrophobic and hydrophilic patches), the dyna mic nature of SDS molecules on SWCNTs, and the complex micro and macrostructure of agarose gels allows for a variety of potential interactive forces (e.g., ionic, van der Waals, hydrophobic). In order to probe the nature of these interactions, the contrib ution from each of these forces are either inhibited or enhanced to determine their relative importance in the selective retention of s SWCNTs onto agarose. By understanding the interaction of SWCNTs with the agarose gel under both
39 equilibrium and nonequil ibrium conditions, we aim to identify the primary force responsible for selective adsorption. Furthermore, we discuss possible reasons why SDS SWCNTs show a larger affinity for agarose gels than SWCNTs coated with other surfactants, such as sodium cholate (SC), as well as a description of the active adsorption sites within the agarose gels. This knowledge should lead to more efficient separations of SWCNTs. 3.2 Methods 3.2.1 Materials Nanopure water was used in all experiments. SDS (>99%) and SC (>99%) wer e purchased from Sigma Aldrich (St.Louis, MO) and used as received. HiPco SWCNTs were obtained from Rice University (Rice HPR 162.3) and used as received. Di ff erent stationary phases were used as the adsorbent, including plain agarose (Sepharose 6 and 4 FF) and agarose function alized with either hydrophobic aliphatic carbon chains (butyl and octyl Sepharose 4 FF), hydrophobic phenyl groups (phenyl Sepharose 6 FF) at both low and high substitution (LS and HS, respectively), or ionic groups (sp and Q Sepharose 6 FF). All the gel s were manufactured by GE. Phenyl Sepharose HS was purchased from Sigma Aldrich, whereas the other gels were obtained from GE Health Care. The average diameter of all the gel beads was 95 m. Table 3 1 summarizes the relevant properties of the gel media used as the stationary phase. It is important to note that Sepharose 4 an d 6 FF are highly cross linked in comparison to the Sepharose 6B used in many studies, providing a more rigid structure.
40 3.2.2 Aqueous SWCNT Suspensions Aqueous suspensions of SWCNTs were prepared as described in our previous work. 109 Briefly 30 40 mg of raw SWCNTs was added to 100 mL of a 1 wt % SDS solution in Nanopure water. The solution was then homogenized for 30 min (IKA T 25 Ultra Turrax) and ultrasonicated (Misonix S3000) for 10 min (120 W) to aid dispersion. After ultrasonication, the resulting mixture was ultracentrifuged (Beckman Coulter Optima L 80 K, SW 28 rotor) at 20 000 rpm (53000 g ). Ultracentrifugation times varied for nonequilibrium (4 h) and equilibrium experiments (1 h) to produce the desired concentration of SWCNTs. A comparison of the absorbance and fl uorescence spectra for both SWCNT suspensions is shown in Figure A 1 ( Appendix A ). Although there is some broadening in the absorption spectra of SWCNTs used in the equilibr ium studies, both suspensions give intense fl uorescence, suggesting a high quality dispersion. 3.2.3 Equilibrium Adsorption The agarose gels were thoroughly rinsed with Nanopure water to remove any residual ethanol used as a preservative prior to their us e. The rinsed gels were then equilibrated with a 1 wt % SDS solution (SDS solution/gel volume ratio of 2:1). Approximately 500 L of surfactant stabilized gel was used for each replicate in separate 15 mL centrifuge tubes. Individual replicates were equilibrated with various concentrations of SWCNT in a constant background solution of 1 wt % SDS. After all components were combined the samples were mixed in a vortex stirrer for 30 s before being placed in a water bath held at 25 C for 24 h to ensure equilibration. After stabilization, samples were centrifuged for 5 min at 5000 rpm to remove any agarose
41 beads from the supernatant. An aliquot of the supernatant (300 L) was then extracted and analyzed by absorption and fl uorescence spectroscopy, as described below. 3.2.4 Nonequilibrium Adsorption Columns were packed with di ff erent compositions of agarose beads up to 6 cm in height. The columns were fi rst stabilized w ith at least fi ve column volumes (CV), approximately 43 mL, of 1 wt % SDS solution. For the experiments with the IEX media, equilibration required more volume; these columns were stabilized with 16 CV. Half a column volume of the suspension was then inject ed into the column. The early fractions of SWCNTs were eluted with 1 CV of 1 wt % SDS solution followed by two CV of 2 wt % SC solution to remove the retained SWCNTs from the column. Each fraction was then characterized by absorption and fl uorescence spect roscopy, as described below. 3.2.5 SWCNT Characterization The initial SWCNT suspensions and the supernatant extracted from the equilibrium studies were characterized by absorption (0.4 cm path) and fl uorescence (1 cm path) spectroscopy on an Applied NanoFl uorescence Nanospectralyzer (Houston, TX) with excitation from 662 and 784 nm diode lasers. E ffl uent from the column was continuously characterized in situ by use of a fl ow cell from Starn a Cells as previously described. Typically, absorption spectra were taken every 20 s as the e ffl uent fl owed through the cell. Mass fractions of SWCNTs eluted during were estimated by use of absorbance values at 626 nm, where the extinction coe ffi cient was calculated on the basis of the one determined by Moore et al 111 at 763 nm (see Appendix A ). The distribution of SWCNT lengths was measured by atomic force microscopy (AFM) following the procedure published by Khripin et al. 52 Silicon wafers were functionalized with 3 (ethoxydimethylsilyl) propylamin e (APDMES, Sigma Aldrich, St. Louis, MO).
42 Before deposition on the substrate, SWCNT samples in 1 wt % SDS were diluted at least 100 with an aqueous solution of 0.2 wt % SC and 20 mM NaSCN. SWCNTs were deposited on the wafer by casting 10 L of the sample, followed by incubation in a closed container for 6 min. After incubation, samples were dried by use of canned nitrogen. Several images were acquired on a Bruker Dimension Icon AFM in the peak force tapping mode (ScanAsyst) with the respective ScanAsyst Ai r probes. By this method, the average SWCNT length was calculated to be 467 nm. A representative AFM image and length distribution histogram are shown in Figure A 2 ( Appendix A ). 3.3. Result and Discussion 3.3.1 Physical and Chemical Structure of Agarose Gels The agarose polymer is the major gelling constituent of agar and contains agarobiose as the monomeric unit 112 as Figure 3 1a sh ows. The porous 3D structure of agarose is due to the self assembly of molecules at the nano and microscale. 112 At the nanoscale, si ngle strands of agarose form double helices that are stabilized by intra and intermolecular hydrogen bonds. Further aggregation of individual helices into bundles of various size and structure result in the characteristic porous 3D nature of agarose (see Figure 3 1b). The porous structure of the resulting beads is dependent upon the concentration of agarose used during production. The pore size decreases slightly as the concentration of agarose used in production increases from 4% to 6%. For example, the a verage pore size for Sepharose 4 and 6 FF is 45 and 29 nm, respectively. 113 Furthermore, electric birefringence studies have reported large, p ermanent dipoles in the range 10 3 10 6 D for agarose. 114 117 Importantly, these studies showed that domains of di ff erent size within agarose align at various time scales when
43 placed in an electric fi eld, suggesting di ff erent dipole moments. This ob servation indicates that agarose gels are formed by the nonuniform and random network shown in Figure 3 1b. Finally, the size of the domains aligned by the applied electric fi eld and, consequently, the dipole moments changed as the concentration of agarose was altered. Initially, the attractiveness of agarose as a size exclusion (SEC) medium in biochemistry stemmed from its hydrophilicity (large number of ether and hydroxide groups), its stability in a wide range of pH values, and concomitant neutrality tha t minimizes the nonspeci fi c binding of proteins. However, residual charges on the surface as well as hydrophobic groups from the manufacturing process can potentially exist, promoting protein binding to the gel according to their hydrophobic character or t heir charge density. 118,119 Consequently, residual contaminant moieties could in fl uence the adsorption of SDS SWCNTs despite pure agarose being neutral and hydrophilic. Although agarose can potentially have small regions of hydrophobic and ionic groups that o ff er selective adsorption, the surface can be additionally functionalized with either hydrophobic or ionic groups to further promote these interactions and aid separation. The glycidyl ether chemistry couples the ligand group to the agarose matrix by reacting with the hydroxyl groups on the backbone, resulting in the structures depicted in Figure 3 1c. 120 It is critical to not e that any functionalization to the base agarose matrix (i.e., phenyl, octyl, etc.) will not signi fi cantly alter the average pore size (see Table 3 1). However, functionalization of the agarose base may alter the magnitude of the permanent dipole moments, which could a ff ect separation.
44 3.3.2 Retention of SDS SWCNTs on Agarose Gels The a ffi nity of agarose toward SDS SWCNTs has already been exploited for the separation of nanotubes. 58,65 67,121 Figure 3 2a shows the elution curves of SDS SWCNTs as they pass through a column packed with one of the base materials (Sepharose 6 FF). Similar to prior experiments with Sepharose 6B, a large percentage of SWCNTs are adsorbed onto the agarose beads. The fraction of nanotubes that pass through the column (peak P1) is highly enri ched in m SWCNTs, as shown by the absorbance spectra in Figure 3 2b, while the nanotubes eluted with 2% SC (peak P2) are primarily s SWCNTs. In general, all the unfunctionalized agarose beads (4 FF, 6 FF, and 6B) are capable of separating SWCNTs. The most appreciable di ff erence observed among the di ff erent variants of agarose is in the shape of ad sorption isotherms (see Figure A 3 in Appendix A ). The di ff erence in adsorption isotherm shape behavior between Sepharose 6 FF and 4 FF con fi rms an inverse relatio nship between the concentration of agarose in the gel matrix and SDS SWCNT retention. 122 This may seem counterintuitive at fi rst, but a possible explanation lies in the different structures (see Figure 3 1b) formed during synthesis, as supported by electric bir efringence studies. 114 117 When one attempts to understand the fundamental mechanisms responsible for the separation of nanotubes via agarose gels, the physical parameters of SWCNTs and agarose must be considered. If one considers the average pore size of the base agarose (45 nm or less) as well as the average length of the SWCNTs used in this study (>100 nm as shown in Figure A 2, Appendix A ), it is unlikely that a significant amount of SWCNTs diffuse into the pore s of the agarose gel. Given that the bead size
45 is the same for all agarose used in this study and that the hydrophobicity of a surface does not significantly aff ect the slip plane, 123 no h ydrodynamic effects should be responsible for changes in retention either. As the pore size of the beads is primarily determined by the percentage of agarose used during production, any functionalization of the backbone does not significantly alter the por e size (see Table 3 1). The agarose gels are equilibrated with significant amounts of SDS (5 CV) prior to separation, so any interaction between the surfactant and functional groups cannot alter the dispersion properties of SWCNTs. Therefore, modification to agarose should affect only the interaction of SWCNTs with the outer surface of the agarose beads. The fact that SDS SWCNT separations are effective with both the beaded and nonbeaded gel forms of agarose supports the assertion that the selective retention of s SWCNT must be governed by surface interaction and not transport through pores. 3.3.3 Probing the Interaction o f SWCNTs with Agarose As described above, the structure of agarose used for these column based separations is complex. The hydrophilic regions are represented by the ubiquitous hydroxyl groups (each monomer contains four OH groups) depicted in Figure 3 1 a, while the potential hydrophobic regions include residues from synthesis. More importantly, these OH groups are also highly polarizable, and it is reasonable that the permanent dipoles observed in electric birefringence measurements 114 117 are associated with these groups. The existence of these dipole regions creates the potential for an attractive interaction with an approaching charge. The interface of SDS SWCNTs is equally intricate. The structure of the surfa ctant shell around nanotubes is dynamic, not well de fi ned, and is expected to be
46 heterogeneous, with some areas of the SWCNT completely exposed to the medium. Hence, the SWCNT interface might present distinct hydrophobic and hydrophilic regions that provid e the possibility of di ff erent interactions with agarose. Simulation studies have shown that the structure of the surfactant shell depends on concentration, 42,110 while buoyancy di ff erences suggest structural variability in surfactant coverage based on the metallic or sem iconducting nature of the SWCNTs. 57,124 Finally, SDS molecules bound to the SWCNTs are highly mobile, as demonstrated by the ability of SDS molecules t o change their assembly under di ff erent mechanical and chemical stimuli, such as shearing, 125 uptake of nonpolar compounds, 109,126 or changes in the ionic strength of the medium. 57,109,124 Given the physical and chemical structure of agarose as well as the complicated interface of surfactant suspended SWCNTs, both long and short range interactions are possible. However, only a fi nite numb er of interactions can exist between the agarose and SWCNTs once their structures are considered. The hydrophobic regions of both the agarose and SWCNTs may yield weak, short range forces important to the separation, such as van der Waals (vdW), hydrophobi c, and interactions. On the other hand, the hydrophilic groups on each enable strong, long range forces, such as ionic or ion dipole interactions. While multiple interactions may be occurring between SWCNTs and agarose, the relative importance of a given forc e can be evaluated by either inhibiting or promoting its signi fi cance during adsorption. We begin by investigating the role of vdW forces by inhibiting ionic interactions. Next, ionic, hydrophobic, and interactions are promoted by systematically modify ing the surface of the stationary phase with aliphatic, phenyl, and charged groups, as shown in Figure 1c.
47 184.108.40.206 Adsorption p redominantly through van der Waals i nteractions While vdW forces are generally weaker than ionic interactions, they are additive along the length of the nanotube in these systems and can provide significant adsorption onto a surface. Furthermore, recent studies using Lifshitz theory quantified the differences in vdW forces between SWCNTs of diffe rent type and chirality. 127,128 These theor etical studies reported significant differences in vdW potentials between m and s SWCNTs, 127 as well as increased attraction of s SWCNTs over m SWCNTs toward polymer surfaces. 128 Although the agarose gels used here have consid erable differences from the polymers in those studies, vdW interactions could be a driving force during the gel based separation of SDS SWCNTs. To investigate the relative importance of vdW interactions during separation, equilibrium studies were conducted with Sepharose 4 and 6 FF. Each gel was equilibrated with identical concentrations of SDS suspended SWCNTs with an increasing electrol yte background up to 80 mM NaCl. Increasing the ionic strength of the solution has several important e ff ects. Most importantly, the increase in charge screening drastically compresses the size of the electric double layer. This compression serves to minimi ze the range and intensity of the strong electrostatic interactions (i.e., ionic or dipole), thereby enhancing the importance of the short range vdW forces. Therefore, if vdW attractive potentials were the primary force driving the retention in these syste ms, an increase in electrolyte concentration should not reduce SDS SWCNT retention. However, Figure 3 3 shows an inverse relationship between electrolyte concentration and retention of SDS SWCNTs. Interestingly, a decrease in retention was also observed du ring our previous mechanistic study in nonequilibrium
48 (column) experiments. 109 While changes to the ionic strength of SDS SWCNT suspensions can alter the structure of the surfactant shell on SWCNT sidewalls, 57,124,129 these transitions were shown to slowly progress as the elec trolyte concentration increased. 57 The monotonic decrease in retention, especially at low salt concentration, indicates that vdW interactions are not the primary force driving the separation of SDS SWCNTs in agarose. Finally, the observed di ff erences in vdW forces between m and s SWCNTs calculated by Lifshitz theory would exist regardless of the surfactant used to stabilize the suspension. The fact that other surfactants do not yield selective adso rption further indicates that vdW forces are not the dominant interaction during separation. 220.127.116.11 Adsorption p redominantly through ionic i nteractions After exclusion of vdW forces, stronger electrostatic interactions are an attractive option, when one considers the potential existence of charged groups on the agarose surface and the charged nature of SDS headgroups used to disperse SWCNTs in solution. Ionic interactions were explored by using agarose beads functionalized with sulfopropyl (sp) and quater nary ammonium (Q) groups, bearing negative and positive charges, respectively, under the working pH conditions (between 6 and 8). Figure 3 4 compares the retention behavior of plain agarose to that of the same gel after addition of ionic groups. The retent ion behavior of Sepharose with charged groups on the backbone is entirely di ff erent than that of the base agarose (control). Negatively charged sp Sepharose reduces the retention of SWCNTs. The behavior of sp Sepharose is not surprising since the negative charges on the gel repel the negative charges on the SDS coated SWCNTs. One interesting observation is that, even though
49 the surface of sp Sepharose is negatively charged, 15% of SDS SWCNTs are still adsorbed, suggesting that there are adsorption sites st rong enough to compete with the repulsion from the negative groups. Interestingly, Q Sepharose also has little retention of SWCNTs despite bearing positive surface charges. It is important to note that the common procedure for all column separations is to equilibrate the surface with the surfactant prior to SWCNT injection. Any positive charges that exist on the agarose (backbone or functional groups) would then be covered with SDS during equilibration. Therefore, the lack of adsorption simply shows that SD S is not extensively displaced by SDS coated SWCNTs. We do note that a small portion of the SWCNTs are irreversibly retained, suggesting that some SWCNTs are able to strongly adsorb onto the surface. In general, there is no driving force for SDS SWCNTs to displace SDS molecules from the surface. If either charge were responsible for the retention of SDS SWCNTs, increasing the number density of that charge would have increased the amount of SWCNTs adsorbed. Since a reduction in retention was observed, electr ostatic attraction (ion exchange) is not the dominant interaction between SDS SWCNT complexes and agarose. 18.104.22.168 Adsorption p redominantly t hrough hydrophobic i nteractions The importance of hydrophobic interactions was tested by using a set of Sepharose beads that have been functionalized with butyl and octyl groups. Figure 3 5 compares the retention and adsorption behavior of plain, butyl and octyl Sepharose under equilibrium and nonequilibrium conditions. Once the Sepharose is functionalized with alip hatic groups, both nonequilibrium and equilibrium adsorption studies show that retention of SDS SWCNTs decreases substantially. During column experiments with
50 butyl Sepharose, only 27% of the injected SWCNTs are retained on the column. This low adsorption a ffi nity is con fi rmed by the limited retention (16 g/g) shown in equilibrium experiments. Decr eased retention is also evident in the octyl Sepharose systems (65% and 45 g/g). The resulting absorbance spectra from none quilibrium studies (see Figure A 5 in Appendix A ) also demonstrate a nearly complete loss of selectivity once Sepharose 4 FF is func tionalized. Finally, it is important to note that functionalization of the base gel creates drastic changes to the shape of the adsorption isotherms presented in Figure 3 5. For example, the multiple plateaus evident in the isotherm for Sepharose 4 FF beco me more distinct after functionalization with butyl and octyl groups. The implications of these changes will be discussed later. The retention characteristics of SWCNTs from both equilibrium and nonequilibrium studies are consistent. Since a systematic inc rease in the concentration and density of groups that promote hydrophobic interactions do not increase SDS SWCNT retention in either system, selective retention is not due to hydrophobic interactions between SDS SWCNTs and agarose. 22.214.171.124 Adsorption p redo minantly through i nteractions The impact of interactions on the adsorption of SDS SWCNTs by agarose was assessed by use of agarose beads functionalized with phenyl groups (phenyl Sepharose). While the addition of phenyl groups to the agarose backb one increases the hydrophobicity of the matrix, phenyl groups also establish interactions with SWCNTs 130 Different degrees of phenyl substitution (see Table 3 1) were used to observe the changes in adsorption behavior as the concentration of phenyl groups on Sepharose 6 FF was changed from 3 6 shows the retention
51 characteristics of SDS SWCNTs on phenyl Sepharose. Similar to the studies of hydrophobic interactions, both nonequilibrium and equilibrium studies of phenyl substituted agarose showed decreases in retention. These decreases in retention, however, were dependent on the surface concentration of the functional groups. At low substitution, phenyl Sepharose retained 54% of SDS SWCNTs, whereas it retained only 25% at high substitution (see Figure 3 6a). This reducti on in retention by almost half occurs while the ligand density is nearly doubled matrix is functionalized with phenyl Sepharose at either degree of substitution, the adsorption isotherms change dramatically (see Figure 3 6b d ). The slope of the isotherms is shallow in comparison with that of Sepharose 6 FF, showing decreased affinity for the surface. These isotherms also show unique nonmonotonic shapes not seen in the unfunctionalized agarose beads, which will be discussed lat er. Similar to the studies of agarose functionalized with aliphatic groups, the absorbance spectra show that any retention that is occurrin g is not selective (see Figure A 6 in Appendix A ). The results from both equilibrium and nonequilibrium studies are s imilar regarding the potential of interaction to drive the separation of SWCNTs in these systems. A systematic increase to the density of phenyl groups on the surface of the gel did not increase SDS SWCNT retention in either system. Therefore, inte ractions between SDS SWCNTs and agarose are not the primary driving force for selective adsorption of SDS SWCNTs. 3.3.4 Nature of Adsorption between SWCNTs and Agarose
52 3.3.4 .1 Role of i on dipole interactions in selective a dsorption Direct measurement of the extent of ion dipole interactions between SDS SWCNTs and agarose gels is di ffi cult for several reasons. By their very nature, ion dipole interactions are mixed systems; therefore, suppressing other forces without a ff ecting ion dipole interactions is un likely. Enhancing their interaction by directly manipulating the charge density of ions (SDS SWCNTs) or dipoles (agarose) is also not feasible. SDS SWCNTs are already coated with a substantial amount of anionic charges and essentially act as macro ions. As a result, an attempt to increase the charge density of SDS on SWCNTs will likely alter the structure of the surfactant on the sidewalls, indirectly a ff ecting the interaction with agarose. Likewise, the per manent dipole moments of agarose 114 117 most likely originate from the ubiquitous placement of highly polarizable OH groups. Increasing the concentration of OH groups beyond the base material is unlikely. While there is inherent di ffi culty in directly measuring io n dipole interactions, these forces remain a strong candidate for adsorption due to SDS SWCNTs acting essentially as macroions and the presence of permanent dipoles in the agarose matrix. 114 117 It is critical to n ote that the charged head groups of SDS used to stabilize the nanotubes must play a crucial role in the separation of SWCNT suspensions. The results presented in Figure 3 3 show that screening the charges on SDS results in reduced SWCNT retention. Previous studies have also shown little to no separation when one attempts to separate SWCNT dispersed in other surfactants or with concentrations of SDS lower than 0.5 wt %. 66,67 Therefore, any force responsib le for
53 retention of SDS SWCNT during separation must account for the fact that the electric potential (charged surface) of SDS SWCNTs is essential for retention on neutral agarose. Interestingly, ion dipole interactions account for this observation. Furthermore, if ionic, hydrop hobic, or interactions were dominant in the selective adsorption of SWCNTs, increasing their density should have yielded higher retention of SWCNTs. Figure 3 7 shows that any modi fi cation to the agarose base signi fi cantly decreases the retention of s S WCNTs, especially in nonequilibrium systems. In fact, a strong inverse relationship is observed with ligand density regardless of the functional group. In nonequilibrium conditions, butyl and phenyl Sepharose HS, which both have a ligand density of 40 mo l/mL, retain a similar amount of SWCNTs despite the ligand groups being di ff erent. Changes to the high ly polarizable OH groups during functionalization will likely alter the overall dipole moment of the matrix, which would reduce retention if ion dipole in teractions were important. As the ligand density increases, more OH groups are altered on agarose, enhancing this e ff ect. The ligand density also a ff ects the selectivity of the matrix. The retention by matrices functionalized with ligand densities higher t han 5 mol/mL is not selective (see Figures A 4 A 6 in Appendix A ). The only functionalized matrix that shows a slight degree of selective retention is octyl Sepharose, which also has the lowest degree of substitution (5 mol/mL). Therefore, the presence of perman ent dipoles appears to be important to retention and selectivity. The results indicate that both the ionic charge on nanotubes and the permanent dipole on agarose gels are important to both retention and selectivity. Therefore, it is logical that ion dipol e interactions play a dominant role in the selective adsorption of
54 SWCNTs, as shown in Figure 3 8a. In some sense, this type of interaction with agarose is similar to those that take place between agarose and other solutes in what is called hydrophilic int eraction chromatography. 131 126.96.36.199 Role of SDS in separation s electivity While ion dipole interactions can account for the adsorption of SDS SWCNTs on agarose gel, questions still remain regarding the nature of selectivity, whereby s SWCNTs are initially retained by the gel and m SWCNTs are eluted. As ion dipole interactions carry no inherent selectivity on their own, the separation must be driven by inherent di ff erences between the s and m SWCNT species. We propose that th e origin of selective adsorption is due to di ff erences in polarizability of SWCNT species. Previous studies have suggested that a charge (i.e., SDS headgroups) in proximity to a polarizable object, such as a SWCNT, can in duce image charges on the SWCNT. 57,132,133 The induced image charges on the SWCNTs serve to screen the SDS headgroups from one another, as well as screening the SWCNTs from other approaching charges (i.e., permanent dipoles on agarose), thereby lowe ring their overall potential. Both theoretical calculations 134 and laboratory AFM measurements 135 have demonstrated large di ff erences in polarizability of m and s SWCNT types, whereby the polarizability of m SWCNTs was 3 orders of magnitude higher than their s SWCNT counterparts. The magnitude of the image charges formed is dependent upon the polarizability of the object; therefore, image charges are more easily induced on m SWCNTs, allowing SDS molecules to pack more tightly around m SWCNTs. 57,124 As a result, the interaction strength between m SWCNTs and the dipoles on agarose are lower due to both the ion dipole repulsion provided by the image charges and the steric
55 repulsion provided by a higher aggregation number of surfactant, as described in our prior work. 109 The combination of these e f f ects produces a much lower a ffi nity of m SWCNTs toward agarose than s SWCNTs, as shown in Figure 3 8b. It is also plausible that the inherent di ff erences in vdW forces for m and s SWCNTs calculated by Lifshitz theory 127,128 help drive the formation of unique sur factant structures surrounding each type of SWCNT (m or s ). The di ff erences in packing of surfactant on the SWCNT sidewalls would subsequently cause similar di ff erences in image charge. It is interesting to note that these e ff ects should exist for all na notubes suspended with ionic surfactants, meaning that agarose should be able to separate any SWCNTs suspended with anionic surfactants. However, selective adsorption is typically observed for only SDS SWCNTs, whereas a SC SWCNT suspension introduced into the column shows almost no retention. If the enthalpic e ff ects described above could solely describe the adsorption of nanotubes, one would expect similar results for SDS and SC SWCNTs. It is possible that the surfactants themselves exhibit di ff erent interactions with the agarose that could explain the di ff erences in adsorption for SDS and SC SWCNTs. However, Figure 3 9 shows that the SDS and SC molecules by themselves have almost no di ff erences in th eir interaction with agarose, indicating that any energetic di ff erence between the surfactant molecules is minimal. These results indicate that any di ff erences seen in SWCNT adsorption must be due to the SWCNT surfactant complex. Clearly there must be diff erences in SDS and SC SWCNT suspensions that promote their variant interactions with agarose. A possible explanation can be obtained by looking at the differences in the shell of both surfactants and its effect on the
56 hydrogen bonds and structure of water molecules. As we previously demonstrated, SDS molecules on SWCNTs are mobile and rearrange in response to chemical and mechanical stimuli. 109,125,126,136 Furthermore, simulation studies indicate that SDS molecules align with the axis of a SWCNT, exposing their hydrophobic tail considerably to the aqueous phase and leaving large areas of the nanotube surface bare. 42,110 In contrast, SC (or sodium deoxycholate) molecules are considered to be much more tightly bound to the surface and do not rearrange in response to chemica l or mech anical stimuli, providing better coverage of the SWCNT surface. Moreover, SC molecules bind to the SWCNT surface with their hydrophobic face in contact with the nanotube surface, while their hydrophilic face is exposed to water. Whenever water mol ecules accommodate nonpolar and hydrophobic molecules, the water molecules in their vicinity are more structured, due to the loss of degrees of freedom and consequently entropy. 133,137 Hence, water should be more structured in the vicinity of SDS SWCNTs than for SCSWCNTs due to the interaction of water with the hydrophobic SDS tails and bare nanotube surface. The adsorption of SDS SWCNTs to agarose will then lead to a net gain of entropy. Although the entropy decreases during adsorption from the r educed SDS degrees of freedom, that entropy loss can be compensated for and surpassed by a gain in entropy from the recovery of degrees of freedom of water. The process is analogous to micellization, where the formation of micelles results in a loss of deg rees of freedom for surfactant molecules inside the micelles but a net gain in entropy due to the recovery of degrees of freedom of water. By the same reasoning, these entropic effects can explain why higher concentrations of SDS can desorb SWCNTs 138,139 and why the solubilization of organic molecules on the surfactant shell reduces the
57 adsorption of SDS SWCNTs. 109 Both higher concen trations of SDS 42,110,124 and the solubilization of organic molecules 126,136 change the assembly of SDS molecules on SWCNTs in such a way that the hydrophobic tails of SDS mol ecules are hidden from the aqueous phase. In summary, adsorption between SDS SWCNTs and agarose occurs through an ion dipole interaction. While the selective structures formed around m and s SWCNTs will have implications on the enthalpic interactions, the entropic di ff erences also must have an important role in the selective retention of SDS SWCNTs. 188.8.131.52 Adsorption isotherm b ehavior Although others have assumed that the adsorption isotherms of SWCNTs on agarose gels are Langmuir type, the isotherms in Figures 5 and 6 cannot be described adequately by a Langmuir isotherm. It is also particularly important to note that the error bars depicted in these fi gures are very small for most data sets. The error bars increase only at the step edge or transition re gion before a stable plateau, where slight concentration di ff erences would yield large changes to retention. A Langmuir isotherm is represented by increasing adsorption until a plateau, or saturation point, is reached, which represents thermodynamic equili brium and complete occupation of the adsorption sites. This theoretical Langmuir shape is driven by the assumption that the absorbent contains a fi xed number of absorption sites of equal energy, resulting in monolayer coverage of the solid adsorbent, and t hat there is no interaction between the solutes. The non Langmuir shape of all isotherms suggest that adsorption sites may have di ff erent locations/conformations producing variou s energy barriers to adsorption. 140 As functional groups are added to the agarose backbone, the di ff erences between these
58 energy sites becomes more clear. These di ff erences could in dicate that regions with distinct magnitudes of permanent dipole moments account for these discrete energy sites. For the octyl Sepharose system presented in Figure 3 5 as an example, the multiple plateaus indicate either that SWCNTs deposit as multiple la yers on the surface of octyl Sepharose or that the number/energy of adsorption sites do not remain constant as the applied concentration of SWCNTs is increased. Interestingly, the adsorption isotherms for pure agarose (see Figures 3 5, 3 6 and Figure A 3 ) do not exhibit Langmuir behavior either, despite its persistent use in the literature. 122,141,142 Adsorption of SWCNTs onto agarose appears to follow isotherms that do not assume homogeneity in the energy of adsorp tion sites (e.g., Freundlich) or the formation of a monolayer on the surface of the gel (e.g., Brunauer Emmett Teller). 140,143 Again, agarose contains multiple ordered structural domains, each with di ff erent dipole moments. By their nature, SDS SWCNT suspensions are heterogeneous in l ength distribution (see Figure A 2 in Appendix A ). Accordingly, the number of adsorption sites (total energy of adsorption) should be proportional to nanotube length. The dynamic nature of the SDS SWCNT interface an d the permanent dipoles of di ff erent magnitude within agarose suggest that binding events of di ff erent energies are probable. Furthermore, the number of di ff erent packing con fi gurations of cylinders (SDS SWCNTs) available during adsorption and the inherent attractive interaction between SWCNTs make cooperative adsorption likely. Although assuming Langmuir behavior can provide some insight into the thermodynamics of the separation process, great care must be taken when attempting to extract speci fi c adsorpti on parameters by use of the relatively simple Langmuir model in complex systems. The complex nature of
59 the isotherms observed in this study indicates that more detailed calorimetric studies are needed to determine the thermodynamics of solute coverage, whi ch is beyond the scope of this study. 3.4 C losing Remarks The development of a simple, large scale process to separate SWCNTs is still needed, and consequently, the selective adsorption of SWCNTs onto agarose remains a promising method. Fully understandin g the mechanism of selective adsorption should lead to more e ff ective separations. In this study, we systematically altered the backbone of agarose to vary the relative importance of ionic, hydrophobic, and interactions during the adsorption between agarose and SWCNTs suspended with SDS. The results demonstrated that any alterations to agarose signi fi cantly reduced retention and selectivity. This inverse behavior and the inherent charge neutrality of agar ose indicate that the large permanent dipole moments exhibited by agarose are critical to the adsorption process. Combined with the importance of the electrical double layer on nanotubes, it is proposed that ion dipole interactions between the anionic char ges on SDS SWCNTs and the permanent dipoles of agarose are the dominant interaction in the adsorption process. The dissimilarities in polarizability of m and s SWCNTs result in di ff erent magnitudes of image charges on the nanotubes, thus altering the pack ing of surfactant on the sidewall. These di ff erent structures also limit the mobility of the surfactant. However, adsorption based solely on enthalpic e ff ects cannot account for the dissimilar behavior of SWCNTs suspended in SDS and other surfactants, such as SC. Therefore, selectivity is considered to be driven by both enthalpic and entropic e ff ects. Finally, the non Langmuir isotherms observed during equilibrium studies
60 indicates that great care must be taken when attempting to extract thermodynamic infor mation without the additional data provided by calorimetric studies.
61 Table 3 1. Characteristics of gel media used in this w ork Medium Typ e Group Ligand Density Pore Size % Agarose Sepharose 6 FF SEC 29 (c) 6 Sepharsoe 4 FF SEC 45 ( c) 4 sp Sepharose IEX ( ) sulfpropyl b 24 (d) 6 Q Sepharsoe IEX (+) quaternary ammonium b 29 (e) 6 phenyl Sepharose (HS) HIC phenyl 40 6 phenyl Sepharose (LS) HIC phenyl 25 35 (f) 6 butyl Sepharose HIC butyl 40 4 octyl Sepharose HIC octyl 5 4 Typical purpose of the gel medium: size exclusion (SEC), hydrophobi interaction (HIC), or ion exchange (IEX) chromatography. (b) Ligand density for charged groups depends on the eluent and is di ffi cult to characterize. (c) Hagel et al 113 (d) DePhillips and Lenho ff 144 (e) Yao and Lenho ff 145 (f) Evans et al. 146
62 Figure 3 1. P hysical and c hemical structure of agarose. A ) Monomeric unit of agarose chains. B ) Polymers organize into double helices and are further stabilized by bundling to form aggregates of various structure and size. Adapted image from Arnott et al. 14 7 C ) Ligands added to agarose backbone after functionalization. The R group represents CH 2 CH(OH)CH 2 chains added through a glycidyl ether coupling reaction. A B C
63 Figure 3 2. Separation of SWCNTs through selective adsorption. A ) Elution curve of SW CNTs suspended in 1 wt % SDS with Sepharose 6 FF as the stationary phase. The SWCNT suspension is injected at time 0. The elution curve is presented in terms of absorbance of effluent normalized by absorbance of initial suspension ) Absorban ce spectra from initial sample and effluent at the first (P1) and second (P2) peaks of the elution curve. Spectra of P1 and P2 have been slightly offset for visual clarity. B A
64 Figure 3 3 Comparison of retention behavior of 1 wt % SDS SWCNT suspension i n plain Sepharose 4 FF and 6 FF at different electrolyte concentrations.
65 Figure 3 4. Retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 6 FF and Sepharose 6 FF functionalized with ionic groups, sp and Q that contai n negative and positive charges, respectively. Bars indicate the mass fraction of SWCNTs eluted in peaks 1 (P1) and 2 (P2) as well as those that are irreversibly retained (not eluted with 2 wt % SC) within the column. All three columns were stabilized with 16 CV of 1 wt % SDS buffer prior to SWCNT injection.
66 Figure 3 5. Retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 4 FF and Sepharose 4 FF functionalized with octyl and butyl groups. A ) Comparison of the mass fraction of SWCNTs eluted in peaks 1 (P1) and 2 (P2) as well as those that are irreversibly retained within the column durin g nonequilibrium studies. B D ) Equilib rium adsorption isotherms for B ) plain Sepharose 4 FF ( ), C ) butyl Sepharose ( ), and D ) octyl Sepharose ( ). Note that the equilibrium concentration (C eq ) is given in milligrams per liter, while the amount of adsorbed SWCNTs (q) is given in micrograms per gram. A B C D
67 Figure 3 6. Retention behavior of 1 wt % SDS SWCNT suspension in plain Sepharose 6 FF and Sepharose 6 FF functionalized with phenyl groups at low and high substitution. A ) Comparison of the mass fra ction of SWCNTs eluted in peaks 1 (P1) and 2 (P2) as well as those that are irreversibly retained within the column during nonequilibrium studies. B D ) Equilibrium adsorption isotherms for B ) plain Sepharose 6 FF ( ) and phenyl Sepharose at C ) low ( ) and D ) high substitution ( ). A B C D
68 F igure 3 7 Relationship between ligand density and the mass of SDS SWCNTs retained by different gel media in nonequilibrium (column) studies
69 Figure 3 8. Mechanism of interaction and selectivity during agarose gel base d separations of SDS SWCNTs. A ) Dipoles on the agarose gel surface enable interaction with negatively charged head groups of SDS on SWCNTs. B ) Higher polarizability or differences in vdW forces alter the surfactant structure around m SWCNTs, thereby minimi zing the interaction potential between m SWCNTs and agarose gels. A B
70 Figure 3 9. Retention of surfactants used in agarose gel separations on Sepharose 6 FF. The figure shows the final ( ) SDS and ( ) SC surfactant concentration that remains in the supern atant after equilibrium. Final concentrations were determined by densitometry.
71 CHAPTER 4 EVALUATION OF CRITIC AL PARAMETERS IN THE SEPARATION OF SINGLE WALL CARBON NANOTUBES THR OUGH SELECTIVE ADSOR PTION ONTO HYDROGELS 4.1 Background Single w all c arbon n a notubes (SWCNTs) have attracted attention in a wide array of applications over the last twenty y ears for their incredible mechanical and electronic properties. 12 O ver th e last decade, significant gains have been reported in separation strategies to isolate as produced SWCNT suspensions into metallic (m ) and semiconduc ting (s ) fractions. 56,57,62,63,105 107,148 One of the most widely studied methods for the electronic sorting of SWCNTs has been their selective adsorption onto hydrogels, which was pioneered by Kataura and coworke rs. 65,66,141,142,149 Using this method, s odium d od e cyl s ulfate (SDS) is used to stabilize SWCNTs that are passed through a column containing dextran or agarose gel. After injection, m SWCNTs can be eluted with addi tional volumes of SDS solution, while s SW CN Ts are retained on the gel. The s SWCNTs are only released from the gel after the e luent is either increased to a high er SDS concentration or changed to a bile salt such as s odium c holate (SC) or s odium d eoxycho late (DOC). Our group was the first to describe the mechanism of this separation technique as the selective adsorption of s SWCNTs to the gel surface based on the structure of the surfactant coating surrounding the SWCNTs, 109 which has subsequently been confirmed by others. 69,71,150 Furthermor e, our group recently investigated the driving forces responsible for the selective retention of s SWCNTs on agarose and highlighted the importance both hydroxyl groups on agarose and SDS used in SWCNT dispersion during the separation process. 11 Continued advancement in this field has led to more refined separation processes that are not only able to sort SWCNT suspension s by
72 electronic type but further i solate the s SWCNT portion into single chirality fractions ( n, m ) using combinations of temperature modification 70 successive columns in series 69 or alterations to solution pH 71,150,151 While a great deal of progress ha s been made usin g this selective adsorption technique researchers have utilized a large variety of commercially available gels and suspension techniques i n a n effort to attain their specific goals. For example, it is well understood that hydrophobic SWCNTs must be stabil ized with surfactants for dispersion in an aqueous medium, 152 followed by ultrasonication and ultracentri fugation to remove metallic catalysts, amorphous carbon, and bundled SWCNTs. 48 However, there is no unified dispersion method in terms of ultrasonication time/power, or ultracentrifugation time/speed. As such, individual resear chers have choosen methods based on equipment restraints or what has been effective for their specific goals. Variability in the ultrasonication times, ultracentrifugation parameters, and mass to volume ratio of the initial SWCNT suspension, has a dramatic effect on the state of the suspension injected into a column. For example, some studies use low ultrasonication output powers (approximately 20 W) for as long at 20 h. 121,153 On the other hand, researchers have also repor ted ultrasonication times as short as 1 min at increased power output. 142 Other groups have found a combination of bath and tip ultrasonication to be effective in de bundling SDS SWCNT suspensions before separation. 69 These same variations are also seen in ultracentrifugation times, which have ranged from 15 min 121,153 to 4 h. 109 Interestingly, Flavel and coworkers completely eliminated the ultracentrifugation step before SDS SWCNT hydrogel separation. 151 These variations can create significant differences in SWCNT length and aggregation state before interaction with h ydrogel
73 adsorption sites, which is likely a key parameter in the separation process. 11 Our experience is that these parameters also have a significant im pact on the lifetime of specific components (e.g., column bed supports) and the reproducibility of the column. A final parameter of interest is the concentration of SWCNTs in the suspension injected into the column. It is very challenging to determine the true concentration of SWCNT dispersed in solution. 154 156 Accordingly, many researchers simply report the mass to volume ratio of the initial suspension, most commonly as 1 mg SWCNT per 1 mL of aqueous solution. 69,121,122,142,150,153 However, a significant portion of SWCNT aggregates, metallic impurities, and amorphous carbon are removed during ultracentrifugation. Therefore, the initial concentration is unrepresentative o f the final concentration injected into the column. Furthermore, the differences in suspension methods make it even more difficult to assess whether observed changes are due to either changes in concentration, dispersion methods, or elution parameters. Th e variability between all of these studies makes it extremely challenging to understand which parameter(s) have the most significant effect on selectivity, throughput, and purity of the resulting fractions. The goal of this work is to use insights from ou r previous column separations to comp are the quality, throughput, and reproducibility in the electronic sorting of SDS suspended SWCNT under a fixed set of conditions for different gels. U sing selected gels, we aim to understand how changes to the experime ntal system alter separation quality. To our knowledge this is the first study that systematically compares the separation process by altering the key parameters of gel medium and SWCNT concentration. The results of this study will be a
74 significant toward tuning the selective adsorption of SWCNTs to specific separation goals (e.g., purity vs. throughput) and developing the method into a large scale process 4.2 Methods 4.2.1 Materials Nanopure w all experiments. Both SDS and DO C ( >99%) were purchased from Sigma Aldrich and used as received. HiPco SWCNTs were obtained from Rice University (HPR 164.1) and used as received. Several hydrogels were used as stationary phase in t his study. These hydrogels include the Sephacryl series ( 100 HR, 200 HR, and 400 HR) and Sepharose series at both 4 and 6 % agarose at different degrees of cross linking (4B, 4B CL, 4FF, 6B and 6 FF) All g el s were manufactured by GE Healthcare. Sepharose 4 and 6 FF were purchased directly from GE, while all other media w ere purchased from Sigma Aldrich. It is important to note that the only major difference between each series of gel used in this study is the level of cross linking. 4.2 .2 Aqueous Suspensio n Preparation and Concentration A queous suspensions were prepared as previously described. 11,109 Briefly, 30 40 mg of raw SWCNT powder was added to 100 mL of a 1 wt % SDS s olution in N anopure water. This solution was mixed at 8 000 rpm (IKA T 25 Ultra Turrax) for 30 min. After homogenization, the solution was allowed to rest for 10 min before ultrasonication (120 W, Misonix S3000) to aid dispersion. The ultrasonication step was repeated a total of 3 times to ensure a well dispersed suspension. The solution was then ultracentrifuged for 4 h at 20 000 rpm (53 000g) to remove metallic catalysts, amorphous carbon, and SWCNT bundles from solution (Beckman Coulter Optima L 80K, SW 28 rotor). The
75 resulting suspension was adjusted to the desired final concentration by dilution with a 1 wt % SDS solution or concentrated using an Amicon 8200 ultrafiltration unit equipped with a regenerated cellulose membrane (30 kDa molecular weight cut off). This size range allows for free surfactant molecules to pass through the filter while surfactant stabilized SWCNTs remain in solution. Representative absorbanc e and fluorescence spectra for the initial and concentrated suspensions used in Figure S1. The spectra show that both suspensions have strong fluorescence and no broadening in the absorption spectra of SWCNTs indicating a high quality suspension of individual nanotubes 4.2.3 Equilibrium Adsorption Isotherms All equilibrium adsorption experim ents were completed similar to those previously described. 11 Each medium was washed thoroughly with Nanopure water to remove any ethanol used for preserv ation. The c leaned gels were then equilibrated with a 1 wt % SDS solution (SDS solution/gel volume ratio of 2:1) Approximately 500 L of the SDS stab i lized gels were placed in separate 2.5 mL centrifuge tubes. Each replicate was mixed with increasing conc entrations of SWCNT in a constant ba ckground concentration of 1 wt % SDS. After adding the SWCNT suspension the samples were mixed in a vortex stirrer for 30 s before being placed in a water bath held at 25 C for 24 h. After equilibration, samples were c entrifuged at 5000 rpm for 5 min to ensure that no gel remained in solution. An aliquot ( 300 L ) of the solution was then extracted and characterized by both absorbance and fluorescence spectroscopy as described below.
76 4.2.4 Column S eparations All column experiments were completed using low pressure chromatography. Glass columns were purchased from Bio Rad Laboratories (inner diameter of 1 cm). The columns were connected to a Bio Rad Econo gradient pump using a flow adaptor. Each column was packed with app roximately 10 mL of the desired medium. The column length varied between 9 11 cm, depending on the rigidity of the selected medium. After packing, columns were stabilized with at least 4 column volumes (CV), approximately 40 mL, of Nanopure water to remove any ethanol used as preservative. The medium was then equilibrated with 4 CV of 1 wt % SDS to prepare the column for separation. SWCNT suspensions were injected into the column at a volume of 2 mL (20 % CV). SWCNT suspensions were used at a concentration of 20 g/mL unless otherwise stated. After SWCNT injection, a portion of the injected sample was collected through elution with 1 CV of a 1 wt % SDS solution. The remaining SWCNT were collected after the introduction of 2 CV of a 1 wt % DOC solution. After collection of all eluted SWCNTs, the columns were re equilibrated with 3 CV of 1 wt % SDS solution in preparation for successive separations. Each column medium was used in three consecutive separations to evaluate reproducibility. The collected fractions were characterized by absorbance and fluorescence spectroscopy, as described below. 4.2.5 SWCNT Characterization The initial SWCNT suspensions and the supernatant extracted from the equilibrium studies were characterized by absorption (0.4 cm path) and f luorescence (1 cm path) spectroscopy on an Applied NanoFluorescence Nanospectralyzer (Houston, TX) with excitation from 662 and 784 nm diode lasers. Effluent from the column was
77 continuously characterized in situ by use of a flow cell from Starna Cells as previously described. 11,109 A bsorption spectra were taken every 20 s as the e ffluent flowed through the cell. Elution profiles of SWCNTs released from the column were esti mated by use of absorbance values at 626 nm to account for the presence of both the m and s SWCNT species. Mass fractions of the collected peaks were calculated using a standard curve plasma along the axis of the SWCNTs in solution and is independent of electronic type. 157 4.3 Results and Discussion As described in the introduction, published studies utilize different gels and contain great variabilit y in other important separation parameters (i.e., SWCNT and SDS concentration, injection volume, ultrasonication time, etc .) The in consistency in these parameters makes it difficult to compare and contrast results in the literature. In an attempt to elimi nate these factors from this comparative study, we have used a well defined suspension preparation strategy that minimizes processing time, resulting in stable suspensions dominated by mono dispersed SDS SWCNTs. Importantly, we have also standardized the i nitial SWCNT concentration throughout all column studies, which were adjusted to 20 mg/L with 1 wt % SDS after ultracentrifugation using an extinction coefficient of 0.043. 155 Normalizing the concentration is a critical step in any attempt to compare the efficiencies of separation using different gels since the isother ms were shown to be non Langmuir in our previous work. 11 Table 1 summarizes all of the SWCNT suspension and elution/separation parameters used in this st udy.
78 4.3.1 Hydrogel Characteristics used in SWCNT Separations Dextran (Sephacryl) and agarose (Sepharose) hydrogels are an attractive material because of the ability to manufacture gels with different pore sizes and rigidity (see Table 2). While SEC is the mechanism of separation for many biochemical applicatio length (> 100 nm) confirms that SEC cannot explain the separation of SDS SWCNTs in hydrogels, as described previously. 11,109 Therefore, direct interaction between the hydrogel surface and the SDS SWCNT complex must be the driver of separation. 11 B oth d extran and agarose have similar monomeric units as shown in Figure 1a. It is well understood that single strands of agarose form double helices stabilized by hydrogen binding, resulting in a relatively rigid structure. 112 Dextran based polymers do not undergo the same intermolecular stabilization that form double helices and, therefore, r equire the addition of cross linkers, such as the widely used methylene bisacrylamide (MBAm). Cross linkers are typically added to both gels (Figure 1b), which improves the rigidity, allowing for increased flow rates and sample throughput. The major differ ence between the gels is evident in the macro structure that forms. While both gels have a coiled structure, agarose has many highly oriented regions that may affect its interaction with other compounds (Figure 1c). As both dextran and agarose hydrogels ha ve been used in the electronic separation of SDS SWCNTs, it is reasonable to assume that the similar properties of the polysaccharides play a major role in the separation process. Both agarose and dextran contain a high concentration of hydroxyl groups tha t are ubiquitous throughout the structure (see Figure 1a). In fact, our recent mechanistic study concluded that these hydroxyl groups play a major role in the selective retention of s SWCNTs. 11 While no
79 mechanistic study linking the importance of hydroxyl groups on dextran gel to SDS SWCNT separation has been completed, it is likely that the hydroxyl groups are equally important due to the structural simil arities with agarose. 4.3.2 Typical Elution Profile of SWCNTs during Selective Adsorption onto Hydrogels The use of both dextran and agarose based gels to separate SDS SWCNTs has become one of the most promising, industrially scalable strategies for sepa rating nanotubes. 67,121,122 A typical elution profile during separation with Sephacryl 200 HR is presented in Figure 2a. After SWCNT injection, significant portions of the SWCNT are adsorbed onto the surface of the gel. The unabsorbed fraction passes through the column using 1 CV of a 1 wt % SDS solution. This initial fraction, herein labeled P1, is highly enriched in m SWCNT as demonstrated by the strong absorbance in the M 11 region and decreased absorbance of the s SWCNTs in both the S 11 and S 22 transitions, as shown in Figure 2b. Continued flow of 1 wt % SDS will not remove the adsorbed SWCNTs from the gel surface. The remaining SWCNTs are only eluted from the column after introduction of 2 CV of 1 wt % DOC. The S WCNTs collected in this fraction (P2) are highly enriched in s SWCNTs, as demonstrated by Figure 2b. Other studies have established that large increases in SDS concentration, 138,139 or other bile salts, including SC, 109 can also remove the adsorbed SWCNTs from the gel surface. 4. 3.3 SWCNT Separations for Various Dextran Gel s Of the commercially available gels used for the separation of SDS SWCNTs, Sephacryl 200 HR has been the most widely utilized. 69,71,142,150 However, other dextran gels including Sephacryl 100 HR, 300 HR, and 400 HR have been used. 68,121,153 Kataura and coworkers were the first to report the effectiveness of this medium to
80 separate SWCNTs by electronic type. 121,153 As shown in Table 2, the average particle size for each Sephacryl gel is 47 m, demonstrating that there should be no hydrodynamic effects altering SDS SWCNT transport through each gel. 11 The only major structural difference between the gels is the average pore size. Beca use the largest pore size for the dextran series is 31 nm, no significant SWCNT diffusion through hydrogel pores will occur 11 and it is reasonable to assume that all dextran gels should perform similarly under the same set of fixed conditions. The equilibrium adsorption results presented in Figure 3 support this conclusion. The slope of each hydrogel isotherm is different but the maximum adsorption for all gels is nearly identical (45 g/g). Interestingly, the initial slope of the isotherms has a distinct pattern, with the SWCNTs having the strongest affinity for the surface of Sephacryl 100 HR followed by 200 and 400 HR. It is also critical to note that the shape of these isotherms are not Langmuir, similar to our previous observations in agarose gels. 11 These isotherm shapes imply a more complex adsorption process than simple monolayer coverage, which should be expected once the heterogeneous nature of the initial SWCNT suspension is considered (i.e., lengths and diameters). Column separations using the dextr an series as an adsorbent demonstrate complimentary behavior to the isotherm data, as shown in Figure 4a. Similar amounts of SWCNTs are eluted in both P1 and P2 fractions for each Sephacryl gel tested. Although there appears to be a slight increase in the amount of SWCNTs eluted in P2 as the pore size is increased from 6.6 (100 HR) to 31 nm (400HR), these differences are not statistically significant.
81 Interestingly, similar behavior is not observed when each of the collected fractions using the dextran ser ies is evaluated for separation selectivity between the m and s SWCNT species. Figure 4 presents both the elution curves (Figure 4b d) and the normalized absorbance spectra of the collected fractions (Figure 4e f). The separation process for Sephacryl 100 and 200 HR are quite similar. In each case, the P1 fraction of nanotubes eluted under constant flow of 1 wt % SDS is highly enriched in m SWCNTs, as illustrated by the increase in M 11 transitions. This behavior is consistent with previously published stud ies using the same medium and expected considering the minimal differences in structure and pore size between 100 and 200 HR. Alternatively, the behavior of 400 HR under identical conditions is not consistent with either 100 or 200 HR. Under constant flow of 1 wt % SDS, the P1 fraction of SWCNTs eluted is enriched in m SWCNTs but also contains a significant portion of larger diameter s SWCNTs, as demonstrated by the significant features in the S 11 or S 22 optical transitions shown in Figure 4e. The change o f eluent to a 1 wt % DOC solution subsequently releases the initially adsorbed SWCNTs. SWCNTs collected in fraction P2 have a corresponding enrichment of the s SWCNT species, as shown in Figure 4f. In the case of 400 HR, the s SWCNT eluted in P2 have a sma ller mass fraction (Figure 4d) and show decreased spectral features in the corresponding abs spectra (Figure 4f), especially for the larger diameter semiconducting nanotubes. The loss of larger diameter s SWCNT to the P1 fraction is also observed in the fl uorescence spectra (see Figure S2) of both fractions when Sephacryl 400 HR is used as the adsorbent.
82 While our previous mechanistic studies focused on agarose based gels, 11, 109 it is likely that the mechanism of separation is analogous in dextran based gels, considering the striking similarities between their structure (see Figure 1). Our prior work highlighted the importance of hydroxyl groups on the outside surface of the gel beads during separation. Considering that the dextran beads used in the separation are uniform in size, any changes to pore size will impact the surface area on the outside of the bead, which is where adsorption of SWCNTs must occur given the size disc repancy between the pore and nanotube length. If it assumed that the OH groups are uniformly dispersed on the outside of the beads, then larger pores will decrease the outer surface area and, hence, the number of active adsorption sites The fact that more cross linker and less dextran are used to make larger pores further reduces the number of active sites on the outer surface. 158 Interestingly, if MBAm cross linkers were the active adsorption sites, as proposed by Strano and coworkers, 69 Sephacryl 400 HR should show increases in SWCNT retention in both column and batch systems. However, the opposite behavior is observed with Sephacryl 400 HR retaining much less SWCNT than ot her members of the dextran series. This is compelling evidence that the MBAm cross linkers do not function as the active adsorption sites during the separation process. On the other hand, these results support the importance of the hydroxyl groups during s eparation since Sephacryl 400 HR has less alkyl dextran (adsorption sites) than both Sephacryl 100 and 200 HR. The reduction in active sites lowers the amount of SWCNTs retained during separation, leading to s SWCNT contamination in the metallic fraction P 1. This change in selectivity is likely not seen when shifting the gel medium from 100 HR to 200 HR because the small increase in average pore size (from 6.6 to 7.7 nm) and
83 subsequent change to the number of active adsorption sites is insufficient to alter SWCNT retention. 4.3.4 SWCNT Separations for Various Agarose Gels Although Sephacryl 200 HR has been one of the most widely used hydrogels in SDS SWCNT separations, several variations of agarose gels have also been successfully utilized. 69,122,141 It has been reported that agarose based gels retain more SDS SWCNTs than their dextran based counterparts in both batch and column studies. 11,109,142,150 The elution curves in F igure 5 confirm this behavior. Under identical elution conditions, the total area and intensity of fraction P1 is greatly reduced when the medium is changed from a dextran to an agarose based gel. Additionally, while the overall area of fraction P2 increa ses with agarose gels, the peaks also become much broader. The broad shape of these peaks may be indicative of the increased volumes of 1 wt % DOC needed to remove adsorbed nanotubes from the gel surface. This phenomenon is likely explained by the increase d adsorbtivity of SDS SWCNT for agarose over dextran gels, as previously described. 122,142 Much like dextran based gels, agarose based gels known as Sepharose are manufactured with a v ariety of specifications by altering the percentage of agarose in the matrix as well as the level of cross linking between polymers. For example, Sepharose 4B is a beaded matrix of minimally cross linked 4 % agarose. Increased cross linking of the agarose matrix improves the rigidity of the structure, allowing for higher flow rates and increased throughput in traditional chromatographic separations. The level of cross linking in agarose is given by Sepharose 4B < Sepharose 4B CL < Sepharose 4 FF. The cross linker has a limited effect on the pore size as it is primarily driven by the percentage of agarose in the beads shown in Table 2 (i.e., 4 %).
84 Interestingly, the 4 % Sepharose series show elution characteristics that depend on the amount of cross linking i n the matrix. Figure 6a is the average mass fractions of SWCNTs eluted in each gel after three separations. It is clear that there is a significant increase in the retention of injected SDS SWCNTs as the amount of cross linking is increased from 4B to 4B C L as evident by the behavior of fraction P1 and the elution curves in Figures 6 b c. However, minimal changes are observed in the mass fraction and elution curves upon further addition of the cross linking agent from 4B CL to 4 FF, as shown in Figure 6d. Fi gure 6e shows that s eparations based on Sepharose 4B do not yield selectivity as high as those seen in the Sephacryl 100 or 200 HR systems While there is enrichment of the m SWCNT species relative to the initial suspension, the strong features in the near infrared spectrum indicate that a wide range of s SWCNTs are eluted as well. Although some selectivity has been demonstrated using this minimally cross linked material in constant flow systems, 122,141,142 Sepharos e 4B does not appear to be fit for industrial scale flow through separations. Once the parent Sepharose 4B material is cross linked, significant changes to the selectivity are observed in the associated absorbance spectra in Figures 6 e g. Although fraction P1 is significantly reduced in terms of mass in the elution profile (see Fig. 6a d) the collected fraction is highly purified in m SWCNTs. The absorbance spectra in Figures 6f g show that the M 11 absorbance transitions are substantially higher than the i nitial suspension and show considerably higher peaks than observed with any of the dextran based gels. Purity assessment for these ultra high purity m SWCNT fractions is extremely challenging using traditional spectroscopy. 159 The dramatic decrease in absorbance in the S 22
85 transitions as well as the complete l oss of spectral features in the S 11 transitions indicate purity levels exceeding 98 % as previously proposed by Tulevski and coworkers. 159 While the SWCNTs collected in fraction P2 do show some enrichment in the s SWCNT types, a significant fraction of m SWCNT contamination is observed Therefore, the highly cros s linked agarose series appear to strongly adsorb the SWCNTs. Indeed, a large portion of the injected nanotubes are retained on to the gel under the initial elution with 1 wt % SDS. This strong adsorption yields highly pure fractions of m SWCNTs but at the expense of contaminating the s SWCNT fraction eluted with DOC. Increasing the amount of agarose to 6% yielded similar observations with respect to the elution profiles, selectivity, and the role that cross linking had on the results (see Figures S3 and S3 ). The 6% agarose series retains fewer SWCNTs than the 4% series, which is consistent with previously reported retention behavior. 11,122 As Figure 7 demonstrates, the Sepharo se 6 FF matrix yielded highly pure (> 98 %) m SWCNT fractions similar to those collected with Sepharose 4 FF. The P1 fraction is a highly pure m SWCNTs fraction that shows no s SWCNT contamination even in the fluorescence spectrum (not shown). However, a p ortion of the m SWCNTs are adsorbed to the gel and are only removed with DOC elution. As a result, the P2 fraction shows M 11 absorbance, indicating m SWCNT contamination within the s SWCNT fraction. The decrease in retention of SDS SWCNTs as the amount of agarose in the matrix increases may seem counterintuitive; however, the behavior is likely driven by changes to the structural domains of agarose at higher concentrations. 11 Interestingly, the
86 average pore size of both Sepharose 4 and 6 FF are similar to Sephacryl 400 HR. Despite the similar dimensions, the agarose matrices are highly selective while the Sephacryl 400 HR matrix is not. These characteristic s also highlight the importance of the major structural differences between the dextran and agarose polymers (See Figure 1). It appears that the intramolecular stabilization that results in double helix formation in agarose has a major impact on how the SD S SWCNTs interact with the hydrogel during separation even when the base units are strikingly similar. 4.3.5 Lifetime of Hydrogels from Repetitive Separations A remaining question in developing the large scale separation of SDS SWCNTs is the stability and lifetime of the hydrogels under repeated use. When shipped from the manufacturer, both agarose and dextran based gels have a clean white color. The intensity of this color is dependent mostly on the amount of dextran or agarose present in the gel, i.e., 4 vs 6 %. After interaction of SDS SWCNTs through the selective adsorption process, the gels begin to darken. Even after washing with high concentrations of SDS or DOC, a portion of the SWCNTs remain on the gel. It is likely that the source of this contami nation is a combination of amorphous carbon, bundled SWCNTs, or irreversibly retained s SWCNTs that clog the gel. 65,109 Regardless of the source, gel stability and lifetime remains a serious obstacle in the future implementation of this method at an industrial scale. In order to evaluate the durability of the gel under repeated use, three SWCNT sep arations were completed on each gel in series. The results of these successive separations are presented in Figure 8. While some decrease in retention is expected throughout the lifetime of the column, the results show that lowered retention occurs immedia tely after the first separation. Figure 8a shows how the elution profile of
87 Sephacryl 200 HR changes after three successive separations. Even after re stabilizing the gel with 3 CV of 1 wt % SDS, the overall retention of all SWCNTs injected into the column drops considerably. This decrease in affinity is not likely because of residual DOC on the gel since bile salts do not have higher affinity for hydrogels than SDS and copious washing of the gel between separations with both water and 1 wt % SDS will remov e any residue not covalently bonded to the matrix. 11 In general, the effect of repeated use shows much less significant decreases in retention for the ag arose based gels, as shown in Figure 8b. In the case of Sepharose 6 FF, there is only a marginal difference from the initial separation after successive separations using the same column. Although the difference is more pronounced for the SWCNTs eluted in fraction P1, this is not entirely unexpected as agarose has repeatedly shown a higher net adsorbtivity for SDS SWCNT than dextran. 122,142 Figure 8c shows the most dramatic change to th e elution profiles from consecutive separations when using the minimally cross linked Sepharose 4B as the adsorbent. After re stabilizing the gel with at least 3 CV of 1 wt % SDS, a subsequent separation using the same column and adsorbent causes a drastic loss in the ability of the gel to separate the injected SDS SWCNTs. The elution curve becomes irregular and shows large increases in retention time for fraction P2. We attribute these dramatic loses to the collapse of the gel structure under these pressur es due to the minimal amount of cross linking. While other studies have not reported such drastic changes when using either Sepharose 4B or 6B under low flow conditions, we anticipate that the internal pressure caused by our pumping system caused considera ble collapse of the gel.
88 4.3.6 Effect of Altering the SWCNT L oading Perhaps the most glaring inconsistency between published reports of the hydrogel based separations of SDS SWCNTs is the role that SWCNT concentration has on purity. While several studies have attempted to elucidate the effects of ultrasonication time, ultracentrifugation time, and loading volumes on the effectiveness of the hydrogels for electronic type separation, many of these studies neglect the fact that variations to the suspension me thod will have an effect on the overall concentration of SDS SWCNTs injected into the column. This inconsistency makes direct comparison between research labs difficult. To evaluate the effect of SWCNT concentration on the product purity, the initial SDS SWCNT suspension was concentrated. These concentrated suspensions were injected into columns packed with Sephacryl 200 HR using SDS SWCNT concentrations ranging from 20 to 70 g/mL. Figure 9a shows that there is little change in the mass eluted in both P1 and P2 fractions as the concentration of SDS SWCNTs is systematically increased. While there may be slight increases in the amount of SWCNTs collected in fraction P1 and corresponding decreases in the SWCNT collected in fraction P2, these changes are not s tatistically significant. Even as the total mass of SWCNTs eluted in each fraction remains constant, there are significant changes to the purity of each fraction as seen by the absorbance spectra presented in Figure 9b and 9c. While an enrichment of m SWCN Ts is observed at every loading concentration, there is clearly an increase in the amount of s SWCNT contamination eluted with the m SWCNT fraction in both absorbance (Figure 9b) and fluorescence (see Figure S5) as the initial loading concentration is incr eased. On the other hand, Figure 9c shows that
89 there is very little difference in the purity of the s SWCNTs collected in fraction P2. Using an approach similar to Hirano et al., 122 the purity of each fraction can be estimated from the peak ratios in the M 11 and S 11 regions, such as the raw absorbance values at A 510 and A 1120 Using an unpurified sample, a baseline ratio of approximately 1.1 can be computed, as shown by the dashed line in Figure 9d. An increase in this ratio is indicative of an enrichment of m SWCNTs in the analyzed fraction while a decreased ratio demonstrates enrichment of the s SWCNT species relative to the initial suspension. At the lowest loadings of SDS SWCNTs, the ratio for fraction P1 is over 2 times higher, indicating a substantial enrichment of m SWCNT species. However, the ratio steadily decreases as the loading concentration of SDS SWCNT is increased, indicating contami nation of s SWCNTs within the fraction. An analysis of the peak ratios in fraction P2 shows a consistent enrichment of s SWCNTs at any nanotube loading. 4.3.7 Selecting a Hydrogel for SWCNT Separations Evaluating the quality and efficiency of hydrogel base d separations of SWCNTs is a complex task. The competing effects of fraction purity, sample throughput, and gel stability must be evaluated in order to properly choose a hydrogel for separation. A summary of these findings for the hydrogels used in this st udy are presented in Table 3. The results of this study indicate that the choice of a suitable hydrogel for the separation of SWCNTs is dependent upon the proposed use of the separated material. For example, all highly cross linked agarose gels (4B CL, 4 F F and 6 FF) create highly pure fractions of m SWCNTs under the separation conditions used here. Additionally, these gels are highly stable as evident by their increased stability and column lifetime. However, the throughput of m SWCNT using these columns i s extremely low, requiring more starting material to be processed to obtain a substantial amount of separated
90 material. The failure of the Sepharose 4B after the first separation highlights the importance of cross linking in establishing a large scale pro cess. While the use of highly cross linked agarose gels obtains very pure m SWCNT fractions, the corresponding s SWCNT fraction contains m SWCNTs because of the high affinity of nanotubes to agarose (see Figures 6 and S4). On the other hand, both Sephacryl 100 and 200 HR create relatively pure fractions of both the m and s SWCNT species. Unfortunately, each of these gels is more susceptible to degradation over successive separations. As Sephacryl 100 and 200 HR are repeatedly used, the purity of the m SWCN T fraction degrades considerably as s SWCNTs cease to adsorb onto the dextran and elute in fraction P1. This behavior is likely due to a portion of the SDS SWCNT becoming irreversibly adsorbed to the column during elution, thereby decreasing the amount of adsorption sites available for subsequent separations. While the purity of the m SWCNT fraction degrades over continued use, the purity of the s SWCNT fraction remains relatively constant as m SWCNT are not retained in the column. Finally, it appears that minimally cross linked agarose gels (Sepharose 4B, 6B) and dextran based gels with larger pore size (Sephacryl 400 HR) are not selective in the separation of SDS SWCNTs, at least not under the elution conditions used here. 4.4 Closing Remarks The developm ent of a high quality and scalable method to separate SWCNTs is still needed. Although selective adsorption with hydrogels is a promising large scale method, great variation in hydrogel selection and SDS SWCNT suspension strategies have been extensively re ported in the literature. These inconsistencies make it difficult to determine whether changes in dispersion state or the elution protocol are responsible for the reported observations. In this study, we have compared the quality, throughput,
91 and stability of the separation process under a fixed set of suspension preparation and elution conditions. The dextran based gels Sephacryl 100 and 200 HR yield the best overall separation of m and s SWCNTs with high throughput, especially during its initial use. How ever, the purity of these separations rapidly decreases with continued gel use. The highly cross linked agarose based gels are more stable but also have some degradation in performance with continued use. Importantly, these gels result in higher purity m S WCNT fractions than the dextran based gels. The results support the conclusion that hydroxyl groups are an integral part of the adsorption of SWCNTs onto the hydrogel surface. In addition, the formation of double helices by agarose seems to strengthen this interaction. In all cases, a portion of the initially injected SDS SWCNT suspension is irreversibly retained on the gel. This results in a reduction of adsorption sites for the selective adsorption of s SWCNTs during subsequent separations, significantly hindering reproducibility. The initial concentration of SDS SWCNTs injected into the column is also a key parameter that affects the quality of the separation. Higher concentrations of injected SWCNTs degrade the purity of the metallic fraction as the acti ve sites for adsorption become overwhelmed. The degradation of the adsorption sites with increased concentration and repeated use raises concerns regarding gel stability and lifetime. In order to develop large scale separations, a simple method to clean th e hydrogels after each use is needed.
92 Table 4 1. Parameters used for all elution studies. Experimental Parameter Experimental Factor Value Suspension Parameters Initial SWCNT loading 0.3 0.4 mg/mL SDS concentration 1 wt % Homogenization time 30 min Homogenization speed 8 000 rpm Ultrasonication time 10 min (3X) Ultrasonication power 1 20 W Ultracentrifugation time 4 h Ultracentrifugation gravity 53 000 g Filtration membrane c ut off 30 kDa Final SWCNT concentration after d ilut ion 20 mg/ L Elution Parameters G el packing 10 mL Column ID 1 cm SWCNT loading 20 % CV 1 wt% SDS eluent volume 1 CV 1 wt% DOC eluent volume 2 CV
93 Table 4 2. Physical parameters of hydrogels used in this study. Pore sizes reported by H agel et al. 113 Base Unit Medium Average Particle Size (m) Average Pore Size (nm) Dextran Sephacryl 100 HR 47 6.6 Dextran Sephacryl 200 HR 47 7.7 Dextran Sephacryl 400 HR 47 31 Agarose Sepharose 4B 45 165 Agarose Se pharose 4B CL 45 165 42 Agarose Sepharose 4FF 45 165 45 Agarose Sepharose 6B 45 165 Agarose Sepharose 6FF 45 165 29
94 Figure 4 1. Physical and chemical structure of the hydrogels used in this study, including A ) monomeric units and B ) cross l inking units added to increase structural stability, and the resulting C ) macrostructure of both dextran and agarose gels. Notice that only agarose forms a double helix structure. A B C
95 Figure 4 2. Separation of 1 wt % SDS SWCNT suspension with a statio na ry phase of Sephacryl 200 HR. A ) Typical elution profile of SWCNT fractions collected during separation. The SWCNT suspension is injected at time zero. The elution curve plots the raw absorbance of e uent normalized by the absorbance of the initial suspe of both m and s SWCNT species. B 626 nm) for the initial suspension and effluent collected in both peaks 1 (P1) and 2 (P2). B A
96 Figure 4 3. Equilibrium adsorption isoth erms for 1 wt % SDS SWCNT suspension on dextran based hydrogels. Note that the equilibrium concentration (Ceq) is given in milligrams per liter while the amount of adsorbed SWCNTs is given in micrograms per gram
97 Figure 4 4. R etention behavior of 1 wt % SDS SWCNT suspension u sing dextran based hydrogels. A ) Comparison of the mass fraction of SWCNTs eluted in P1 and P2 fractions during separation. B D ) Elution curves of SWCNTs in columns with a stationary phase of Sephacryl 100, 20 0 and 400 HR, respectively. E F = 626 nm) for the initial suspension and effluent collected in both P1 and P2 fractions for each stationary phase in the dextran series. Notice the s SWCNT contamination in fraction P1 wh en Sephacryl 400 HR is used as the stationary phase. B C D A E F
98 Figure 4 5 Elution curves of 1 wt % SDS SWCNT suspensions with different hydrogels. Notice the decreased intensity and peak area of fraction P1 as the stationary phase changes from dextran to an ag arose based gel.
99 Figure 4 6. Retention behavior of 1 wt % SDS SWCNT suspension using 4 % agarose based gels A ) Comparison of the mass fraction of SWCNTs eluted in P1 and P2 fractions during separation. B D ) Elution curves of SWCNTs in columns with a stationary phase of Sepharose 4B, 4 B CL, and 4 FF, respectively. E F ) effluent collected in both P1 and P2 fractions for columns with a stationary phase of Sepharose 4B, 4B CL, and 4 FF, respectively. Note that the absorbance data collected in fraction P1 for the highly cross linked gels (4B CL and 4 FF) has been smoothed to minimize noise as a result of the low concentrations of SWCNT collected. A B C D E F G
100 Figure 4 7. effluent collected in both P1 and P2 fractions for columns with a stationary phase of Sepharose 6 FF. Note that the absorbance data collected in fraction P1 for the highly purified m SWCNTs has been smoothed to minimize noise as a result of the low concentration of SWCNT collected
101 Figure 4 8. Changes to retention behavior of 1 wt % SDS SWCNT suspension during three consecutive separations for different hydrogels. After collection of fraction P2, each column was re stabilized with 3 CV of 1 wt % SDS in preparation for the next separation. Elution curves of SWCNTs in colum ns with a stationary phase of A) Sephacryl 200 HR, B) Sepharose 6 FF, and C ) Sepharose 4B. Note that t he drastic changes in elution curves for Sepharose 4B is attributed to the collapse of the gel struc ture A B C
102 Figure 4 9. The effect of SWCNT loading concentration on the retention behavior of 1 wt % SDS SWCNT suspension using Sephacryl 20 0 HR as the stationary phase. A ) Comparison of the mass fraction of SWCNTs eluted in P1 and P2 the initial suspension a nd effluent collected in both B) P1 and C ) P2 fractions. D ) The effect of SWCNT loading concentration on the purity of the eluted P1 and P2 fractions. Purit y is qualitatively evaluated using a ratio of raw absorbance at 510 nm and 1120 nm, which has a value of approximately 1.1 for the initial suspension. A ratio above and below 1.1 indicates metallic and semiconducting rich SWCNT fractions, respectively. A B C D
103 Table 4 3 Comparison of hydrogel attributes in the sep aration of 1 wt% SDS SWCNTs through selective adsorption. The signs (+) and ( ) indicate benefits and limitations, respectively. Medium Metallic Fraction (P1) Semiconducting Fraction (P2) Gel Stability Adsorption Strength Purity Throughput Reproducibility Purity Throughput Reproducibility Sephacryl 100 HR + + ++ ++ + + + Sephacryl 200 HR + + ++ ++ + + + Sephacryl 400 HR + --+ -+ + Sepharose 4B -+ --+ --+ Sepharose 4B CL ++ + ++ + ++ +++ Sepharose 4FF ++ + ++ + ++ +++ Sepharose 6B -+ --+ --++ Sepharose 6FF ++ + ++ + ++ ++
104 CHAPTER 5 EFFECTS OF THE ELECT RONIC CHARACTER OF S INGLE WALL CARBON NANOTUBES ON THE BIO LOGICAL RESPONSE OF A FRESHWATER GR EEN ALGAE 5.1 Background Since their discovery, single wall carbon nanotubes (SWCNTs) have been extensively researched by both scientists and engineers due to their unique properties (e.g. strength, high adsorption capacity, controllable conductivity, et c.) and their potential to be used in for both industrial applications and consumer products. 12 However, as is the co ncern for all emerging technologies, increased production and subsequent inclusion of SWCNTs in consumer products would likely lead to increased environmental and human exposure. Upon introduction to waste streams, SWCNTs could reach natural systems and po tentially interact with ecological receptors and impact the biosphere. In fact, aquatic systems, such as rivers and lakes, behave as primary environmental sinks by integrating pollutants from atmospheric deposition, terrestrial surface runoffs, and groundw ater discharges, raising concerns as to potential effects of SWCNTs that might accumulate in these systems. 160,161 Over the last decade, significant research has been conducted to evaluate the potential toxicity of SWCNTs 80,162 167 However, published results on the toxic effects of SWCNTs have been rather controversial, in that, for the same test organism, some studies point to severe acute toxicity while others show little to no toxic effect. 168 170 In fact, such inconsistencies are common in the burgeoning field of nanotoxicology. With exposure to t he different types of engineered nanomaterials, different research groups are left to determine which traditional tests best fit their needs. Assessing the potential
105 toxicity of SWCNTs is further complicated when one considers the wide variety of surface modifications, particularly the use of surfactant molecules, to aid SWCNT dispersion in aqueous media. 33 Researchers have used a la rge number of amphiphilic molecules for SWCNT stabilization including traditional surfactants like sodium dodcyle sulfate, bile salts including sodium cholate, or nonionic surfactants and polymers such as plurionc acids. While SWCNTs may have negative biol ogical impacts on aquatic organisms themselves, several surfactants commonly used in SWCNT stabilization have also proven to be acutely toxic to aquatic organsisms. 161 Therefore, any attempt to elucidate the toxicity of SWCNT must first evaluate the potential for used surfactant to contribute tox SWCNTs (i.e, stabilizing agents, length, diameter, electronic type, etc.) adds another level of complexity which has implications for toxicity studies. A great deal of work has been com pleted attempting to understand the effects of SWCNT stabilizing agents, length, diameter, and levels of residual metal catalysts on the biological responses of test organisms. In contrast, very little research has been conducted on the role of SWCNT elec tronic type (i.e., metallic (m ) or semiconducting (s )) in the potential toxic response. To our knowledge, only one study, using E. Coli as test organism, directly investigated the toxicity of metallic and semiconducting SWCNTs, following a separation of m SWCNT and s SWCNT species by density gradient ultracentrifugation. 81 However, the efficiency of the method used in the above study to produce large quantities of sufficient quality m SWCNT and s SWCNT fractions has not been well established. In fact, current research on large scale separat ion of m SWCNT and s SWCNT fractions relies on the selective adsorption of SWCNTs on
106 hydrogels surfaces during column separation. 65,66 Our research team has been worki ng to understand the mechanisms of SWCNT separation using hydrogel columns in order to scale up the process for industrial applications. 11,109 As a result, we have been ab le to produce SWCNT fractions of sufficient purity and concentration for use in toxicity studies. The goal of this study is therefore to investigate the potential adverse effects of as prepared SWCNTs mixtures, and well separated fractions of SWCNTs by t heir electric character on a model aquatic organism, the freshwater green algae, Pseudokirchneriella subcapitata. This is the first study that combines SWCNT separation by hydrogel selective adsorption and toxicity assessment. Using comparative laboratory studies, the biological response of P. subcapitata exposed to (i) as produced SWCNTs (mixture), (ii) m SWCNTs, and (iii) s SWCNTs was evaluated to determine differences in biological impact. 5.2 Methods 5.2.1 Materials Nanopure w all experiments. Surfactants (>99%), including sodium dodecyl sulfate (SDS) and sodium cholate (SC) were purchased from Sigma Aldrich and used as received. HiPco SWCNTs were obtained from Rice University (HPR 164.1) and used as received. Hydrogels used f or separation of initial SWCNT suspensions were Sepharose 6 FF, and Sephacryl 200 HR. Both gels are manufactured by GE, but Sephacryl 200 HR was purchased from Sigma Aldrich.
107 5.2.2 Aqueous Suspension Preparation Aqueous suspensions were prepared as de scribed previously. 11,88,109 Briefly, 30 to 40 mg of raw SWCNT powder were added to 100 mL of a 1 wt % desired surfactant solution prepared in Nanopure water. The suspension was mixed at 8000 rpm (IKA T 25 Ultra Turrax) for 30 min. After homogenization, the solution was allowed to rest for 10 min before ultrasonication (120 W, Misonix S3000) to improve the dispersion quality. The ultrasonication step was repeated 3 times to ensure a well dispersed suspension. N anotube bundles were then removed from the suspensions by ultracentrifugation (Beckman Coulter Optima L 80K, SW 28 rotor) for 4 h at 20000 rpm (53000g). The supernatants containing well suspended SWCNTs were characterized by absorbance and fluorescence sp ectroscopic methods as described below. 5.2.3 Surfactant Exchange on SWCNT Surfaces To produce large quantities of the different electronic fractions of SWCNTs using hydrogels, suspensions prepared in SDS were necessary. 66,67 However, our previous study has shown that SDS alone is highly toxic to several aquatic organisms including P. subcapitta the model organism to be used in this study. 161 Therefore, following the separation of the SWCNT fractions, an extra step was used to remove the tox ic SDS surfactant from SWCNT suspensions and replace it with sodium cholate (SC), which is less toxic to P. subcapit a ta Surfactant exchange and concentration was achieved using an Amicon 8200 ultrafiltration cell equipped with a regenerated cellulose mem brane (MW cut off 30 kDa). This system allows for SDS molecules to pass through the membrane under a pressure head of ultra high purity nitrogen gas. After an initial concentration step, the suspension was mixed with approximately 150 mL of a 1 wt % SC s olution and stirred for at least 20 minutes to allow for equilibration. Introduction of
108 a 1 wt % SC solution to the previously concentrated SWCNT suspension creates a concentration gradient resulting in an energetic driving force removing SDS from SWCNT s idewalls to equilibrate with the surrounding solution. Simultaneously, the SC in solution was driven to equilibrate with the SWCNT sidewalls. Several ultrafiltration passes eliminated SDS from the SWCNT sidewalls while keeping a well dispersed suspension in 1 wt % SC. These concentration and dilution procedures were then repeated at least three times having a serial dilution effect removing SDS from nanotubes sidewalls and replacing it with SC. After surfactant exchange, stabilized fractions were charac terized as described below. 5.2.4 SWCNT Separation into m SWCNTs and s SWCNTs All column experiments were completed using a method adapted from Kataura and coworkers, 6 5,66 and as described in our previous publications. 11,109 Glass columns purchased from Bio Rad with inner diameter of 2.5 cm were used in all studies. The columns were c onnected to a Bio Rad Econo gradient pump using a flow adaptor. Each column was packed with approximately 40 mL of the desired medium. Column length was approximately 8 cm in height based on the on the rigidity of the medium used in separation. After pac king, columns were stabilized with at least 4 column volumes (CV), approximately 160 mL, of Nanopure water to remove any ethanol used as preservative or potential containments not covalently bonded to the matrix. The medium was then equilibrated with 4 CV of 1 wt % SDS to prepare the column for separation. SWCNT suspensions were injected into the column at a volume of 8 mL (20 % CV). It is critical to note that the original suspension must be suspended in SDS for this separation process to be affective. 66,67 After suspension injection, 1 CV of 1 wt %
109 SDS was passed through the column at a constant flow rate of 1 mL/min. Using this eluent, a portion of the suspension highly enriched in m SWCNTs is elut ed from the column and collected. The remaining adsorbed nanotubes were eluted from the column by changing the eluent to a 2 wt % SC solution. This portion of collected nanotubes is highly enriched in the s SWCNT species. After each run, the column was r e stabilized with at least 3 CV of 1 wt % SDS in preparation for subsequent separations. All collected fractions containing either m SWCNTs or s SWCNTs were combined and concentrated using the Amicon ultrafiltration cell as described in the previous secti on to produce fractions at appropriate concentration and surfactant background for use in toxicity studies. These fractions were characterized with absorbance and fluorescence spectroscopy as described below. 5.2.5 SWCNT Characterization The initial SWC NT suspensions (i.e. mixtures of m SWCNTs and s SWCNTs) and separated fractions were characterized by absorbance, (0.4 cm path) and fluorescence (1 cm path) spectroscopy Absorbance and fluorescence measurements were completed on an Applied NanoFluorescenc e Nanospectralyzer (Houston, TX) with excitation from 662 and 784 nm diode lasers. Effluent from the column was continuously characterized in situ by use of a flow cell from Starna Cells. 11,109 A bsorption spectra were taken every 20 s as the effluent flowed through the cell. Elution profiles of SWCNTs released from the column were estimated by use of absorbance values at 626 nm to account for the presence of both the m SWC NT, and s SWNCT species Concentrations of initial suspensions, and concentrated fractions used in this study were characterized using absorbance measurements at 280 nm. This wavelength
110 plasma along the axis of the SWCNTs in solution, and is independent of electronic type. 157 5.2.6 Algal Growth in SWCNT Containing Culture Media Pure cultures of P. subcapitata were obtained from Hydrosphere Research (Alachua, FL) and cultivated under constant light exposure at room temperature. Algal toxicity tests w ere conducted as described by Youn et al., 88 and adapted from the U.S. Environmental Protection Agency ( US EPA) P. subcapitata 96 h growth inhibition method 1003.0. 171 In dose exposure studies, a given volume of a sterilized culture medium was transferred to separate 150 mL Erlenmeyer flasks that had been sterilized by autocl aving. 171 Individual flasks were spiked with increasing aliquot volumes of appropriate SC suspended SWCNT to produce a SWCNT concentration gradient. However, this approach results in increasing background SC concentration, therefore, adding a second variable in addition to changing SWCNT concentrations. Accordingly, SC concentrations in all culture media were maintained constant by adding necessary aliquot vol umes of the SC stock solution to each treatment prior to bringing the final volume to 50 mL using the culture medium solution. Following vigorous mixing, each flask was inoculated with 1 mL of pure algal suspension. All growth experiments were conducted in quadruplicate and under continuous illumination (~86 E m 2s 1). Changes in chlorophyll a (Chl a) concentrations were used as indicator of general biological responses and were monitored over time using a Turner Quantech Digital Filter Fluorometer. Ave rage algal growth rates determined from control samples were used as reference values to compare to the growth rates determined from the different SWCNT treated culture media.
111 5.3 Results and Discussion 5.3.1 Initial Screening of Surfactant Toxicity Any attempt to elucidate the potential for SWCNTs to cause toxicity to test organisms must first evaluate the potential of surfactant used in SWCNT dispersion to affect biological response. In order to determine the potential biological impact of surfactants used in this study to effect algal growth, an initial surfactant screening was completed. The results of a screening study designed to determine the concentrations of SDS and SC that do not impact the growth of P. subcapitata in SWCNT free growth media a re presented in Figure 1. When using SDS as dispersing agent for SWCNTs, a 1 wt % SDS solution corresponds roughly to a molar concentration of ~36 mM. Figure 5 1a shows that all tested SDS concentrations negatively impacted the growth of P. subcapitata For instance, a SDS concentration of 0.075 mM resulted in algal growth, that was 50% less than levels observed in control treatments (Figure 5 1a). Therefore, SDS should not be used as SWCNT stabilizing agent in suspensions prepared for the determination of intrinsic toxicity of SWCNTs. In contrast, the results of the concentration range finding study using SC as surfactant show that toxicity to P. subcapitata is expressed only for SC concentrations > 4 mM (Figure 5 1b). Therefore, all toxicity studies presented in this study using SC suspended SWCNTs have been completed in culture media with SC background concentration < 4 mM in order to assess the biological impact of SWCNTs and not that of the surfactant used in dispersion.
112 5.3.2 Biological Response o f Test Organism Exposed to as Prepared SWCNTs Following the determination of the adequate background concentrations of SC needed for toxicity studies, the effect of non separated SWCNTs (i.e. combined mixed m and s SWCNTs) on P. subcapitata growth was eva luated. The results are presented in Figure 5 2. The background SC concentration in these experiments was 1mM, which is less than 4mM, identified as threshold SC concentration (Figure 5 1b). It is clear from Figure 5 2 that algal growth decreased with inc reasing SWCNT concentration. Concentrations of SWCNTs at and above 0.25 mg/L show significant biological impact, with a dose of 0.75 mg/L inhibiting ~50% of algal growth. These results are consistent with previously published data using P. subcapitata a nd SWCNTs dispersed in either SC, 88 or other non toxic surfactants. 164 5.3.3 Separation of SDS SWCNTs to Produce m SWCNT and s SWCNT Fracti ons As described in the methods sections, the separation of SWCNTs using hydrogels must use SDS as a dispersant for proper application. 66,67 Accordingly, r epresentative spectra of SDS SWCNT suspension s used in this study show a well dispersed system characterized by a great deal of mono dispersed SWCNTs. (Figure C 1) A typical profile for the column separation of a SDS SWCNT suspension is presented in Figure 5 3. The elution profile shown in Figure 5 3a is obtained by in situ measurement of absorbance at 626 nm as SWCNTs are eluted from the column. This wavelength is chosen in order to account for the presence of both the m SWCNT and s SWCNT species as they absorb strongly in different spectral regio ns. The normalized absorbance spectra of collected SWCNT fractions using this method are presented in Figure 5 3b. The first fraction (P1) eluted under flow of 1 wt % SDS is highly enriched in
113 m SWCNTS relative to the initial suspension as evidenced by t he strong absorbance in the 400 620 nm range, and which can be attributed to the M 11 transitions of the m SWCNT species. The remaining SWCNTs are eluted from the column through introduction of a 2 wt % SC solution. The SWCNTs collected in this fraction, (P2) are highly enriched with the semi conducting s SWNCTs, as shown by the strong absorbance in both the S 11 (900 1350 nm) and S 22 optical transitions (650 850 nm). While selective absorption using hydrogels is a very powerful method of the separation of SWCNTs by electronic type, two major roadblocks exist for direct use of these collected fractions in dose response studies. The first is the use of toxic SDS surfactant in the initial suspension and column separation. It has been consistently demonstrat ed in the laboratory that the use of SDS in initial suspension preparation is critical to the success of the selective adsorption process. 66,67 Furthermore, our previous mechanistic work highlighted t he role of SDS on both the entropic and enthalpic drivers of the separation process. 11 Using the standard separation process, m SWCNT will elute coated with toxic SDS surfactant that must be removed b efore any dose response study can be completed. While the s SWCNT fractions are collected under flow of 2 wt % SC, there is potential for small amounts of SDS to be collected with the s SWCNT fractions. These small SDS concentrations have proven harmful to the test organisms used here, (See Figure 5 1), and must also be removed before use in dose response studies. A second roadblock for direct use of these separated fractions is the concentration of each collected solution. After initial sample injectio n, large volumes of both 1 wt % SDS and 2 wt % SC are needed to remove and collect the separated fractions from the packing material in the column. As such, both the m
114 SWCNTs and s SWCNTs are significantly diluted when compared to the initial suspension. In order to collect a significant portion of separated SWCNT fractions for use in bioassays, this separation process must be repeated frequently. All separated fractions of both m SWCNTs and s SWCNTs enriched fractions can be combined, resulting in large volumes of pure, but dilute solutions. To overcome each of these roadblocks, collected fractions of both m SWCNT and s SWCNTs were separately combined, concentrated, and surfactant exchanged to a 1 wt % SC through the use of the Amicon ultrafiltration ap proach as outlined in the methods section. 5.3.4 Validation of Surfactant Exchange Method In order to ensure that this concentration and surfactant exchange process effectively removes toxic SDS from collected fractions, and has minimal effect on the biological response of test organisms, a 1 wt % SDS SWCNT was prepared as outlined in the methods section. This suspension was surfactant exchanged to 1 wt % SC using the ultrafiltration method. Figure 5 4 shows the normalized fluorescence and absorbance spectra of the initial SDS SWCNT suspension and the resulting suspension after surfactant exchange. The fluorescence spectra in Figure 5 4a shows that after surfactant exchange, the suspension remains extremely stable with a great deal of individually su spended SWCNTs. The shifts in the peak positions can be explained by the changes to the microenvironment around SWCNTs. 126,172 As all the atoms on SWCNTs are on the surface, they function as sensors to any changes in their microenvironment. 125,136 The exchange of SDS for SC alters the microenvironment around SWCNT to shift the spectrum, which has been reported by others. 111,173 This same blue shift is evident in the absorbance spectra presented in Figure 5 4b. While these types of shifts have also been attributed to slight aggregation of the suspension,
115 the strong fluorescence intensities indica te well that these shifts are likely due to surfactant exchange. The obtained surfactant exchanged SWCNT suspension was then used in a 96 hr dose response studies with P. subcapitata to determine the potential effects on growth. The results were compared to those obtained from the baseline dose response study presented earlier (Figure 5 2). First, it is clear that there is no difference in biological response of P. subcapitata when grown in culture media containing SC suspended SWCNTs, or when grown in su spensions that have undergone surfactant exchange through serial dilution using ultrafiltration (Figure 5 5). In fact, the data indicate that surfactant exchange may even have a slight mitigation effect on SWCNT toxicity, although there is no statistical difference for the number of samples used here. This result confirms the efficiency of the Amicon 8200 ultrafiltration method to concentrate and surfactant exchange the SWCNT fractions after column separation. 5.3.5 Characterization of Separated m SWCNT a nd s SWCNT Fractions All fractions of m SWCNTs obtained from the separation process were combined resulting in a large volume ( > 300 mL ) of pure, but very dilute solution. This solution was concentrated and SDS exchanged to 1 wt % SC as described in the method section. An identical approach was used for the corresponding s SWCNT fractions. The resulting SWCNT fractions, now in 1 wt % SC were characterized by both fluorescence and absorbance spectroscopy and the results are presented in Figure 5 6. The raw fluorescence spectra (Figure 5 6a) clearly shows the presence of well stabilized, mono dispersed nanotubes in the s SWCNT fraction. Again, the peak shift can be explained by changes in the microenvironment around SWCNTs, as they are now coated with S C. 173 Additionally, the low intensity and lack of sharp peaks in the m
116 SWCNT fraction are evidence of little to no s SWCNT contamination. The absor bance spectra in Figure 5 6b further supports the purity of each collected fraction. Relative to the initial suspension, the m SWCNT fraction shows an increase in absorbance peaks in the M 11 transitions and significant signal decreases in both the S 11 and S 22 Alternatively, the processed s SWCNT fraction shows large signal increase in the S 11 and S 22 ranges, and a corresponding decrease in the M 11 range. These concentrated, high purity fractions were used in subsequent dose response studies. 5.3.6 B iological Responses of P. subcapitata to Type Separated SWCNT Fractions The results of the dose response assays using the separated SWCNTs are presented in Figure 5 7. At a concentration of 0.5 mg/L and below, m SWCNTs create no significant loss in algal viability when compared to control samples. In contrast, s SWCNTs resulted in significant negative impact of the algal growth regardless of the concentration tested. At the lowest tested concentration of 0.25 mg s SWCNTs/L, the algal growth was only ~46 % of what was observed in control culture media. Increasing the concentration of s SWCNTs to 0.75 mg/L resulted in total algal growth inhibition. These results suggest that the observed toxicity of the non separated SC suspended SWCNT (see Figure 5 2) is driven primarily by the s SWCNT species. In order to explain the potential driver of these observed differences in toxicity response, a clear understanding of the mechanism of column separation is needed. Our previously published work has focused on both the mechanism of separation in these column systems, as well as the source of selectivity. 11,109 It has been proposed that the significant differences in polarizability be tween the m SWCNT and s SWCNT species, 134,13 5 effects the nature and structure of surfactant coating on SWCNT
117 sidewalls. The increased polarizability of m SWCNTs allows from increased surfactant coverage along their sidewalls, relative to their s SWCNT counterparts. 57,124 This increased surfactant packing on m SWCNTs results in a both a net electrostatic and steric repulsion force from the hydrogel surfa ce. 11,109 Therefore, m SWCNTs are eluted under flow of 1 wt % SDS, while the s SWCNT species are retained. Importantly to the results presented here, the hy drogels used i n separation are organic molecule s consisting of long polymer like chains of sugars, i.e., carbohydrates which are chains of two or more sugars, and are an important structural feature of both cell membranes and cell walls Therefore, the increased toxici ty of the s SWCNTs, and lower toxicity of the m SWCNTs species may be explained by differences in the initial attraction of each SWCNT type to surfaces (i.e., cell membrane or cell wall) of the test organism used here. Additionally, these results suggest the need to understand the response of test organisms as driven by the nature of the surfactant SWCNT complex, and not simply the characteristics of the SWCNTs used in experimentation. In a previous study, Vecitis et al. investigated the viability of E. C oli when exposed to separated SWCNT fractions. They found that the biological response of E. coli was dependent upon the amount of m SWCNTs in the mixture. 81 Specifically, the authors reported that the m SWCNTs were significantly more toxic to E. Coli than their s SWCNT counterparts when us ing type separated tubes purchased from NanoIntegris, which has been separated through Density Gradient Ultracentrifugation (DGU). 81 The opposite behavior has been reported here. To ensure that inconsistency between these two studies was not driven by the choice of test organism used in bi oassays, type separated SWCNTs purchased from NanoIntegris were used in algal growth studies
118 using P. Subcapitata Using separated SWCNTs from NanoInegris, the results previous published by Vecitis et al. 81 are reproducible using P. Subcapitata as a test organism (See Figure C 3). This res ult is compelling evidence that the differences in biological response are not driven by different test organisms used during analysis. After eliminating test organism identity as a possible factor in the observed inconsistency between these two studie s, inherent differences between SWCNT materials used in each analysis must be considered. Importantly, there are critical differences between the SWCNTs used during toxicological evaluation. First, the SWCNTs used in this study were produced by the HiPco method, resulting in tubes with a very narrow diameter of roughly 1 nm. Vecitis et al. used nanotubes manufactured by Arc Discharge, a process which results in nanotubes of a slightly larger average diameter (ca.1.5 nm). 81 Finally, two different separation methods were used. Vecitis et a l. purchased separated SWCNTs (NanoIntegris) produced through DGU 81 and not through selective adsorption on hydrogels used in this study. DGU separations methods rely on differences in buoyancy of surfactant coated SWCNTs in co surfactant solutions of SDS and SC to separate the m SWCNT an d s SWCNT species. Additionally, the SWCNTs used by Vecitis et al. underwent secondary preparatory steps, including precipitating the purchased SWCNTs from solution, and re suspending them through bath sonication in appropriate solvents shown to be non t oxic to E Col i. 81 The results of bo th of these studies are compelling evidence that important differences exist between the biological responses of organisms when exposed to m and s SWCNTs. Further research is needed to fully understand how both SWCNT production methods and separation str ategies affect toxicity results. In fact, the affinity of one
119 SWCNT fraction for hydrogel surface and the lack of attraction from the other suggest that m SWCNTs and s SWCNTs may cycle differently in the environment, with potential different impacts on bi ota. 5.3.7 Mitigation of s SWCNT Toxicity As evident by mechanism of separation during the selective adsorption of as produced SWCNTs, the amount and structure of surfactant coating on SWCNT sidewalls can have a significant effect on how these particles i nteract with surfaces, including hydrogel organic polymers. 11,109 These differences in surfactant structure may also play a major role in observed toxicity of s SWCNTs as compared to m SWCNTs as previously discussed. Importantly, previously published studies have shown that the toxicity of SWCNTs may be mitigated through increasing the concentration of non toxic surfactant in culture media. 88 Increasing surfactant concentration likely results in more robust surfactant coverage along SWCNT sidewalls, minimizing their potential interaction with cell surfaces (i.e., cell membranes and cel l wall). To investigate the ability of increasing SC concentrations to mitigate the observed s SWCNT toxicity, P. Subcapitata cultures were exposed to identical concentrations of s SWCNTs (0.2 mg/L) in an increasing SC background. The results presented i n Figure 5 8 show significant increases in algal viability as the SC concentration is increased. This result is compelling evidence that that the nature and extent of surfactant coating along SWCNT sidewalls has a major impact on the potential for SWCNTs to cause negative biological impact once released to natural systems.
120 5.3 .8 Hydrogel Based Separation Technique as Model for Predicting the Affinity of NPs for Specific Bio molecules The toxicological evaluation of engineered nanoparticles (NPs), inc luding SWCNTs is a difficult undertaking. The incredibly small size of these materials raises new challenges for environmental scientists and toxicologists on how to adequately study their fate and transport in natural systems. Additionally, the use of st abilizing agents including amphiphillic surfactants and water soluble polymers adds complexity to the issue. For most NPs, it is likely that traditional toxicity assays developed for bulk chemicals may not be adequate for the determination of their impact s on organisms. Researchers have also used quantitative structure activity relationships (QSAR) to predict risks associated with the introduction of NPs in the environment. 174 176 While QSAR approaches are extreme ly beneficial, they require additional external experimental data to both validate and refine prediction abilities. The results presented in this paper offer an opportunity for use of bench scale laboratory studies to serve as predictive tools for NP toxic ity. Our results show that the m SWCNTs are far less toxic to P. subcapitata than their s SWCNT counterparts. The mechanism of separation of SWCNTs in column systems can be used to help develop approaches that predict the potential for toxicity of NPs ba sed on the composition of the column packing materials used as stationary phase, and the affinity of a given NP for such materials. During separation conducted in this study, m SWCNT are not attracted to the hydrogel surface due to their robust coverage wi th surfactant. 11,109 In contrast, s SWCNTs with a less robust surfactant coverage are strongly adsorbed on hydrogel surfaces during separation. 11,109 Interestingly, it is also this latter fraction which induces a significant adverse biological
121 impact on P. subcapitata (Figure 5 7). More importantly, the observed toxicity of s SWCNTs can be mitigated through increases in non toxic surfactant concentration resulting in more robust surfactant coverage along nanotubes sidewalls (Figure 5 8). These observations suggest that the toxicity of certain NPs could be predicted by simple interaction stu dies between bio molecules commonly found in cell membranes and NP of interest. With regard to SWCNTs, recent advancements to the selective adsorption method for SDS SWCNT separation have begun to produce not only electronically separated SWCNTs (m and s ), but also further speciation of the s SWCNT fraction into specific (n,m) types by using several columns in series 69 temperature adjustments, 70 or changes to solution pH. 71,150,151 Each of these methods alters separation properties to adjust the shape and structure of surfactant coverage along SWCNT sidewalls, and therefore stre ngth of interaction between the SWCNTs and the hydrogel surface. If the structure of surfactant coverage is a major driver of the observed toxicity of SWCNTs, there should be a corresponding gradient of biological responses. Specific (n,m) types that have the weakest interaction with the hydrogel surface (i.e. semi metals) should present limited biological impact, while those (n,m) types that interact the strongest with the hydrogel surface would cause harm to living organisms. Overall, the above observat ions do open a path to future research focusing on the predictions of the potential impacts of SWCNTs and other NPs on biota. 5.4 Closing Remarks The study of SWCNT toxicity is a complex task due to the heterogeneity of the initial powders in terms of l ength, diameter, metallic catalyst content, electronic configuration, and chemical stabilizing agents used in aqueous suspension preparation.
122 It is critical for researchers to develop systematic approaches to evaluate the potential for each of these factu res to drive biological response in traditional toxicity evaluations. In this study, we have demonstrated that the surfactant used in stabilization of SWCNTs is an important factor in evaluating their potential toxicity in environmental systems. In labor atory studies, the use of SDS as dispersing agent should be avoided as it enhances the toxicity of SWCNTs. Additionally, the developed separation process based on selective adsorption of SWCNTs onto hydrogels, allows for efficient fractionation of SDS susp ended SWCNTs into pure m SWCNT and s SWCNT fractions. Using our mechanistic understanding of this process, we have substantially increased the purity and throughput of the separation, creating fractions of sufficient purity and concentration for use in bi oassays. Our analysis demonstrates that the toxicity of as produced SWCNT suspensions is driven by the presence of the s SWCNT fraction. Our results suggest that methods used for production and separation of SWCNTs can have a significant impact on their interactions with organisms. Finally, this selective adsorption with hydrogels may have potential for development of a bench scale technique for NP toxicity screening by determining the affinity of NPs for specific bio molecules common in biological membr anes.
123 Figure 5 1. Effect of increasing concentrations of surfactants used in the aqueous dispersions of SWCNTs on the growth of P. subcapitata in a standard 96 hour chronic algal assay. A) Effects of increasing SDS concentration. B) Effects o f increasing SC concentration. Columns show changes in biomass measured as chlorophyll a Bars correspond to error bars based on 4 replicates. The dotted horizontal line represents the growth of P. subcapitata in the control culture media with no surfac tant added. (*) indicates significant difference from the control at 95% confidence. A B
124 Figure 5 2. Effect of increasing concentrations of non separated SWCNTs (i.e. mixture) on the growth of P. subcapitata in a standard 96 hour chronic algal a ssay as measured by chlorophyll a. The final concentration of sodium cholate (SC) in all treatments was adjusted to 1.0 mM. The dotted horizontal line represents the growth of P. subcapitata in the control culture media only SC added. (*) indicates signi ficant difference from the control at 95% confidence. Figure 5 3. Spectra of separated SDS suspended SWCNTs using hydrogel packed columns. The packing medium used in separation is Sepha cryl 200 HR. A ) Elution profile of separation process. Data collected at 626 nm to account for the presence of both m and s SWCNT species. B ) Normalized absorbance spectra of fractions collected in both P1 and P2. P1 shows clear enrichment of m SWCNT species, and loss of s SWCNT features relative to the initial suspension. Alternatively, P2 shows increases in s SWCNT features and loss of m SWCNT character. Spectra normalized at 626 nm to eliminate concentration effects. A B
125 Figure 5 4. Evaluation of the efficiency of Amicon Ultrafiltration method in surfact ant exchange from SDS to SC on SWCNT surfaces. A ) Fluorescence spectra of initial SDS SWCNT suspension and the same suspension after surfactant exchange of 1 wt % SC using an excitation wavelen gth of 662 nm. B ) Normalized absorbance spectra of initial S DS SWCNT suspension and the same suspension after surfactant exchange of 1 wt % SC. Spectra have been normalized at 626 nm to eliminate concentration induced effects. A B
126 Figure 5 5. Effect of suspension preparation on the growth of P. subcapitat a in a standard 96 hour chronic algal assay. A 1 wt % SDS SWCNT suspension was exchanged to a 1 wt % SC solution through ultrafiltration. Bars show algal growth of this suspension compared to a traditionally prepared 1 wt % SC SWCNT suspension. The fina l concentration of SC in all treatments was adjusted to 1.0 mM. The dotted horizontal line represents the growth of P. subcapitata in the culture media containing only SC and no SWCNTs. (*) i ndicates significant difference from the control at 95% confide nce.
127 Figure 5 6 Characterization of separated SWCNT fractions us ed in dose response studies. A ) Fluorescence spectra of initial SDS SWCNT suspension, and concentrated m and s SWNCT fractions in 1 wt % SC using an exci tation wavelength of 662 nm. B ) Normalized absorbance spectra of initial SDS SWCNT suspension and concentrated m and s SWNCT fractions in 1 wt % SC. Spectra have been normalized at 626 nm to eliminate concentration induced effects. A B
128 Figure 5 7. Effect of increasing concentrations of separated SWCNTs (m SWCNTs and s SWCNTs) on the growth of P. subcapitata in a standard 96 hour chronic algal assay. The final concentration of SC in all treatments was adjusted to 2 mM. Changes in biomass measured as chlorophyll a are shown as bars based on SWCNT type. Doses of 0.75 ppm s SWCNT resulted in complete growth inhibition. The dotted horizontal line represents the growth of P. subcapitata in the control culture media containing only SC and no SWCNTs. (*) indicates significa nt difference from the control at 95% confidence.
129 Figure 5 8. Effect of increasing concentrations of SC at a fixed concentration of s SWCNTs on the growth of P. subcapitata in a standard 96 hour chronic algal assay. The final concentr ation of s SWCNTs in all treatments was adjusted to 0.2 mg/L. Changes in biomass measured as chlorophyll a are shown as bars based on SWCNT type. The dotted horizontal line represents the growth of P. subcapitata in the control culture media containing o nly SC and no SWCNTs. (*) indicates significant difference from the control samples at 95% confidence.
130 CHAPTER 6 C ONCLUSIONS 6.1 Summary of Findings Research in nanoscience and the manufacturing of engineered nanoparticles (NPs) are grow ing exponentially, resulting in a wide variety of applications and novel products. Lessons learned from past ground breaking technologies show that such development brings not only societal benefits, but also unintended consequences on both the environme nt and organisms. Unfortunately, nanotechnology is not an exception. However, unlike most ground breaking and scientific revolutions, the growth of nanotechnology is accompanied by research on the potential adverse effects of nanotechnology products on bo th ecological functions and human health. One of the most widely studied NPs is the Single Wall Carbon Nanotube (SWCNT), and this research was driven primarily by the conflicting reports on the potential toxicity of SWCNTs. It was hypothesized that discrep ancies in results obtained from bioassays are likely due to the use of heterogeneous raw materials in terms of length, diameter, and electronic configuration of the SWCNTs. While the biological impacts of the length and diameter of SWCNTs have been resear ch extensively, the electronic characteristics of SWCNTs have been overlooked. Extensive literature search during this study showed only one peer reviewed article focusing on the electronic speciation of SWCNTs and the toxicity attributable to each of the identified SWCNT species. 81 Using a conceptual approach illustrated in Figure 6 1, this study elucidated the mechanisms of interaction between SWCNTs and a n organic polymer and how such mechanisms could be extrapolated to forecast the biological impacts of SWCNTs.
131 A selective adsorption me thod for SWCNT fractionation into metallic and semi conducting species was used. While the method has been studied quite extensively in the literature, very little is known on the physicochemical processes that drive the separation. With the goal of prod ucing highly concentrated and very pure fractions of both m SWCNTs and s SWCNTs needed for toxicity studies, the driving forces at play during the separation process were investigated. Looking at different forces that could be linked to the selective rete ntion of s SWCNTs on hydrogel surfaces, it was concluded that ion dipole interactions between the negatively charged SDS head groups and dipoles on hydrogel surfaces are the driving force during separation. It has also been proposed that the selectivity d uring hydrogel separation could be attributable to differences in the polarizability of the m SWCNT and s SWCNT species, with the higher polarizability of m SWCNTs, resulting in a more robust surfactant coverage along nanotube sidewalls and strong repulsiv e forces between the m SWCNT/surfactant complex and the hydrogel surface. Alternatively, the lower polarizability of the s SWCNT species does not result in surfactant coverage sufficient to result in a net repulsive force, leading to a selective retention during column separation. In an attempt to standardize, as well as maximize the quality and throughput of separated SWCNT fractions, a comparative study was completed examining how manipulation of important separation parameters affects the effectiveness of separation. Tested parameters included the type of hydrogel used during separation, column lifetime and the initial concentration of SDS suspended SWCNT to be used. The results of this study indicated that the choice of hydrogel used in separation is a n extremely important parameter. In general, dextran based hydrogels with small pore size provide
13 2 the best combination of fraction quality and throughput. However, dextran based gels with large pore volumes show no selectivity in separation (Sephacryl 40 0 HR). In contrast, agarose based gels produced very pure m SWCNT fractions, but with extremely low throughput. Therefore, the choice of hydrogel to be used in the separation process should depend on the intended use of separated material. Finally, it w as determined that the initial loading concentration of SDS SWCNTs affects the purity of the m SWCNT fraction. As the initial concentration of SDS SWCNTs injected in to the separation column increases, a large portion of s SWCNT begin to contaminate the m SWCNT fraction. Initial concentrations within the range of 20 to 30 ppm were found to be best for both SWCNT fraction purity and throughput. After optimizing the separation process, pure factions of m SWCNTs and s SWCNTs were produced, characterized an d tested in dose response studies using the freshwater green algae P. subcapitata as model test organism. It has been demonstrated that there is a significant difference in the biological response when P. subcapitata is exposed to identical concentrations of the two SWCNT species, with s SWCNT being significantly more toxic than m SWCNT. This observed difference in biological responses may be explained based on knowledge of the SWCNT separation mechanisms in hydrogel packed columns. During separation, m SWCNTs which are highly coated with the surfactant are poorly attracted for adsorption onto hydrogel surfaces. This increased surfactant coverage of the m SWCNT species may inhibit, or drastically reduce their attraction to organism membranes as evident b y their reduce impact to the test organisms used in this work. Alternatively, the s SWCNTs are highly attracted to the bio like surfaces used in separation, which may explain their toxic
133 impact on P. subcapitata Since the hydrogels used in this study ca n be seen as proxy for long chains of polymerized sugars similar to those present in cell membranes and cell walls (e.g. carbohydrates), these results suggest then that s SWCNTs do likely have a higher affinity for the sugar polymer in cell membranes, lead ing to toxicity. The actual mechanisms of toxicity (e.g. piercing, penetration in the lipid bilayer and oxidation of the hydrophobic tails of the phosphor lipid molecules, transfer in the cytoplasm, generation of reactive oxygen species, etc.) of s SWCNTs to P. subcapitata were not investigated. Overall, these findings point to the possibility of predicting the toxicity of SWCNTs through interaction with biomolecules which chemical composition mimics to some extent that of compounds found in plasma memb ranes. 6.2 Engineered Nanoparticle Toxicity Management The results presented in this work demonstrate the difficulty in truly understanding the potential threat that NPs may cause to living organisms when released to natural systems. The inherent charac teristics of the produced material, the use of surface modifications and chemical stabilizers, as well as the physiochemical transformations which likely to occur after introduction to the environment further, complicate the issue. However, the evaluation of NP toxicity in environmental systems should not be considered only as a challenge, but also as an opportunity for interdisciplinary research between materials scientists, environmental chemists, and engineers. Traditionally, each of these fields has b een viewed as separate disciplines, where materials scientist and engineers develop new products or novel applications with emerging technology, and environmental scientist and engineers focus on the end
134 of life concerns of these newly developed products. This compartmentalized approach is both inefficient and infective as it results in pollution consequences that require a retroactive action instead of a proactive careful design of initial products to minimize/eliminate their environmental impacts. Figu re 6 2 summarizes what a near ideal approach should be for a sustainable manufacturing and use of NPs. Nanotechnology appears to be here to stay, and the development of nanoparticle based products will continue to increase, thereby increasing the potentia l for both intentional (e.g. landfills) and non intentional (diffuse sources) environmental pollution as nano products move from cradle to grave. Well controlled laboratory studies do give good insights into potential NPs might have on the environment. H owever, after introduction to the environment, NPs are likely to undergo bio physicochemical transformations which should be taken into account in laboratory studies. The mechanistic understanding of how NPs affect living organisms under in situ condition s (e.g. experiments which takes into account the complexity of natural systems such as water chemistry) can lead to production/stabilization process that may eliminate/mitigate NP toxicity and inform the regulation process. Establishing feedback loops and connections as shown in Figure 6 2 might be critical for a safe development and implementation of nanotechnology. 6.3 Future Directions The work presented in this dissertation has laid the foundation for more studies on the biological impacts of SWCNTs, w hich can be extended to other NPs. Throughout this study, a single test organism was used in the evaluation of SWCNT toxicity. For a complete understanding of the effects of the separated SWCNT fractions on biota, additional test organisms must be evalua ted. It is possible that the test
135 organisms used here is especially sensitive to s SWCNTs, while a different organism may not exhibit the same sensitivities. Additionally, the effect of surfactant concentration on type separated SWCNTs toxicity may also be evaluated. Previous studies have indicated that toxic doses of carbon based NPs may be driven by the concentration of both SWCNTs and surfactant or polymers used in dispersion. 88 With regard to this study, it is possible that the toxicity of s SWCNTs could be drastically reduced or simply eliminated increasing the concentrations of used non toxic surfactant (i.e., SC). Another approach to toxicity mitigation of SWC NTs is the addition of natural dissolved organic matter (NDOM) isolates to prepared suspensions. The interaction of SWCNTs with NDOM is well documented in the literature, however little is known on the nature of the interaction mechanisms involved. The r esults presented here show that both polymers and ligands interact with SWCNTs differently based on their electronic type (m or s ). Therefore, it is plausible that NDOM isolates will interact with both m and s types differently creating alternate fate and transport paths in environmental systems. Finally, the ability of hydrogel interaction and separation as a predictive tool for NP cytotoxicity should be validated through controlled dose response studies using single chirality SWCNT suspensions acqui red through advancement to the separation process. It is likely that a gradient of biological impact will be observed with SWCNTs which interact the strongest with the hydrogel surfaces causing the most significant biological impact.
136 Figure 6 1. Expe rimental approach used in this study for the determination of the biological responses of a model test organism following the separation of an initial heterogeneous SDS SWCNT and production of pure m SWCNT and s SWCNT fractions. Mechanisms of s SWCNT sorpt ion on hydrogel surfaces used to imply the mode of interaction between algal cells and s SWCNTs, which resulted in toxic effects on P. subcapitata
137 Figure 6 2 Flow chart illustrating the research paradigm for future studies evaluating the toxicity of nanoparticles and assessing risks of engineering and using nanoparticles. A key feature to this approach is the feedback loop which allows toxicity results to info rm the manufacturing/design processes for development of lees toxic nanomaterials. Obtained results could then be used in support of the development of regulatory guidelines and safe nanotechnology applications
138 APPENDIX A SUPPORTING INFORMATI ON FOR CHAPTER 3 Elution Curve Cumulative Mass Calculation The challenge in estimating the cumulative mass of SWCNTs collected during experimentation lies in the difficulty of detecting both m and s SWCNTs using a single wavelength. Traditionally, m SWCN Ts absorb light strongly at shorter wavelengths (400 600 nm), while the s SWCNT species absorb light strongly at wavelengths longer that 600 nm. Therefore, choosing a detection wavelength at longer wavelengths will overestimate the contribution of s SWCNTS while detection at wavelengths in the shorter range will overestimate the contribution of m SWCNTs. Therefore, a detection wavelength of = 626 nm was used during experimentation as the best estimate of both m and s SWCNT contribution. The mass fracti ons are calculated assuming a constant flow rate of 1 mL/min and integrating the area under the curve using an extinction coefficient modified from a previously publish study by Moore et al. The extinction coefficient calculated by Moore et al (0.043) was extrapolated to an suspension of known concentration.
139 Figure A 1. Characterization of suspensions used in Chapter 3. A ) Absorbance spectr a of 1 wt% SWCNT suspensions used in this study. Suspensions used for equilibrium an d non equililbrium studies were ultracentrifuged for 1 and 4 h, respectively. The spectra have been normalized at = 763 nm for comparison. B ) NIR uoresence spectra of the same suspensions used during experimentation. Excitation was completed at = 662 nm. The spectra have been normalized at = 1123 nm for comparison. A B
140 Figure A 2 Lengths of SWCNTs used in Chapter 3. A) AFM image of the SDS SWCNT suspensio n. B) Histogram of th e nanotubes lengths showed an average of 467 nm. A B
141 Figure A 3. Equilibrium adsorption isotherms for SWCNTs in 1 wt% SDS with various agarose gels. The gel media are plain Sepharose 6 FF ( ) and 6B ( ), and Sepharose 4FF ( )
142 Fig ure A 4. Separation using functionalized Agarose Gels. A) Elution curves (chromatograms) of SWCNTs suspended in 1 wt% SDS using plain Sepharose 6 FF and Sepharose 6 FF functionalized with ionic groups as the stationary p hase. The SWCNT suspension is injected at time zero. The elution curves are presented in terms of the absorbance of the efflu ent normalized by the absorbance of the initial suspension. All absorbance data points are at = 626 nm. B) Absorbance spectra fr om the initial sample and the effluent at the fi rst (P1) and second (P2) peaks of the elution curve. Spectra of P1 and P2 have been s lightly off set for visual clarity. Notice the loss of separation selectivity after gel functionalization. A B
143 Figure A 5 Sepharose separation using functionalized gels. A ) Elution curves (chromatograms) of SWCNTs suspended in 1 wt% SDS using Sepharose 4 FF and Sepharose 4 FF functionalized with octyl and butyl groups as the stationary phase. The SWCNT suspens ion is injected at time zero. The elution curves are presented in terms of the absorbance of the effluent normalized by the absorbance of the initial suspension. All absorbance data points are at = 626 nm. B ) Absorbance spectra from the initial sample an d the effluent at the first (P1) and second (P2) peaks of the elution curve. Spectra of P1 and P2 have been slightly offset for visual clarity. While the octyl Sepharose matrix shows minimal signs of selectivity, butyl Sepharose shows a complete loss of se lective retention. A B
144 Figure A 6. Functionalized Phenyl Series A ) Elution curves (chromatograms) of SWCNTs suspended in 1 wt% SDS using Sepharose 6 FF and Sepharose 6 FF functionalized with phenyl groups at low (LS) and high (HS) sub stitution as the stationary phase. The SWCNT suspension is injected at time zero. The elution curves are presented in terms of the absorbance of the effluent normalized by the absorbance of the initial suspension. All absorbance data points are at = 626 n m. B ) Absorbance spectra from the initial sample and the effluent at the first (P1) and second (P2) peaks of the elution curve. Spectra of P1 and P2 have been slightly offset for visual clarity. Notice the loss of separation selectivity after gel functiona lization B A
145 APPENDIX B SUPPORTING INFORMATI ON FOR CHAPTER 4 Figure B 1 Characterization of SDS SWCNT suspensions used in this study. The c oncentrated suspension was obtained using a n Amicon 8200 Ultrafiltration unit equipped with a regenerat ed cellulose membrane (MW cut off 30kDa ). A ) Fluorescence spectrum ex = 6 6 2 nm ) And B ) Vis NIR absorbance spectrum of SDS SWCNT before and after concentration. The absorbance s pectra have been normalized at 763 nm to eliminate concentration induced chang es. F igure B 2 Representative fluorescence spectra ex = 6 6 2 nm ) of SWCNT fractions collected usi ng Sephacryl 400 HR as an adsorbent. s SWCNT s clear ly contaminate the P1 fraction, causing a corresponding decrease in their concentration in the P2 fractio n B A
146 6 % Agarose Series Separation profiles completed using the 6 % agarose series behave in a similar fashion to their 4 % counterparts in both mass fraction analysis and separation selectivity. Figure B 4 is a summary of separation process using the 6 % Sepharose agarose series. Again, as the parent matrix is supplemented with more cross linking material, a greater amount of SWCNTs are selectively absorbed to the gel surface as demonstrated by the behavior of P2 in Figure B 4 a. Column studies were only completed using the most highly cross linked material (Sepharose 6 FF) as previous studies using the 4 % agarose series showed no significant difference in mass retention. It is also important to note that while the behavior of the 6 % Sepharose series b ehaves similarly to the 4 % series, it initially retains less SDS SWCNTs than the 4 % series (See Figure 4 6 in main text ). This is consistent with previously reported behavior found in the literature whereby larger amounts of agarose decrease SDS SWCNT r etention. 11,122 A similar pattern of selectivity while increasing the amount of cross linker is also seen in the UV absorbance of the collected fractions shown in Figure B 4 d e. The base matrix of Sepharose 6B shows very little selectivity when used in these low pressure systems. While P1 shows slight indication of m SWCNT enrichment, there are abundant s SWCNTs features in the both the Visible and NIR regions of the spectru m. Contamination of s SWCNTs in P1 is also confirmed in the fluorescence spectra of the collected fractions. ( see Figure B 5) Alternatively, when using the Sepharose 6 FF matrix, P1 is a pure m SWCNTs fraction that shows no s SWCNT contamination on the flu orescence spectrum. However, a portion of the m SWCNTs are adsorbed to the gel after injection, and are removed from the gel surface
147 under flow of 1 wt % DOC along with the s SWCNTs. As a result, P2 shows significant absorbance in the 400 620 nm range indi cating that a portion of the m SWCNTs do not elute from the gel with the flow of 1 wt % SDS solution. Figure B 3 Retention behavior of 1 wt % SDS SWCNT suspension using 6 % agarose based hydrogels at diffe rent levels of cross linking. A ) Comparison of the mass fraction of SWCNTs eluted in P1 and P2 fractions during separatio n. B C ) Elution curves of SWCNTs in columns with a stationary phase of Sepharo se 6B and 6 FF, respectively. D E ) Normalized absorbance spectra tial suspension and effluent collected in both P1 and P2 fractions for columns with a stationary phase of Sepharose 6B and 6 FF, respectively. Note that the absorbance data collected in fraction P1 for the highly cross linked gels (6 FF) has been smoothed to minimize noise as a result of the low concentration of SWCNTs collected. A B C D E
148 Figure B 4 Representative fluorescence spectra ex = 6 6 2 nm ) of SWCNT fractions collect ed using Sepharose 6B as an adsorbent Fluorescence data clearly shows a lack of selec tivity in term s in separating the m from the s SWCNTs.
149 Figure B 5. Fluorescence spectra of SWCNT fractions collected in P1 using Sephacryl 200 HR as an adsorbent with increasing initial SDS SWCNT concentration. Spectra are collected with excita tion 6 2 nm and B ) 784 nm. There is a clear link between the loading concentration and s SWCNT contamination in fraction P1. A B
150 APPENDIX C SUPPORTING INFORMATI ON FOR CHAPTER 5 Figure C 1. Characterization of initial suspensions used i n separation and of SDS SWCNTs for us e in toxicological analysis. A ) Normalized fluorescence spectra of both suspensions using exci tation wavelength of 662 nm. B ) Absorbance spectra of both suspensions normalized at 763 nm to eliminate concentration effe cts. Figure C 2. Aliquot volumes of SWCNT suspensions used in this study. As produced SWCNT suspensions result in a dark black color, while isolated fractions of s SWCNTs and m SWCNTs appear blue/green and red, respectively. Photo courtesy of the aut hor. B A
151 Figure C 3. Effect of separated SWCNTs (m SWCNTs and s SWCNTs) on the growth of P. subcapitata in a standard 96 hour chronic algal assay using SWCNT purchased from NanoIntegris. The final SWCNT concentration in both treatments was 0.25 mg/L w hile final SC concentration was constant at 1mM. Changes in biomass measured as chlorophyll a. m SWCNT treatments resulted in completed algal inhibition. The dotted horizontal line represents the growth of P. subcapitata in the control culture media cont aining only SC and no SWCNTs. (*) indicates significant difference from the control at 95% confidence
152 LIST OF REFERENCES (1) Prez, S.; Farr, M. l.; Barcel, D. TrAC Trends in Analytical Chemistry 2009 28 820 832. (2) Kennett D. J.; Kennett, J. P.; West, A.; Mercer, C.; Hee, S. S. Q.; Bement, L.; Bunch, T. E.; Sellers, M.; Wolbach, W. S. Science 2009 323 94. (3) Verma, H. C.; Upadhyay, C.; Tripathi, A.; Tripathi, R. P.; Bhandari, N. Meteoritics & Planetary Science 2002 37 901 909. (4) Murr, L. E.; Soto, K. F.; Esquivel, E. V.; Bang, J. J.; Guerrero, P. A.; Lopez, D. A.; Ramirez, D. A. Jom 2004 56 28 31. (5) Woodrow Wilson International Center for Scholars and the Pew Charitable Funds: 2009. (6) Peralta Videa, J. R.; Zhao, L.; Lopez Moreno, M. L.; de la Rosa, G.; Hong, J.; Gardea Torresdey, J. L. Journal of Hazardous Materials In Press, Corrected Proof (7) Thayer, A. M. In Chemical and Engineering News 2010; Vol. 88, p 10 16. (8) Yao, D.; Chen, Z.; Zhao, K.; Yang, Q.; Zhang, W. Procedia Environmental Sciences 2013 18 149 156. (9) Scown, T. M.; van Aerle, R.; Tyler, C. R. Critical Reviews in Toxicology 40 653 670. (10) Mueller, N. C.; van der Bruggen, B.; Keuter, V.; Luis, P.; Melin, T.; Pronk, W.; Reisewitz, R .; Rickerby, D.; Rios, G. M.; Wennekes, W.; Nowack, B. Journal of Hazardous Materials 2012 211 212 275 280. (11) Clar, J. G.; Silvera Batista, C. A.; Youn, S.; Bonzongo, J. C. J.; Ziegler, K. J. J. Am. Chem. Soc. 2013 135 17758 17767. (12) De Volder, M. F. L.; Tawfick, S. H.; Baughman, R. H.; Hart, A. J. Science 2013 339 535 539. (13) Wilder, J. W. G.; Venema, L. C.; Rinzler, A. G.; Smalley, R. E.; Dekker, C. Nature 1998 391 59 62. (14) Wang, H.; Ren, F.; Liu, C.; Si, R.; Yu, D.; Pfefferle, L. D .; Haller, G. L.; Chen, Y. J. Catal. 2013 300 91 101. (15) Shi, D.; Resasco, D. E. Chem. Phys. Lett. 2011 511 356 362. (16) Hersam, M. C. Nat. Nanotechnol. 2008 3 387 394. (17) Liu, J.; Hersam, M. C. MRS Bull. 2010 35 315 321.
153 (18) Siddique R.; Mehta, A. Construction and Building Materials 2014 50 116 129. (19) Moaseri, E.; Karimi, M.; Maghrebi, M.; Baniadam, M. International Journal of Solids and Structures 2014 51 774 785. (20) Derycke, V.; et al. Comptes Rendus Physique 2009 10 3 30 347. (21) Merchan Merchan, W.; Saveliev, A. V.; Kennedy, L.; Jimenez, W. C. Progress in Energy and Combustion Science 36 696 727. (22) Liu, W. W.; Chai, S. P.; Mohamed, A. R.; Hashim, U. Journal of Industrial and Engineering Chemistry 2013 (23) An do, Y. Fullerene Science and Technology 1994 2 173 180. (24) Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Ohtsuka, Y.; Sen, R.; Suzuki, S.; Achiba, Y. Carbon 2000 38 1691 1697. (25) Bandow, S.; Asaka, S.; Saito, Y.; Rao, A. M.; Grigorian, L.; Richter, E. ; Eklund, P. C. Physical Review Letters 1998 80 3779. (26) Sen, R.; Ohtsuka, Y.; Ishigaki, T.; Kasuya, D.; Suzuki, S.; Kataura, H.; Achiba, Y. Chemical Physics Letters 2000 332 467 473. (27) Schauerman, C. M.; Alvarenga, J.; Landi, B. J.; Cress, C. D .; Raffaelle, R. P. Carbon 2009 47 2431 2435. (28) Carbon nanotubes : science and applications ; Meyyappan, M., Ed.; CRC Press: Boca Raton, FL, 2005. (29) Anna, M.; et al. Journal of Physics: Condensed Matter 2003 15 S3011. (30) Khler, A. R.; Som, C .; Helland, A.; Gottschalk, F. Journal of Cleaner Production 16 927 937. (31) Gilbert, D. N. Nanoscale 2 1306. (32) Height, M. J.; Howard, J. B.; Tester, J. W.; Vander Sande, J. B. Carbon 2004 42 2295 2307. (33) Vaisman, L.; Wagner, H. D.; Marom, G. Advances in Colloid and Interface Science 2006 128 130 37 46. (34) Chen, J.; Hamon, M. A.; Hu, H.; Chen, Y. S.; Rao, A. M.; Eklund, P. C.; Haddon, R. C. Science 1998 282 95 98.
154 (35) Mickelson, E. T.; Chiang, I. W.; Zimmerman, J. L.; Boul, P. J.; L ozano, J.; Liu, J.; Smalley, R. E.; Hauge, R. H.; Margrave, J. L. The Journal of Physical Chemistry B 1999 103 4318 4322. (36) Banerjee, S.; Hemraj Benny, T.; Wong, S. Advanced Materials 2005 17 17 29. (37) Garg, A.; Sinnott, S. B. Chemical Physics L etters 1998 295 273 278. (38) Bahr, J. L.; Yang, J.; Kosynkin, D. V.; Bronikowski, M. J.; Smalley, R. E.; Tour, J. M. Journal of the American Chemical Society 2001 123 6536 6542. (39) Shvartzman Cohen, R.; Levi Kalisman, Y.; Nativ Roth, E.; Yerushalm i Rozen, R. Langmuir 2004 20 6085 6088. (40) Shvartzman Cohen, R.; Nativ Roth, E.; Baskaran, E.; Levi Kalisman, Y.; Szleifer, I.; Yerushalmi Rozen, R. Journal of the American Chemical Society 2004 126 14850 14857. (41) Moore, V. C.; Strano, M. S.; Ha roz, E. H.; Hauge, R. H.; Smalley, R. E.; Schmidt, J.; Talmon, Y. Nano Letters 2003 3 1379 1382. (42) Tummala, N. R.; Striolo, A. ACS Nano 2009 3 595 602. (43) Islam, M. F.; Rojas, E.; Bergey, D. M.; Johnson, A. T.; Yodh, A. G. Nano Letters 2003 3 269 273. (44) Wang, R. K.; Chen, W. C.; Campos, D. K.; Ziegler, K. J. Journal of the American Chemical Society 2008 130 16330 16337. (45) Haggenmueller, R.; Rahatekar, S. S.; Fagan, J. A.; Chun, J.; Becker, M. L.; Naik, R. R.; Krauss, T.; Carlson, L.; Kadla, J. F.; Trulove, P. C.; Fox, D. F.; DeLong, H. C.; Fang, Z.; Kelley, S. O.; Gilman, J. W. Langmuir 2008 24 5070 5078. (46) Musumeci, A. W.; Waclawik, E. R.; Frost, R. L. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2008 71 140 142. (47) Jorio, A.; et al. New Journal of Physics 2003 5 139. (48) 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.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E. Science 2002 297 593 596. (49) Bachilo, S. M.; Strano, M. S.; Kittrell, C.; Hauge, R. H.; Smalley, R. E.; Weisman, R. B. Science 2002 298 2361 2366. (50) Fu, X.; Wang, J.; Ding, J.; Wu, H.; Dong, Y.; Fu, Y. Composites Science an d Technology 2013 87 170 173.
155 (51) Xi, Y.; Mallavarapu, M.; Naidu, R. Materials Research Bulletin 2010 45 1361 1367. (52) Khripin, C. Y.; Tu, X.; Heddleston, J. M.; Silvera Batista, C.; Hight Walker, A. R.; Fagan, J.; Zheng, M. Anal. Chem. 2013 85 1382 1388. (53) Chun, J.; Fagan, J. A.; Hobbie, E. K.; Bauer, B. J. Analytical Chemistry 2008 80 2514 2523. (54) Fagan, J. A.; Becker, M. L.; Chun, J.; Hobbie, E. K. Adv. Mater. 2008 20 1609 1613. (55) Khripin, C. Y.; Arnold Medabalimi, N.; Zheng, M ACS Nano 2011 5 8258 8266. (56) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C. Nature Nanotech. 2006 1 60 65. (57) Niyogi, S.; Densmore, C. G.; Doorn, S. K. J. Am. Chem. Soc. 2009 131 1144 1153. (58) Tanaka, T.; Jin, H. H.; Miyata, Y.; Kataura, H. Applied Physics Express 2008 1 (59) Mesgari, S.; Poon, Y. F.; Yan, L. Y.; Chen, Y.; Loo, L. S.; Thong, Y. X.; Chan Park, M. B. J. Phys. Chem. C 2012 116 10266 10273. (60) Lu, J.; et al. Small 2007 3 1566 1576. (61) Miya ta, Y.; Maniwa, Y.; Kataura, H. J. Phys. Chem. B 2006 110 25 29. (62) Tu, X.; Manohar, S.; Jagota, A.; Zheng, M. Nature 2009 460 250 253. (63) Nish, A.; Hwang, J. Y.; Doig, J.; Nicholas, R. J. Nat. Nanotechnol. 2007 2 640 646. (64) Chattopadhyay, D.; Galeska, I.; Papadimitrakopoulos, F. J. Am. Chem. Soc. 2003 125 3370 3375. (65) Liu, H.; Feng, Y.; Tanaka, T.; Urabe, Y.; Kataura, H. J. Phys. Chem. C 2010 114 9270 9276. (66) Tanaka, T.; Urabe, Y.; Nishide, D.; Kataura, H. Appl. Phys. Express 20 09 2 125002 1. (67) Tanaka, T.; Jin, H.; Miyata, Y.; Fujii, S.; Suga, H.; Naitoh, Y.; Minari, T.; Miyadera, T.; Tsukagoshi, K.; Kataura, H. Nano Lett. 2009 9 1497 1500.
156 (68) Gui, H.; Li, H.; Tan, F.; Jin, H.; Zhang, J.; Li, Q. Carbon 2012 50 332 33 5. (69) Tvrdy, K.; Jain, R. M.; Han, R.; Hilmer, A. J.; McNicholas, T. P.; Strano, M. S. ACS Nano 2013 7 1779 1789. (70) Liu, H.; Tanaka, T.; Urabe, Y.; Kataura, H. Nano Lett. 2013 13 1996 2003. (71) Flavel, B. S.; Kappes, M. M.; Krupke, R.; Hennric h, F. ACS Nano 2013 7 3557 3564. (72) Elsaesser, A.; Howard, C. V. Advanced Drug Delivery Reviews 2012 64 129 137. (73) Rzigalinski, B. A.; Strobl, J. S. Toxicology and Applied Pharmacology 2009 238 280 288. (74) Aschberger, K.; Johnston, H. J.; S tone, V.; Aitken, R. J.; Tran, C. L.; Hankin, S. M.; Peters, S. A. K.; Christensen, F. M. Regulatory Toxicology and Pharmacology 2010 58 455 473. (75) Hadrup, N.; Lam, H. R. Regulatory Toxicology and Pharmacology 2014 68 1 7. (76) Grass, R. N.; Limba ch, L. K.; Athanassiou, E. K.; Stark, W. J. Journal of Aerosol Science 2010 41 1123 1142. (77) Aruguete, D. M.; Hochella, M. F. Environmental Chemistry 2009 7 3 9. (78) Kang, S.; Pinault, M.; Pfefferle, L. D.; Elimelech, M. Langmuir 2007 23 8670 86 73. (79) Kang, S.; Herzberg, M.; Rodrigues, D. F.; Elimelech, M. Langmuir 2008 24 6409 6413. (80) Kang, S.; Mauter, M. S.; Elimelech, M. Environmental Science & Technology 2009 43 2648 2653. (81) Vecitis, C. D.; Zodrow, K. R.; Kang, S.; Elimelech, M ACS Nano 2010 4 5471 5479. (82) Chou, C. C.; Hsiao, H. Y.; Hong, Q. S.; Chen, C. H.; Peng, Y. W.; Chen, H. W.; Yang, P. C. Nano Letters 2008 8 437 445. (83) Kolosnjaj Tabi, J.; Hartman, K. B.; Boudjemaa, S.; Ananta, J. S.; Morgant, G.; Szwarc, H.; Wilson, L. J.; Moussa, F. ACS Nano 2010 4 1481 1492.
157 (84) Mutlu, G. k. M.; Budinger, G. R. S.; Green, A. A.; Urich, D.; Soberanes, S.; Chiarella, S. E.; Alheid, G. F.; McCrimmon, D. R.; Szleifer, I.; Hersam, M. C. Nano Letters 2010 10 1664 1670. (85) Alpatova, A. L.; Shan, W.; Babica, P.; Upham, B. L.; Rogensues, A. R.; Masten, S. J.; Drown, E.; Mohanty, A. K.; Alocilja, E. C.; Tarabara, V. V. Water Research 2010 44 505 520. (86) Bakota, E. L.; Aulisa, L.; Tsyboulski, D. A.; Weisman, R. B.; Hartger ink, J. D. Biomacromolecules 2009 10 2201 2206. (87) Zhang, Y.; Ali, S. F.; Dervishi, E.; Xu, Y.; Li, Z.; Casciano, D.; Biris, A. S. ACS Nano 2010 4 3181 3186. (88) Youn, S.; Wang, R.; Gao, J.; Hovespyan, A.; Ziegler, K. J.; Bonzongo, J. C. J.; Bitto n, G. Nanotoxicology 2012 6 161 172. (89) Bennett, S. W.; Adeleye, A.; Ji, Z.; Keller, A. A. Water Research 2013 47 4074 4085. (90) Roberts, A. P.; Mount, A. S.; Seda, B.; Souther, J.; Qiao, R.; Lin, S.; Ke, P. C.; Rao, A. M.; Klaine, S. J. Environme ntal Science & Technology 2007 41 3025 3029. (91) Kim, K. T.; Klaine, S. J.; Lin, S.; Ke, P. C.; Kim, S. D. Environmental Toxicology and Chemistry 29 122 126. (92) Yu, Z. G.; Wang, W. X. Water Research 2013 47 4179 4187. (93) da Rocha, A. M.; Ferr eira, J. R.; Barros, D. M.; Pereira, T. C. B. o.; Bogo, M. R.; Oliveira, S. r.; Geraldo, V.; Lacerda, R. G.; Ferlauto, A. S.; Ladeira, L. O.; Pinheiro, M. V. B.; Monserrat, J. M. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physi ology 2013 165 460 467. (94) Cheng, J.; Flahaut, E.; Cheng, S. H. Environmental Toxicology and Chemistry 2007 26 708 716. (95) Smith, C. J.; Shaw, B. J.; Handy, R. D. Aquatic Toxicology 2007 82 94 109. (96) Sargent, L.; Shvedova, A.; Hubbs, A.; Sa lisbury, J.; Benkovic, S.; Kashon, M.; Lowry, D.; Murray, A.; Kisin, E.; Friend, S.; McKinstry, K.; Battelli, L.; Reynolds, S. Environmental and Molecular Mutagenesis 2009 50 708 717. (97) Davoren, M.; Herzog, E.; Casey, A.; Cottineau, B.; Chambers, G.; Byrne, H. J.; Lyng, F. M. Toxicology in Vitro 2007 21 438 448. (98) Panessa Warren, B. J.; Maye, M. M.; Warren, J. B.; Crosson, K. M. Environmental Pollution 2009 157 1140 1151.
158 (99) Casey, A.; Herzog, E.; Lyng, F. M.; Byrne, H. J.; Chambers, G.; Da voren, M. Toxicology Letters 2008 179 78 84. (100) Pichardo, S.; et al. Toxicology in Vitro 2012 26 672 677. (101) Dong, L. F.; Witkowski, C. M.; Craig, M. M.; Greenwade, M. M.; Joseph, K. L. Nanoscale Research Letters 2009 4 1517 1523. (102) Murr ay, A. R.; Kisin, E.; Leonard, S. S.; Young, S. H.; Kommineni, C.; Kagan, V. E.; Castranova, V.; Shvedova, A. A. Toxicology 2009 257 161 171. (103) Dresselhaus, M. S., Dresselhaus, G., and P.C. Eklund Science of Fullerenes and Carbon Nanotubes ; Academi c Press: New York, 1996. (104) Komatsu, N.; Wang, F. Materials 2010 3 3818 3844. (105) Zheng, M.; Jagota, A.; Semke, E.; Diner, B.; Mclean, R.; Lustig, S.; Richardson, R.; Tassi, N. Nat. Mater. 2003 2 338 342. (106) Dalton, A. B.; Stephan, C.; Colem an, J. N.; McCarthy, B.; Ajayan, P. M.; Lefrant, S.; Bernier, P.; Blau, W. J.; Byrne, H. J. J. Phys. Chem. B 2000 104 10012 10016. (107) Khripin, C. Y.; Fagan, J. A.; Zheng, M. J. Am. Chem. Soc. 2013 135 6822 6825. (108) Ju, S. Y.; Utz, M.; Papadimit rakopoulos, F. J. Am. Chem. Soc. 2009 131 6775 6784. (109) Silvera Batista, C. A.; Scott, D. C.; McLeod, S. M.; Ziegler, K. J. J. Phys. Chem. C 2011 115 9361 9369. (110) Xu, Z.; Yang, X.; Yang, Z. Nano Lett. 2010 10 985 991. (111) Moore, V. C.; St rano, M. S.; Haroz, E. H.; Hauge, R. H.; Smalley, R. E.; Schmidt, J.; Talmon, Y. Nano Lett. 2003 3 1379 1382. (112) Tako, M.; Nakamura, S. Carbohydr. Res. 1988 180 277 284. (113) Hagel, L.; Ostberg, M.; Andersson, T. J. Chromatogr. A 1996 743 33 42. (114) Stellwagen, J.; Stellwagen, N. C. Biopolymers 1994 34 187 201. (115) Stellwagen, J.; Stellwagen, N. C. Biopolymers 1994 34 1259 1273. (116) Stellwagen, J.; Stellwagen, N. C. Nucleic Acids Res. 1989 17 1537 1548. (117) Dormoy, Y.; Candau S. Biopolymers 1991 31 109 117.
159 (118) Aqueous Size Exclusion Chromatography ; Elsevier: Amsterdam, 1988. (119) Scopes, R. K. Protein Purification: Priciples and Practice ; Springer Verlag: New York, 1994. (120) Stellan H. Jerten, J. R.; Pahlman, S. J. Chromatogr 1974 101 281 288. (121) Liu, H.; Nishide, D.; Tanaka, T.; Kataura, H. Nat. Commun. 2011 2 309. (122) Hirano, A.; Tanaka, T.; Kataura, H. J. Phys. Chem. C 2011 115 21723 21729. (123) Schaeffel, D.; et al. Physical Review E 2013 87 051 001. (124) Duque, J. G.; Densmore, C. G.; Doorn, S. K. J. Am. Chem. Soc. 2010 132 16165 16175. (125) Silvera Batista, C. A.; Weinberg, P.; Butler, J. E.; Ziegler, K. J. J. Am. Chem. Soc. 2009 131 12721 12728. (126) Wang, R. K.; Chen, W. C.; Campos, D. K.; Ziegler, K. J. J. Am. Chem. Soc. 2008 130 16330 16337. (127) Rajter, R. F.; Podgornik, R.; Parsegian, V. A.; French, R. H.; Ching, W. Y. Phys. Rev. B 2007 76 045417. (128) Hobbie, E. K.; Ihle, T.; Harris, J. M.; Semler, M. R. Phys. Rev. B 2012 85 245439. (129) Niyogi, S.; Boukhalfa, S.; Chikkannanavar, S. B.; McDonald, T. J.; Heben, M. J.; Doorn, S. K. J. Am. Chem. Soc. 2007 129 1898 1899. (130) Chen, R. J.; Zhang, Y.; Wang, D.; Dai, H. J. Am. Chem. Soc. 2001 123 3838 3839. (131) Tan, T.; Su, Z. G.; Gu, M.; Xu, J.; Janson, J. C. Biotech. J. 2010 5 505 510. (132) Lustig, S. R.; Jagota, A.; Khripin, C.; Zheng, M. J. Phys. Chem. B 2005 109 2559 2566. (133) Israelachvili, J. N. Intermolecular and Surface Forces ; Academic Press: San Di ego, CA, 1991. (134) Kozinsky, B.; Marzari, N. Phys. Rev. Lett. 2006 96 166801. (135) Lu, W.; Xiong, Y.; Hassanien, A.; Zhao, W.; Zheng, M.; Chen, L. Nano Lett. 2009 9 1668 1672.
160 (136) Silvera Batista, C. A.; Ziegler, K. J. Langmuir 2011 27 11372 11380. (137) Chandler, D. Nature 2002 417 491 491. (138) Inori, R.; Okada, T.; Arie, T.; Akita, S. Nanotechnology 2012 23 7. (139) Lukaszczuk, P.; Rummeli, M. H.; Knupfer, M.; Kalenczuk, R. J.; Borowiak Palen, E. Mat. Res. Bull. 2012 47 687 691. (140) Foo, K. Y.; Hameed, B. H. Chem. Eng. J. 2010 156 2 10. (141) Hirano, A.; Tanaka, T.; Urabe, Y.; Kataura, H. J. Phys. Chem. C. 2012 116 9816 9823. (142) Hirano, A.; Tanaka, T.; Kataura, H. ACS Nano 2012 6 10195 10205. (143) Giles, C. H.; Smit h, D. J. Colloid Interf. Sci. 1974 47 755 765. (144) DePhillips, P.; Lenhoff, A. M. Journal of Chromatography A 2000 883 39 54. (145) Yao, Y.; Lenhoff, A. M. Journal of Chromatography A 2006 1126 107 119. (146) Evans, D. R. H.; Macniven, R. P.; La banca, M.; Walker, J.; Notarnicola, S. M. Journal of Chromatography A 2008 1177 265 271. (147) Arnott, S.; Fulmer, A.; Scott, W. E.; Dea, I. C. M.; Moorhouse, R.; Rees, D. A. Journal of Molecular Biology 1974 90 269 284. (148) Fagan, J. A.; Khripin, C. Y.; Silvera Batista, C. A.; Simpson, J. R.; Hroz, E. H.; Hight Walker, A. R.; Zheng, M. Adv. Mater. 2014 (149) Tanaka, T.; Urabe, Y.; Nishide, D.; Kataura, H. J. Am. Chem. Soc. 2011 133 17610 17613. (150) Hirano, A.; Tanaka, T.; Urabe, Y.; Kataura H. ACS Nano 2013 7 10285 10295. (151) Flavel, B. S.; Moore, K. E.; Pfohl, M.; Kappes, M. M.; Hennrich, F. ACS Nano 2014 (152) Strano, M. S.; Moore, V. C.; Miller, M. K.; Allen, M. J.; Haroz, E. H.; Kittrell, C.; Hauge, R. H.; Smalley, R. E. J. Nanos ci. Nanotechnol. 2003 3 81 86. (153) Liu, H. P.; Tanaka, T.; Kataura, H. Phys. Status Solidi B. 2011 248 2524 2527. (154) Shtein, M.; Pri bar, I.; Regev, O. Analyst 2013 138 1490 1496.
161 (155) Moore, V. C., Rice University, 2005. (156) Jeong, S. H.; Kim, K. K.; Jeong, S. J.; An, K. H.; Lee, S. H.; Lee, Y. H. Synthetic Metals 2007 157 570 574. (157) Murakami, Y.; Einarsson, E.; Edamura, T.; Maruyama, S. Phys. Rev. Lett. 2005 94 087402. (158) Sephacryl Series Resin Chemical Pr operty Request 579286496 L252 According to the manufacturer, the pore size of the Sephacryl HR series is determined by mixing proper ratios of alkyl dextran and MBAm. In general, larger pore sizes are created by increasing the amount of MBAm cross linker and decreasing the amount of alkyl dextran during production. Therefore, Sephacryl 400 HR would have a larger amount of MBAm when compared to both Sephacryl 100 and 200 HR. (159) Tulevski, G. S.; Franklin, A. D.; Afzali, A. ACS Nano 2013 7 2971 2976. (160) Gao, J.; Youn, S.; Hovsepyan, A.; Llaneza, V. n. L.; Wang, Y.; Bitton, G.; Bonzongo, J. C. J. Environmental Science & Technology 2009 43 3322 3328. (161) Gao, J.; Llaneza, V.; Youn, S.; Silvera Batista, C. A.; Ziegler, K. J.; Bonzongo, J. C. J. En vironmental Toxicology and Chemistry 2011 31 210 214. (162) Fang, J.; Lyon, D. Y.; Wiesner, M. R.; Dong, J.; Alvarez Environmental Science & Technology 2007 41 2636 2642. (163) Liu, S.; Wei, L.; Hao, L.; Fang, N.; Chang, M. W.; Xu, R.; Yang, Y.; Chen Y. ACS Nano 2009 3 3891 3902. (164) C. Blaise, F. G. J. F. F. P. E. Environmental Toxicology 2008 23 591 598. (165) Casals, E.; Vzquez Campos, S.; Basts, N. G.; Puntes, V. TrAC Trends in Analytical Chemistry 2008 27 672 683. (166) Kim, K. T.; Klaine, S. J.; Lin, S.; Ke, P. C.; Kim, S. D. Environmental Toxicology and Chemistry 2010 29 122 126. (167) Boyle, D.; Fox, J. E.; Akerman, J. M.; Sloman, K. A.; Henry, T. B.; Handy, R. D. Aquatic Toxicology 2014 146 154 164. (168) Jia, G.; Wang, H.; Yan, L.; Wang, X.; Pei, R.; Yan, T.; Zhao, Y.; Guo, X. Environmental Science & Technology 2005 39 1378 1383. (169) Wick, P.; Manser, P.; Limbach, L. K.; Dettlaff Weglikowska, U.; Krumeich, F.; Roth, S.; Stark, W. J.; Bruinink, A. Toxicology Letters 200 7 168 121 131.
162 (170) Shvedova, A. A.; Kisin, E. R.; Mercer, R.; Murray, A. R.; Johnson, V. J.; Potapovich, A. I.; Tyurina, Y. Y.; Gorelik, O.; Arepalli, S.; Schwegler Berry, D.; Hubbs, A. F.; Antonini, J.; Evans, D. E.; Ku, B. K.; Ramsey, D.; Maynard, A .; Kagan, V. E.; Castranova, V.; Baron, P. American Journal of Physiology Lung Cellular and Molecular Physiology 2005 289 L698 L708. (171) U.S.EPA 2002 (172) Silvera Batista, C. A.; Wang, R. K.; Weinberg, P.; K.J., Z. Physical Chemistry Chemical Phy sics 2010 12 6990 6998. (173) Haggenmueller, R.; et al. Langmuir 2008 24 5070 5078. (174) Winkler, D. A.; Mombelli, E.; Pietroiusti, A.; Tran, L.; Worth, A.; Fadeel, B.; McCall, M. J. Toxicology 2013 313 15 23. (175) Gajewicz, A.; Rasulev, B.; Din adayalane, T. C.; Urbaszek, P.; Puzyn, T.; Leszczynska, D.; Leszczynski, J. Advanced Drug Delivery Reviews 2012 64 1663 1693. (176) Willighagen, E.; Karlsson, H.; Grafstrom, R.; Fadeel, B. Toxicology Letters 2012 211 S62 S63.
163 BIOGRAPHICAL SKETCH Justin G. Clar was born in Rochester, New York and was interested in environmental studies from a young age. He attended the University of Richmond serv ed with AmeriCorps St. Louis before returning to academia in 2008. He received his Masters Degree in 2010 from the University of Florida studying Arsenic M obilization durin g Aquifer Storage and Recovery Operations After becoming interested in the advanc ements of nanotechnology he stayed at the University of Florida to pursue his PhD under the guidance of Jean Claude J. Bonzongo and Kirk J. Ziegler
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ENBC3R90U_MB18Q9 INGEST_TIME 2014-10-03T22:15:36Z PACKAGE UFE0046498_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC