<%BANNER%>

The Application of Liquid-Liquid Interactions with Single-Walled Carbon Nanotubes

Permanent Link: http://ufdc.ufl.edu/UFE0024930/00001

Material Information

Title: The Application of Liquid-Liquid Interactions with Single-Walled Carbon Nanotubes
Physical Description: 1 online resource (110 p.)
Language: english
Creator: Wang, Randy
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: dispersion, emulsions, fluorescence, interfaces, nanotubes, spectroscopy, surfactants
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study covers three important research topics related to the application of liquid-liquid interaction with single-walled carbon nanotubes (SWNTs). The first topic describes the removal of SWNT bundles from liquid suspensions of nanotubes. The key to this work includes the use of liquid-liquid interfaces to trap the SWNT bundles due to the free energy change of the system during the process. SWNTs pack into crystalline ropes that form bundles due to strong van der Waals attraction. Bundling diminishes mechanical and electronic properties because it could interrupt the electronic structure of the nanotubes. Also, the electronic devices based on as-grown nanotubes with mixture of individual nanotubes and nanotube bundles make the electrical response unpredictable. We developed a new simple process to remove bundles by liquid-liquid interaction. SWNTs bundles are trapped at the interface because bundles stabilize the emulsions. Eliminating the use of ultracentrifugation to remov SWNT bundles enables large-scale production with reduced production costs and time savings. The second topic presented the swelling effect of the surfactant layer surrounding SWNTs with nonpolar solvents. Solvatochromic shifts in the absorbance and fluorescence spectra are observed when surfactant-stabilized aqueous SWNT suspensions are mixed with immiscible organic solvents. When aqueous surfactant-suspended SWNTs are mixed with certain solvents, the spectra closely match the peaks for SWNTs dispersed in only that solvent. These spectral changes suggest the hydrophobic region of the micelle surrounding SWNTs swells with the organic solvent when mixed. The solvatochromic shifts of the aqueous SWNT suspensions are reversible once the solvent evaporates. However, some surfactant-solvent systems show permanent changes to the fluorescence emission intensity after exposure to the organic solvent. The intensity of some large diameter SWNT (n, m) types increase by more than 175%. These differences are attributed to surfactant reorganization, which can improve nanotube coverage, resulting in decreased exposure to quenching mechanisms from the aqueous phase. The third topic describes the further study of the solvatochromism of the SWNTs. Since SWNTs are encapsulated with microenvironments of nonpolar solvents, it provides a new method to measure the photophysical properties of nanotubes in environments with known properties. Fluorescence and absorbance spectra of SWNTs show solvatochromic shifts in 16 nonpolar solvents, which are proportional to the solvent induction polarization. The photophysical properties of SWNTs were used to determine the relationship between the longitudinal polarizability and other nanotube properties.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Randy Wang.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Ziegler, Kirk.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024930:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024930/00001

Material Information

Title: The Application of Liquid-Liquid Interactions with Single-Walled Carbon Nanotubes
Physical Description: 1 online resource (110 p.)
Language: english
Creator: Wang, Randy
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: dispersion, emulsions, fluorescence, interfaces, nanotubes, spectroscopy, surfactants
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study covers three important research topics related to the application of liquid-liquid interaction with single-walled carbon nanotubes (SWNTs). The first topic describes the removal of SWNT bundles from liquid suspensions of nanotubes. The key to this work includes the use of liquid-liquid interfaces to trap the SWNT bundles due to the free energy change of the system during the process. SWNTs pack into crystalline ropes that form bundles due to strong van der Waals attraction. Bundling diminishes mechanical and electronic properties because it could interrupt the electronic structure of the nanotubes. Also, the electronic devices based on as-grown nanotubes with mixture of individual nanotubes and nanotube bundles make the electrical response unpredictable. We developed a new simple process to remove bundles by liquid-liquid interaction. SWNTs bundles are trapped at the interface because bundles stabilize the emulsions. Eliminating the use of ultracentrifugation to remov SWNT bundles enables large-scale production with reduced production costs and time savings. The second topic presented the swelling effect of the surfactant layer surrounding SWNTs with nonpolar solvents. Solvatochromic shifts in the absorbance and fluorescence spectra are observed when surfactant-stabilized aqueous SWNT suspensions are mixed with immiscible organic solvents. When aqueous surfactant-suspended SWNTs are mixed with certain solvents, the spectra closely match the peaks for SWNTs dispersed in only that solvent. These spectral changes suggest the hydrophobic region of the micelle surrounding SWNTs swells with the organic solvent when mixed. The solvatochromic shifts of the aqueous SWNT suspensions are reversible once the solvent evaporates. However, some surfactant-solvent systems show permanent changes to the fluorescence emission intensity after exposure to the organic solvent. The intensity of some large diameter SWNT (n, m) types increase by more than 175%. These differences are attributed to surfactant reorganization, which can improve nanotube coverage, resulting in decreased exposure to quenching mechanisms from the aqueous phase. The third topic describes the further study of the solvatochromism of the SWNTs. Since SWNTs are encapsulated with microenvironments of nonpolar solvents, it provides a new method to measure the photophysical properties of nanotubes in environments with known properties. Fluorescence and absorbance spectra of SWNTs show solvatochromic shifts in 16 nonpolar solvents, which are proportional to the solvent induction polarization. The photophysical properties of SWNTs were used to determine the relationship between the longitudinal polarizability and other nanotube properties.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Randy Wang.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Ziegler, Kirk.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024930:00001


This item has the following downloads:


Full Text

PAGE 1

APPLICATION OF LIQUID-LIQUID IN TERACTIONS WITH SINGLE-WALLED CARBON NANOTUBES By RANDY KAI-WEI WANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

2009 Randy Kai-Wei Wang 2

PAGE 3

To my family in Taiwan, who, always suppor t all my efforts especially for my PhD 3

PAGE 4

ACKNOWLEDGMENTS I sincerely express my gratitude to my advisor, Professor Kirk J. Ziegler, for his technical guidance throughout my graduate research. I will always be grateful for his patience and encouragement. I also thank my committee members, Professor Ranga Narayanan, Professor Peng Jiang, and Professor Ant Ural for their co ntribution to my propos al presentation and dissertation defense. I thank our group members, Justin Hill, Carlos Silvera, Fahd Rajab, Wei-Chiang Chen, and Cheng-Yin Lin for giving me lots of suggesti on and assistance to my research. I was very fortunate to be a member of this group. Finally, and most importantly, I express my tha nks and gratitude to my parents and friends in Taiwan, who have always supported me in sp irit along this journey. I also appreciate my dearest girl friend Jiun-Yi Tsai who has always been my great est support during my life in United States. My parents and friends gave me the courage to pursue my dreams, and Jiun-Yi encouraged me to be a successful researcher with an aggressive attitude in life and work. I am always going to be grateful for having them by my side. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................8 ABSTRACT ...................................................................................................................................13 CHAPTER 1 INTRODUCTION................................................................................................................. .15 1.1 The Electronic St ate of SWNTs.......................................................................................15 1.2 Characterization of SWNTs..............................................................................................16 1.3 The Surfactant Structure on SWNTs................................................................................17 1.4 Pickering Emulsions........................................................................................................ .18 1.5 Bundle Removal and Type Separation of SWNTs...........................................................19 1.6 Applications of SWNTs....................................................................................................20 2 INTERFACIAL TRAPPING IN REMOVING SWNT BUNDLES......................................25 2.1 Experimental Procedure....................................................................................................2 6 2.1.1 Dispersion................................................................................................................26 2.1.2 Interfacial Trapping.................................................................................................27 2.1.3 Characterization.......................................................................................................27 2.2 Results and Discussion..................................................................................................... 28 2.2.1 Selective Separation of Bundles..............................................................................28 2.2.2 Dispersion Quality...................................................................................................30 2.2.3 Improving Separation Effectiveness.......................................................................32 2.3 Mechanism........................................................................................................................34 3 INTERFACIAL TRAPPING OF SWNT BUNDLES SUSPENDED WITH ANIONIC SURFACTANTS.................................................................................................................... 46 3.1 Experimental Procedure....................................................................................................4 8 3.1.1 Dispersion and Interfacial Trapping........................................................................48 3.1.2 Characterization.......................................................................................................48 3.2 Results and Discussion..................................................................................................... 49 3.2.1 Selective Separation of Bundles..............................................................................49 3.2.2 Improving Separation Effectiveness.......................................................................50 5

PAGE 6

4 SWELLING THE MICELLE CORE SU RROUNDING SWNTS WITH WATERIMMISCIBLE ORGANIC SOLVENTS................................................................................57 4.1 Experimental Procedure....................................................................................................5 9 4.1.1 Dispersion................................................................................................................59 4.1.2 Characterization.......................................................................................................60 4.2 Results...............................................................................................................................60 4.3 Discussion.........................................................................................................................64 4.3.1 Swelling Effects......................................................................................................64 4.3.2 Possible Quenching Mechanism.............................................................................68 5 SOLVATOCHROMIC EFFECTS OF SWNTS IN NONPOLAR MICROENVIRONMENTS....................................................................................................78 5.1 Experimental Procedure....................................................................................................8 0 5.1.1 Dispersion................................................................................................................80 5.1.2 Characterization.......................................................................................................80 5.2 Results and Discussion..................................................................................................... 81 5.2.1 Solvatochromic Shifts.............................................................................................81 5.2.2 SWNT-Solvent Interaction......................................................................................83 6 CONCLUSIONS AND RECOMMENDATIONS.................................................................90 6.1 Removing SWNT Bundles via Interfacial Trapping........................................................90 6.2 Nonpolar Microenvironment Around SWNTs.................................................................91 6.3 Solvatochromism of SWNTs............................................................................................91 6.4 Recommendations.............................................................................................................92 APPENDIX A SUPPORTING INFORMATION FOR INTERFACIAL TRAPPING OF SINGLEWALLED CARBON NANOTUBE BUNDLES...................................................................93 A.1 Optical Linearity ......................................................................................................... ....93 A.2 Correction of Absorbance Spectra of SWNTs.................................................................93 B SUPPORTING INFORMATION FOR SOLVATOCHROMISM OF SINGLEWALLED CARBON NANOTUBES.....................................................................................96 B.1 Measured Solvatochromic Shifts.....................................................................................96 B.2 Deconvolution of Spectra.................................................................................................96 B.3 Nonlinear Optimization Model .......................................................................................97 LIST OF REFERENCES.............................................................................................................103 BIOGRAPHICAL SKETCH.......................................................................................................110 6

PAGE 7

LIST OF TABLES Table page 2-1 Dispersion Quality Comparison of Aqueous SWNT Suspensions....................................45 3-1 Dispersion Quality Comparison of Aqueous SWNT Suspensions....................................56 3-2 Comparison of the processing parameters for the pH-adjusted two-step interfacial trapping (pH 6.75) and ultracentrifugation methods for producing high quality SWNT suspensions............................................................................................................56 4-1 Spectral changes of surfactant-coated SWNTs mixed with organic solvents....................77 4-2 Fluorescence intensity changes for various SWNT ( n, m ) types after exposure to ODCB or chloroform.........................................................................................................77 5-1 Solvent parameters......................................................................................................... ....89 B-1 Criterion for assigning the we ighting factors for both inte nsity and amount of overlap with other peaks...............................................................................................................101 B-2 Combined weighting factors ( w = w Intensity w Overlap ) assigned for each ( n m ) type as a function of solvent.........................................................................................................101 B-3 Optimization results for all fitted parame ters when solved as individual systems and collectively.......................................................................................................................102 7

PAGE 8

LIST OF FIGURES Figure page 1-1 SWNTs can be viewed as a grapheme la yer wrapped into a tubular structure. The vector connecting two points on the plane can be described by the indices n and m ........22 1-2 Density of electronic states for a single nanotube. Color arrows indicate the optical transitions; black arrows represent the nonrad iative relaxation of the electron (in the conduction band) and hole (in the valence band)..............................................................22 1-3 Possible adsorption structure of SD S surfactants on a SWNT: A) SWNT encapsulated in a cylindrical structure. B) Hemimicellar structure of SDS surfactants on a SWNT. C) Random distribution of SDS surfactants on a SWNT.............................23 1-4 Toluene/water emulsions stabilized by surfactant-suspended SWNTs.............................23 1-5 The SWNT-based transistor. As ultrat hin wires, SWNTs could free up space in microchips for more devices, as well as solving heat and stability problems. At a little over a nanometer in diameter, this SWNT makes lines drawn by state-of-the-art photolithography look huge in comparison.......................................................................24 2-1 A) Absorbance and B) fluorescence (ex = 662 nm) spectra for the interfacial trapping process of Gum Arabic-suspended SW NTs using an initial SWNT mass concentration of 0.03 mg/mL. The control sp ectra (black lines) are the SWNTs after homogenization and ultrasonica tion. The inset shows the in terfacial trapping process in a separatory funnel. The control sa mple is then either subjected to ultracentrifugation (green lines ) or interfacial traps (red lines). The fluorescence from specific (n, m ) types are labeled.........................................................................................38 2-2 A) Absorbance and B) fluorescence (ex = 662 nm) spectra of SWNTs at a concentration of 0.005 mg/mL. The control sp ectra (black lines) are the SWNTs after homogenization and ultrasonication. The control sample is then subjected to interfacial traps.............................................................................................................. .....38 2-3 A) Normalized Raman spectra of the RBMs (ex = 785 nm) for Gum Arabicsuspended SWNTs after homogenization and ultrasonication (black line), interfacial trapping (red line), and ultracentrifugation (green line). The suspension was prepared from an initial concentration of 0.2 mg/m L raw SWNTs. B) Steady values of the shear viscosities for the Gum Arab ic-suspended SWNTs produced after homogenization and ultr asonication (control), in terfacial trapping, and ultracentrifugation. The suspensions were pr epared at similar concentrations to enable comparison. The final concentrati on of the control, interfacial trapping suspension, and ultracentrifuged samples were 0.02, 0.026, and 0.025 mg/mL, respectively........................................................................................................................39 8

PAGE 9

2-4 AFM images of Gum Arabic SWNT su spensions after A) homogenization and ultrasonication and B) after bundle removal by interfacial trapping. C) Histogram of nanotube diameters measured from the AFM images. The suspension was prepared from an initial concentra tion of 0.2 mg/mL raw SWNTs..................................................40 2-5 Fluorescence emission intensity fluctuations of the (7, 6) nanotube as a function of the volume ratio ( R ). The sample was prepared from an initial mass concentration of A) 0.03 or B) 0.005 mg/mL raw SWNTs. Excitation at 662 () and 784 nm ( ). Conductivity measurements suggest the formation of oil-in-wat er (o/w) emulsions (shaded area) for R < 1 and water-in-oil emulsions (w/o) for R > 1. The dashed lines represent the fluorescence spectra for the suspension after homogenization and ultrasonication....................................................................................................................41 2-6 A) Absorbance spectra of Gum Arabic-suspended SWNTs and B) fluorescence to absorbance ratio fluctuations as a function of the volume ratio (R ). The fluorescence intensity was measured for the (7, 6) nanot ube and divided by th e absorbance at the excitation wavelength (dashed line at 662 nm ). The suspension was prepared from an initial concentration of 0.03 mg/mL raw SWNTs.............................................................41 2-7 A) Fluorescence (ex = 662 nm), B) ab sorbance, and C) Raman RBM spectra (ex = 785 nm) of Gum Arabic-suspended SWNTs fo r one-step and two-step interfacial trapping separations compared to ultracen trifugation. The suspensi ons were prepared from an initial concentra tion of 0.2 mg/mL raw SWNTs..................................................42 2-8 Overall process of removing SWNT bundl es from aqueous suspensions via liquid liquid interfaces. Prior to in terfacial trapping, the raw SW NTs are first homogenized with a highshear mixer and ultrasonicated to coat SWNTs with surfactant (not shown), resulting in a mixture of i ndividually suspended SWNTs and SWNT bundles. A) Toluene is added to the aqueous phase and mixed to form emulsions, B) SWNT bundles are trapped at the emulsion interfaces, C) creaming and coalescence of emulsions separates the bundled SWNT s, and D) SWNT bundles are removed from the bulk aqueous fluid...............................................................................................43 2-9 Adsorption process of an individual or bundle of nanotubes at th e interface of the oil and water phases. A) Prior to interfacial trapping, individual or bundled nanotubes are dispersed in the aqueous phase. Afte r mixing, the nanotubes assemble at the interface. B) Diagram showing the end of a nanotube at the interface where R is the radius of an individual or bundled nanotube and is the contact angle measured into the water phase...................................................................................................................44 3-1 A) Radial breathing mode (RBM) of Raman spectra, B) Vis-NIR absorption spectra, C) NIR-fluorescence spectra (662 nm excitation), and (d) Fluorescence / absorbance ratio (calculated from 662 nm) of SC-suspe nded SWNTs. (1), (2), (3), (4) and (5) represent the control, ultracentrifuged, one-step interfacial trapped, two-step interfacial trapped, and three-step interf acial trapped SC-SWNTs, respectively..............53 9

PAGE 10

3-2 A) The pH-dependent changes to the zet a-potential of SC-SWNTs B) The effect of pH on the intensity/absorbance ratio after interfacial tr apping. The surface charge was adjusted by changing the pH of the susp ensions prior to IT. C) The effect of surface charge on the intensity of (6,5), (8,3), and (10,5) SWNT types after interfacial trapping. All inte nsities are normalized to the maximum intensity. D) pHdependent Raman spectra (784 nm excitation) of the radial breathing modes (RBMs) of SWNTs. Spectra are normalized to the G band. E) The effect of surface charge on the Raman aggregation ratio..............................................................................................54 3-3 The effect of solvent on the fluoresce nce intensity of ultracentrifuged SC-SWNTs (662 nm excitation). The ultracentrifuged SWNTs (1) are first mixed with toluene (2). The toluene was then removed by evaporation (3).....................................................55 3-4 Comparison of SC-SWNT suspensions prepared by the pH-adjusted (pH 6.75) twostep IT process (1), ultracentrifugation (2), and dilution of the IT process (3) obtained from (1). The ratio of fluores cence intensity to absorbance for SWNT suspensions excited at A) 662 and B) 784 nm. C) Raman spectra of the SWNT RBMs. D) Absorption spectra of the suspensions.............................................................55 4-1 Comparison of NIR fluores cence spectra (ex = 662 nm) of surfactant-coated SWNT suspensions before ( 1 and 3) and after mixing with ODCB ( 2 and 4). A) SDBSsuspended SWNTs and B) SDS-suspended SWNTs.........................................................71 4-2 Comparison of normalized NIR fluorescen ce spectra (ex = 662 nm) of surfactantcoated SWNT suspensions to SWNTs dispersed in only ODCB ( 1). A) SDBSSWNTs ( 2) and B) SDS-SWNTs ( 3) after mixing with ODCB........................................71 4-3 A) NIR fluorescence spectra (ex = 66 2 nm) of SDS-coated SWNT suspensions mixed with ODCB during solvent evaporation. B) Time-dependent recovery of (7, 6) and (8, 3) peak intensities of SDS-SWNTs mixed with ODCB........................................72 4-4 NIR fluorescence spectra of SDBSand SD S-coated SWNT suspensions after being mixed with the immiscible organic solvents: hexane ( 2), benzene (3), ODCB ( 4), or chloroform (5). The initial surfactant-SWNT suspension ( 1 ) is plotted for comparison. Excitation from 662 and 784 nm lasers.........................................................73 4-5 NIR fluorescence spectra of SDBSa nd SDS-coated SWNT suspensions after evaporation of the immiscible organic solvents: hexane ( 2 ), benzene ( 3 ), ODCB ( 4), or chloroform ( 5). The surfactant-SWNT suspension exposed to air for 24 h ( 1) is plotted for comparison. Note that both SDSand SDBS-coated SWNT suspensions showed minor changes to the emission sp ectra after being op en to air for 24 h. Excitation from 662 and 784 nm lasers.............................................................................74 10

PAGE 11

4-6 A) Absorbance and B)D) Raman spectra (ex = 785 nm) of SDS-coated SWNT suspensions after evaporation of the organic solvent: hexane ( 2), benzene ( 3), ODCB ( 4), or chloroform (5). The surfactant-SWNT suspension exposed to air for 24 h ( 1) is plotted for comparison. The radial breathing modes (RBMs) for specific ( n, m ) SWNT types are shown in B) while C) a nd D) show the Raman spectrum of the G and G band of the SDS-coated SWNTs susp ensions, respectively. The dashed lines indicate the positions of the surfactant suspensions prior to mixing with organic solvents. All spectra in C) and D) are offset for clarity.....................................................75 4-7 A) Swelling of the hydrophobic core of the micelle surrounding SWNTs. B)-E) The effect of pH on the fluorescence intensity of the (7, 6) SWNT (ex = 662 nm) and (10, 5) SWNT (ex = 784 nm) in SDS-suspensions after being mixed with the immiscible organic solvents benzene and ODCB and after evaporation of the solvent.......................76 5-1 PL spectra ( E 11 emission) for the A) (7, 6) a nd B) (10, 5) SWNT types and C) E 11 and D) E 22 absorbance for the (7, 6) SWNT type measured in microenvironments of various nonpolar solvents..................................................................................................87 5-2 A) The PL emission spectra for the ( 10, 5) SWNT in a hexane/ODCB mixture at different mole fract ions of ODCB: ( 1) x ODCB = 0.0; ( 2) x ODCB = 0.2; ( 3) x ODCB = 0.4; ( 4) x ODCB = 0.6; ( 5) x ODCB = 0.8; ( 6) x ODCB = 1.0. The dashed line is the position of the E 11 emission energy in air. The B) peak shif t (relative to air no t the initial SDBS suspension) and C) intensity changes for multiple ( n, m ) types at different mole fractions of ODCB in a hexa ne/ODCB mixture. The dashed line in C) are guides to show the trend....................................................................................................................87 5-3 Solvatochromic shifts of various ( n, m ) SWNT types in nonpol ar solvents with different solvent induction polarization, f ( 2 )..................................................................88 5-4 Comparison of the peak positions for the (7,6) SWNT type suspended in surfactants and suspended in sodium dodecylbenzenesulfonate (SDBS) but surrounded by various nonpolar solvents. Sodium cholat e (SC), sodium dodecylbenzenesulfonate (SDBS), sodium dodecyl sulfate (SDS), and cetyl trimethylammonium bromide (CTAB)..............................................................................................................................88 A-1 A) Absorbance at 763 nm of the (7, 6) SWNT type as a function of nanotube concentration. Initial SWNT concentrati on of 1 mg/mL. B) Fluorescence emission intensity of the (7, 6) SWNT type (Ex = 662 nm) as a function of nanotube concentration. Initial SWNT concentration of 1 mg/mL...................................................95 A-2 A) Corrected absorbance of ultracentrif uged, singleand two-step interfacial trapped SWNT suspensions after subtracting th e non-resonant background. B) Corrected absorbance of interfacial trapped SWNT suspensions at different oil-to-aqueous volume ratios after subtracting the non-resonant background...........................................95 11

PAGE 12

B-1 PL emission spectra of SWNTs excited at A) 662 and B) 784 nm and C) absorbance spectra of SWNTs in nonpolar microenvir onments. D) PL emission spectra (Ex. = 784 nm) of SWNTs in various hexane/ ODCB mixture microenvironments. All spectra in a-c are offset for clarity an d arranged by dielectric constant..........................100 B-2 Deconvolution of PL emission spectra for excitation at A) 662 and B) 784 nm. Weighting factors were assigned to each p eak based on the amount of overlap with other peaks and the intensity. Those peaks highlighted in A) show examples of the three weighting factors associ ated with overlap while B) shows those associated with intensity. The green, blue, and purple curv es were assigned values of 1, 0.5, and 0.25, respectively.............................................................................................................101 12

PAGE 13

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy APPLICATION OF LIQUID-LIQUID INTER ACTIONS WITH SINGLE-WALLED CARBON NANOTUBES By Randy Kai-Wei Wang August 2009 Chair: Kirk J. Ziegler Major: Chemical Engineering This study covers three important research topics related to the application of liquid-liquid interaction with single-walled ca rbon nanotubes (SWNTs). The firs t topic describes the removal of SWNT bundles from liquid susp ensions of nanotubes. The key to this work includes the use of liquid-liquid interfaces to trap the SWNT bundles due to the free energy change of the system during the process. SWNTs pack into crystalline ropes that fo rm bundles due to strong van der Waals attraction. Bundling diminishes mechanical and el ectronic properties because it could interrupt the electronic structure of the nanotubes. Als o, the electronic devi ces based on as-grown nanotubes, which contains a mi xture of individual nanotubes and nanotube bundles, make the electrical response unpredictable. We develope d a new simple process to remove bundles by liquid-liquid interaction. SWNTs bundles are trap ped at the interface be cause bundles stabilize the emulsions. Eliminating the use of ultracen trifugation to remove SWNT bundles enables large-scale production with reduced production costs and time savings. The second topic presented the swelling effect of the surfactant layer surrounding SWNTs with nonpolar solvents. Solvatochromic shifts in the absorbance and fluo rescence spectra are observed when surfactant-stabilized aqueous SW NT suspensions are mixed with immiscible 13

PAGE 14

organic solvents. When aqueous surfactant-suspended SWNTs are mixed with certain solvents, the spectra closely match the peaks for SWNTs di spersed in only that solvent. These spectral changes suggest the hydrophobic region of the micelle surrounding SWNTs swells with the organic solvent when mixed. The solvatochromic shifts of the aqueous SWNT suspensions are reversible once the solvent evaporates. Howeve r, some surfactant-solvent systems show permanent changes to the fluorescence emission inte nsity after exposure to the organic solvent. The intensity of some large diameter SWNT ( n, m ) types increase by more than 175%. These differences are attributed to surfactant reorganization, which can improve nanotube coverage, resulting in decreased exposure to quench ing mechanisms from the aqueous phase. The third topic describes the further study of the solvatochr omism of the SWNTs. Since SWNTs are encapsulated with microenvironmen ts of nonpolar solvents, it provides a new method to measure the photophysic al properties of nanotubes in environments with known properties. Fluorescence and absorbance spectra of SWNTs show solvatochromic shifts in 16 nonpolar solvents, which are proportional to the solvent induction polarization. The photophysical properties of SWNT s were used to determine the relationship between the longitudinal polarizability a nd other nanotube properties, 11,|| 1/( R 2 E 11 3 ). 14

PAGE 15

CHAPTER 1 INTRODUCTION Single-walled carbon nanotubes (SWNTs) are ultrathin and impressively long macromolecules made of pure carbon with structur es as regular and symmetric as crystals. They can be utilized in devices at a molecular scal e and offer the possibili ty of shrinking devices further because the speed, densit y, and efficiency of devices a ll rise rapidly as the minimum feature size decreases. SWNTs have a tensile strength 20 times that of steel, carry 1000 times more current density than copper wires, a nd transport charges down the nanotube without significant scattering. [ 1 ] There are approximately 30 different ( n, m ) types (approx. 1/3 metallic; 2/3 semiconducting) produced in the raw HipCo SWNT materials. The index ( n, m ) defines the length and angle ( ) of the SWNTs roll-up vector on a graphene sheet shown in Figure 1-1 [ 2] The difference between metallic and semiconducting nanotubes is only small changes in the angle by which the graphene sheet is rolled into a nanotube. Therefore, small changes in lead to different ( n, m ) types, resulting in different electr ical properties for both metallic and semiconducting types since each different (n, m ) has specific energy states. 1.1 The Electronic State of SWNTs Generally, all nanotubes can be described by the indices | n m | = 3 q. If q is an integer, the tubes are assigned as metallic or semimetallic. Otherwise, they are semiconducting nanotubes with geometry-dependent bandgaps. Therefore, each ( n, m ) type of semiconducting nanotubes generates near-infrared (NIR) fluorescence of a specific wavele ngth; the intensity of the fluorescence is based on the concentration of each (n, m ) semiconducting tube. The density of electronic states (DOS) desc ribes the number of electronic states at a particular energy and is used to understand the op tical transitions of SWNTs. As shown in Figure 1-2 one-dimensional materials, such as SWNTs, show sharp peaks in the DOS, which are called 15

PAGE 16

van Hove singularities and are similar to molecula r energy levels. Therefor e, optical transitions of SWNTs occur betwee n matching peaks in the 1D DOS. When a SWNT absorbs energy ( E 22 ), the electron is excited from th e valence band to the conduction band. Then, after the electron loses energy through nonradiative relaxation to the lowest con duction band energy state, the electron recombines with the hole in the vale nce band, yielding radia tive fluorescence emission ( E 11 ) (It is also the bandgap of SWNTs). [ 3] Because the energy states vary for each (n,m) type, E 11 and E 22 vary with SWNT diameter. 1.2 Characterization of SWNTs Most applications employing the unique electronic, thermal, optical, and mechanical properties of individual SWNTs [ 4-7 ] and any separation processes will require the large-scale manipulation of stable suspensions at high we ight fraction. Individual SWNTs suspensions will bring nanotube science into be tter contact with fundamental research on self-assembly in complex fluids. Unfortunately, na notubes bundle easily and are difficult to suspend as a result of substantial van der Waals attractions between tubes. [ 8] Bundled nanotubes also perturb the electronic structure of semi conducting SWNTs, quenching thei r intrinsic fluorescence, [ 2, 9, 10] limiting all attempts to separate the tubes by si ze or type or to use them as individual macromolecular species. Generally, optical absorption, [ 11] NIR fluorescence, [ 2, 9, 10] and Raman spectroscopy [ 12] have been the most common measurement te chniques to characterize and evaluate the dispersion quality of SWNT suspensions. NIR fluorescence provides direct evidence of the dispersion of individual SWNTs because only individual semiconducting nanotubes emit fluorescence radiation. [ 9] The Raman Radial Breathing Mode s of nanotubes are sensitive to disturbances from their surroundings. Therefore, these modes result in an aggregation peak, 16

PAGE 17

which is measure of the extent of dispersion. [ 13] In contrast, optical absorption spectroscopy has limited sensitivity to detect the bundling in SWNT suspensions. [ 14] The surrounding media of SWNTs can influe nce the optical measurements, such as absorption and fluorescence, by showing the charact eristic peak shifts in spectra. It has been shown that the optical transition energy of SWNTs in various solvents and surfactants shifts to lower energies with increasing dielectric constant of the surrounding media. [ 15, 16] This environmentally induced energy shift is due to the rearrangement of the media molecules to solvate the dipoles of SWNTs, causing a change in the reaction field. Therefore, each solvent yields unique emission or adsorpti on spectra called the solvatochromic shift. Investigation of this solvatochromic shift should allow us to understand the photophysics of SWNTs in different dielectric environments and help us characte rize the surfactant stru cture surrounding SWNTs. 1.3 The Surfactant Structure on SWNTs SWNTs can be dispersed in aqueous media wh en coated by adsorbed surfactants. This simple non-covalent method is able to disperse both organic and inorganic particles since the hydrophilic head groups help partic les with limited solubility di sperse homogeneously in an aqueous solution. Sodium dodecyl sulfate (SDS) and sodium dodecylbenzene sulfonate (SDBS) are commonly used ionic surfactants to generate high quality dispersions of individual SWNTs due to the electrostatic repulsion. OConnell et al. [ 9] have suggested that SDS forms a cylindrical structure around na notubes. Others have shown th at either a hemimicellar [ 17] or random distribution of SDS surfactant forms on th e nanotubes without pref erential arrangement of the head and tail groups when the surf actants adsorb on the sidewall of SWNTs [ 18 ] Figure 1-3 shows the schematic representa tion of possible adsorption stru ctures of SDS surfactants. [ 18] Nonionic surfactants such as natural [ 19] and artificial polymers [ 15] are also used to produce SWNT suspensions. These two kinds of surfactant have bulky hydrophilic groups that 17

PAGE 18

are able to prevent the aggregat ion of nanotubes due to the enhanc ed steric stabilization provided by longer polymeric groups. However, the real surfactant structure remains unknown whenever the ionic or nonionic surfactants are used. Wallace et al. [ 20] have suggested that the concentration of surfactant determines the mech anism of adsorption on the SWNTs. They have also suggested that the chirality of the SW NTs can influence the wrapping angle of the surfactants on the SWNT at low concentration. Therefore, it is clear that the concentration of surfactants, which affect the micelle structure, plays an important role on understanding the dispersion mechanism of the SWNT in the aqueous media. 1.4 Pickering Emulsions Nanoparticles, such as SWNTs, can not only be suspended by the surfactant but also act as surfactants to stabiliz e the emulsions. These particle-sta bilized emulsions, also known as Pickering emulsions, [ 21] are extremely stable and have been used to self-assemble nanoparticles at the liquid-liquid interfaces, [ 22, 23] to separate the particles, [ 24] and to prepare unique porous structures. [ 25] Figure 1-4 is the typical Pickering emulsions stabilized by Gum Arabicsuspended SWNTs. In general, the stability of the Pickering emulsions is determined by the extent to which the particles are wetted by th e two immiscible fluids, particle sizes, concentration, and mu tual interaction. [ 26] Particles with higher hydr ophilicity (having smaller contact angles ow ) will preferentially stabilize oilin-water emulsions, and more hydrophobic particles (having larger contact angles ow ) will tend to stabilize the water-in-oil emulsions. These emulsions also provide a new route to de velop large-scale purifi cation of carbon nanotube suspensions, which is described further in Chapters 2 and 3. 18

PAGE 19

1.5 Bundle Removal and Type Separation of SWNTs The general method for produci ng individually suspended nanot ubes involves high-shear homogenization, ultrasonication, and ultracentr ifugation to remove the SWNT bundles from individual SWNTs. [ 9] However, ultracentrifugation is not suitable for large-scale production because of its limited scale, high price, and tim e-consuming process. This leads to negative economic impacts. Therefore, a faster and hi gher throughput technique to remove nanotube bundles will be required for the successful genera tion of individual SWNTs for further potential applications. The mixture of metallic and semiconducting type s in raw SWNTs is another drawback for the applications since it limits the performance of devices because devices based on as-grown SWNTs make the electrical response unpredictable. Separation processes offer the possibility of dividing the raw SWNTs material into the me tallic and semiconducting nanotubes. Researchers [ 27] have shown that electrophoresis can be utilized to obtain a degree of separation. Rinzler and coworkers [ 28] have suggested that the complexation of metallic SWNTs with bromine is able to separate the mixture of tubes in tw o types. In addition, Zheng et al. [ 29] have developed the DNA-based chromatography technique to achieve separation. Recently, Density differentiation has also been discovered to sort SWNTs with different electronic structure. [ 30] Although the methods described above are able to separate SWNT in type, the major problem with these techniques is that they are onl y analytical-scale techniques. Therefore, a large-scale production of separation method will be necessary to generate specific SWNTs type. In order to remove the SWNT bundles a nd enable type sepa ration, liquid-liquid interactions are utilized since it is a low-cost, time-saving and easy-to-scale-up process. As described in Section 1.4, solid particles adsorb at liquid-li quid interfaces stabilizing Pickering emulsions, which can be an appropriate method to remove SWNT bundles. In addition, the 19

PAGE 20

interfaces have high capacities to open a new rout e for large-scale technique. Hence, it provides an opportunity to remove the bundles dispersed in the aqueous phase from individual tubes. 1.6 Applications of SWNTs The development of purification and type separation techniques will open up a multitude of applications in the material s science, microelectronic, bion anotechnology, and medical fields. Individual semiconducting SWNTs with specific ( n, m ) types can be utilized in the construction of microelectronic transistors with known band gaps. Semiconducting SWNTs have been used as a channel through which electron flow in fieldeffect transistors (F ETs) between two metal electrodes with SWNTs as shown in Figure 1-5 The gate electrode can change the conductivity of the nanotube channel in a FET by a factor of one million or more comparable to silicon FETs. Also, because of its nanoscale size, the nanotube FETs should switch reliably using much less power than silicon-based devices. In general, the requirement for th e devices is to be made much smaller because they could pack more devices into the same area. However, the problem to limit shrinking metal wires further is that there is st ill no good way to remove the heat produced by the devices, so packing them in more tightly will only lead to rapid overh eating. SWNTs could solve the problem since scientists have found that nanotubes could conduct heat nearly as well as diamond or sapphire. Therefore, nanotubes could efficiently cool very dense arrays of devices. Individual semiconducting SWNTs were also sugge sted to be utilized as biosensors to replace traditional dyes suffering from bleachin g, degradation, and toxicity problems. [ 31] They are expected to detect single enzyme or prot ein molecules accurately, pr oviding high-sensitive biosensing applications. Strano and coworkers [ 32] have developed a SWNT-based implantable biosensor, which is able to detect the -D-glucose in solution phase based on their inherent 20

PAGE 21

fluorescent properties. Therefore, the ability to synthesize or se parate the SWNTs by type should aid the development of biosensors. For metallic SWNTs, it can be used as ball istic conductors because they have little electrical resistance which is usually a disadvant age in metal wires. Since electrons can travel long distances without scattering, they can maintain their quantum states. It is also appropriate to utilize them as optical fibers for telecommunication over long dist ances since the signal will not be decayed. Also, metallic nanotubes can tr ansport electric current greater than 10 9 A/cm 2 which will vaporize metals, such as copper or gold. [ 33] Metallic SWNTs can also be applied to photovoltaic cells since they can be as conductive transparent electrodes, which are a crucial component in solar cells. 21

PAGE 22

Figure 1-1: SWNTs can be viewed as a grapheme layer wrapped into a tubular structure. The vector connecting two points on the plane can be described by the indices n and m [ 2] -5 -4 -3 -2 -1 0 1 2 3 4 5 0246810 v1c1v2 Density of Electronic StatesEnergyc2 conduction valence E22absorption E11 emission e-hole -5 -4 -3 -2 -1 0 1 2 3 4 5 0246810 conduction v1c1v2 Density of Electronic StatesEnergyc2 E22absorption E22absorption E11 emission E11 emission e-hole valence Figure 1-2: Density of electronic states for a single nanotube. Co lor arrows indicate the optical transitions; black arrows represent the nonrad iative relaxation of the electron (in the conduction band) and hole (in the valence band). [ 2] 22

PAGE 23

A B C Figure 1-3: Possible adsorpti on structures of SDS surfact ants on a SWNT: A) SWNT encapsulated in a cylindrical structure. B) Hemimicellar structure of SDS surfactants on a SWNT. C) Random distribution of SDS surfactants on a SWNT. [ 18 ] Figure 1-4: Toluene/water em ulsions stabilized by surf actant-suspended SWNTs. [ 34] 23

PAGE 24

Figure 1-5: The SWNT-based transistor. As ul trathin wires, SWNTs could free up space in microchips for more devices, as well as solving heat and stability problems. At a little over a nanometer in diameter, this SWNT makes lines drawn by state-of-the-art photolithography look huge in comparison. [ 1] 24

PAGE 25

CHAPTER 2 INTERFACIAL TRAPPING IN REMOVING SWNT BUNDLES SWNTs pack into bundles because of the strong van der Waals attractive forces. [ 8, 35] High-energy ultrasonication is often used to ove rcome these forces to obtain individual SWNTs in solution. However, many common solvents have insufficient solvation forces to suspend SWNTs, yielding low degrees of solubility. [ 36] These suspensions consist of many small bundles and relatively few individual SWNTs. The difficulties associated with obtaining individually dispersed SWNTs li mits nanotube applications, [ 35, 37, 38] leading researchers to develop a multitude of functionalization sc hemes to achieve nanotube suspensions. [ 36, 38-40] Noncovalent functionalization r outes are preferred in many a pplications to preserve the properties of SWNTs. However, researchers have noted that these approaches yield a solution containing individual nanotubes as well as nanotube bundles. [ 2] The bundling of nanotubes diminishes the mechanical properties in nanocom posites since the SWNTs slip past one another under stress. [ 41] Bundled nanotubes also perturb the electronic structure of semiconducting SWNTs, quenching their in trinsic fluorescence. [ 2, 9, 10] The conventional method to disperse individual nanotubes in aqueous solutions is by high-shear homogenization in various surfactant solutions [ 15, 19] followed by ultrasonication and ultracentrifugation. [ 9] However, ultracentrifugation is a time-consuming approach to the removal of SWNT bundles. Therefore, alternative routes to remove bundles from SWNT suspensions are needed for economic, largescale dispersion. In this chapter, we describe a simple a lternative approach to ultracentrifugation using liquidliquid interfaces to remove nanotube bundles from the aqueous SWNT suspension. This technique has already been combined with density gradient ultracentrifugation [ 30] to significantly improve field-effect transistor performance. [ 42 ] The adsorption of colloidal 25

PAGE 26

particles at interfaces was first ch aracterized by Pickering in 1907. [ 21] These systems have been used to self-assemble part icles at the interface, [ 22, 23] to separate particles, such as ampicillin and phenylglycine crystal mixtures in water/alkanol systems, [ 24, 43] and to prepare unique porous structures. [ 25] Pickering systems have also demons trated the large-scale separation of bioparticles, achieving efficiencies greater than centrifugation. [ 44] Wang, Hobbie and coworkers [ 45, 46] were the first to show SWNT-based st abilization of emulsions. Bare nanotubes were used as amphiphobic surfactants that stab ilized toluene/water emulsions for months. [ 45] Later, DNA-wrapped SWNTs were shown to stabili ze emulsions for the synthesis of colloidal particles. [ 46] Stabilized emulsions were also seen in length-based separations of functionalized SWNTs. [ 47 ] More recently, researchers have begun to use SWNT-based Pickering emulsions for other applications. Asuri et al. [ 48 ] demonstrated that interfacial SWNTs decreased transport limits and improved catalytic activ ity of two-phase reactions lead ing to increased bioreactivity. Others used polymerization reactions [49, 50] or nanotube interactions [ 51] to prepare nanotube capsules for supports, contro lled release capsules, [ 51] and lubricating additives. [ 25] The method and mechanism of the interfacial trapping process are also provided for removing nanotube bundles from aqueous suspensions. Near-infrared (NIR) fluorescence, vis NIR absorbance, and Raman spectroscopy as well as atomic force microscopy (AFM) and rheology are used to characterize the SWNT susp ensions. New approaches to improve dispersion quality are also discussed. The suspensions prep ared by interfacial trapping have comparable dispersion quality to those prepared by ultracentrifugation. 2.1 Experimental Procedure 2.1.1 Dispersion Nanotube suspensions were prepared with a given initial mass (typi cally between 6 and 20 mg) of raw SWNTs (Rice HPR 145.1) and mixe d with 200 mL of an aqueous Gum Arabic 26

PAGE 27

(Sigma-Aldrich) surfactant solution [ 19] (1 wt %) to form an initial concentration of 0.03.2 mg/mL. High-shear homogenization (IKA T-25 Ultra-Turrax) for 1 h and ultrasonication (Misonix S3000) for 10 min aided dispersion. Cont rol samples were ultracentrifuged at speeds between 20,000 and 26,000 rpm (Beckman Coulter Optima L-80 K) for 4 h to remove nanotube bundles. 2.1.2 Interfacial Trapping Toluene (Acros, 99%) was added to the aqueous SWNT suspension to form a two-phase system. The two-phase system was then shaken vigorously for 30 s to increase interfacial area. Phase separation occurs within 1 min; however, the solutions were allowed to settle for 30 min to ensure that steady state was achieved for spectroscopy. All expe riments were conducted at an organic to aqueous SWNT suspension vol ume ratio of 0.1 unless otherwise indicated. 2.1.3 Characterization The aqueous phase was carefully removed after interfacial trapping to prevent further emulsification. The aqueous phase was char acterized by visNIR absorbance and NIR fluorescence spectra using an Applied NanoFl uorescence Nanospectrolyze r (Houston, TX) with excitation from 662 and 784 nm diode lasers. A concentrated Gum Arabic SWNT suspension was homogenized and ultrasonicated and then in crementally diluted with 1 wt % Gum Arabic solution to determine the linear regions of th e absorbance and fluorescence spectra (Figures A-1 (A) and A-1 (B) in the Appendix A ). Raman spectra were recorded using a Renishaw Invia Bio Raman with excitation from a 785 nm diode laser. SWNT suspensions were also spin-coated onto mica to collect AFM images on a Digital Instrument s Dimension 3100. Diameter analysis was then performed using SIMAGIS software. [ 52] A minimum of 700 nanotubes were measured for statistically relevant results. Th e rheology of SWNT suspensions was examined using an ARES LS -1 strain-controlled rheometer (TA Instru ments). A cone and plate fixture with 27

PAGE 28

a diameter of 50 mm and an angl e of 0.04 rad was used to measur e the steady shear viscosity at rates between 1 and 100 s -1 As an additional step in preparing the interfacial trapped samples for rheological characterization, the Gum Arabic solution was initially centrifuged at 26000 rpm for 30 min before adding nanotubes. Solutions of Gum Arabic, prior to centrifugation, showed variations in rheology, consis tent with other reports. [ 53, 54] The additional processing produced a stable solution of Gum Arabic with a viscosity of 1.68 0.05 cP as measured over the entire range of shear rates. For rheo logical characterization, SWNT su spensions were ultracentrifuged for 2 h to prepare similar final concentrations for analysis (determined using the extinction coefficient [ 55] and absorbance at 763 nm). 2.2 Results and Discussion 2.2.1 Selective Separation of Bundles After mixing toluene with a homogenized a nd ultrasonicated aqueous suspension, the system forms a homogeneous gray solution of emulsions. An interphase of emulsions and a bulk aqueous phase forms within a minute as seen in the inset of Figure 2-1 Absorbance spectra of the bulk aqueous phase are shown in Figure 2-1 (A). The homogenized and ultrasonicated sample (control) has high absorbance due to th e concentration of both individual (as evidenced by the interband transition peaks) and bundled SWNTs. The absorbance of the suspension has clearly decreased after in terfacial trapping while the spectral features have blue-shifted and are better resolved. These changes in dicate the removal of nanotubes from the aqueous phase and a higher fraction of individual SWNTs. [ 56] When the control suspension is ultracentrifuged, the absorbance of the aqueous phase is significan tly lower demonstrating that ultracentrifugation removes a significant portion of the nanotubes (bot h bundled and individual) at the relatively low concentrations used here. 28

PAGE 29

Fluorescence spectra of the aque ous phase are shown in Figure 2-1 (B). For comparison, the spectra after homogenization and ultrasonication are shown (contro l) as well as the spectra using conventional ultracentrifuga tion rather than inte rfacial trapping. The spectra show that ultracentrifugation results in a decrease in fluorescence intensity indicative of the removal of individual nanotubes. However, the fluorescence intensity after interfacial trapping increases when compared to the control sample. It is also important to note that the peak positions for both the interfacial trapped and u ltracentrifuged suspensions are id entical, indicating that the environment surrounding the nanotube has not changed after mixing with the solvent. Fluorescence intensities are related to the concentration of individual nanotubes. Fluorescence spectra also provide a sensitive probe to the aggregat ion state of the aqueous phase. [ 9, 14] Typically, higher intensity peaks in the spectra indicate better dispersion quality [ 14] since energy transfer can o ccur between adjacent nanotubes in bundles resulting in either intensity changes or complete quenching of the fluorescence. [ 2, 10] This energy transfer mechanism can also occur between individual SWNTs. Several groups [ 57-59] have observed that energy transfer between SWNTs results in fluorescence changes. The fluorescence intensity of individually suspended SWNTs may decay as the volume fraction increases as observed in solid composites. [ 60] Concentrated SWNT suspensions coul d also potentially reabsorb emitted photons thereby reducing the measured emission in tensity. However, the dilute initial SWNT concentration (0.03 mg/mL) is in the linear region of both fluorescence and absorbance measurements (see Figures A-1 (A) and A-1 (B) in the Appendix A ). Therefore, the increase in fluorescence intensity observed in Figure 2-1 (B) is likely due to reduced quenching after interfacial trapping. Removal of nanotubes from the solution, especially SWNT bundles that 29

PAGE 30

quench fluorescence, would allow more indi vidual, semiconducting SWNTs to emit photons, resulting in increased fluorescence intensities. SWNT suspensions with a much lower concentr ation were used during interfacial trapping (0.005 mg/mL) to test whether the enhan cement of the fluorescence seen in Figure 2-1 was due to the removal of SWNT bundles. The lower initial concentration results in fewer collisions between nanotubes and, thus, fewer potential quenc hing events. At these c oncentrations, changes to the number of quenching bundles in the susp ension should have a reduced effect on the spectra. Figure 2-2 (A) shows that fluorescence at low conc entrations has only a slight change after interfacial trapping, whereas the absorbance in Figure 2-2 (B) has diminished significantly. Since the fluorescence remains relatively constant the clear decrease in absorbance can only be attributed to the removal of bundles or other carbonaceous impurities from the aqueous phase, suggesting that selective bundle removal occurs. 2.2.2 Dispersion Quality Quantifying the fraction of i ndividually suspended SWNTs or bundles in an aqueous solution remains difficult. [ 14] Rheology, [ 14, 55 ] fluorescence, [ 9, 10, 14] Raman, [ 14, 61] absorbance, [ 56] and small-angle neut ron scattering (SANS) [ 14, 46, 62, 63] have already been successfully used to study the interactions of surfactant-coated SWNTs in aqueous solutions. Visual methods such as AFM are also used to obtain qualitative and quantitative information about the dispersion. [ 17, 52 ] The group from NIST recommended that multiple techniques be used to quantify dispersion. [ 14] Several techniques that descri be the dispersion quality of the aqueous suspension after interfacial trapping are discussed below. A higher initial SWNT concentration of 0.2 mg/mL was used to investig ate the dispersion quality, and the results are summarized in Table 2-1 30

PAGE 31

The relative intensity of the interband transi tion peaks in the absorbance spectra and their width are related to th e dispersion quality. [ 9, 14, 56] Tan and Resasco [56] showed that the resonance ratio could be used to characterize the fraction of i ndividual SWNTs in solution. This ratio is calculated by comparing the area for the resonant band with the nonresonant band (carbonaceous impurities and -plasmon). The interfacial trapping method shows improved dispersion characteristics and improved frac tions of individually suspended SWNTs. Figure 2-3 (A) is the Raman spectra of the SWNT radial breathing modes (RBMs). The peak at ~270 cm -1 represents the shift of (10, 2) nanotubes due to bundling, providing a measure of the extent of nanotube aggregation. [ 13] The homogenized and ultrasonicated suspension has a very intense aggregation peak in comparison to the (11, 3) nanotubes at ~234 cm -1 As seen in Figure 2-3 and Table 2-1 the ratio of these peaks drops dram atically after interfacial trapping, indicating removal of bundled SWNTs. Figure 2-3 (B) compares the steady values of the shear viscosities for the three different SWNT suspensions. The viscosity of all systems n early matches at the highest shear rate of 100 s -1 demonstrating that the small difference in concentration across the three systems has a limited impact on the results. Differences in the rheology are apparent at lower shear rates. Except for the lowest shear rate, the control sample has a nearly constant viscosity as a function of shear rate. The ultracentrif ugation and interfacial trapping me thods produce samples that shear thin. The decrease in viscosity with increas ing shear rate is more pronounced for the ultracentrifugation sample but agrees within meas urement error with the result from the sample prepared using interfacial trapping. The favorab le comparison implies that our method produces a suspension of nanotubes similar in flow charac teristics, and hence dispersion quality, to the ultracentrifugation method. 31

PAGE 32

Parts a and b of Figure 2-4 show AFM images of the hom ogenized and ultrasonicated SWNT suspension compared to the suspension af ter interfacial trapping while the corresponding histograms of nanotube diameters are shown in Figure 2-4 (C). As expected, the homogenized and ultrasonicated SWNT suspension has many bundled SWNTs and some individually suspended SWNTs (diameters less than ~3 nm), yielding a broad diameter distribution with an average of 12.8 nm. After interf acial trapping, the suspension ha s significantly fewer bundled SWNTs and a sharper distri bution in diameters with an average of 4.1 nm. The fluorescence intensity provides a measure of the concentration of individually suspended SWNTs but does not account for varia tions in total concentr ation (i.e. fraction of individual SWNTs). On the othe r hand, the absorbance provides an overall measure of the entire ensemble of individual and bundled SWNTs, providing a means of normalizing the fluorescence intensity. The ratio of fl uorescence to absorbance ( F / A ), therefore, is another measure of dispersion quality. This ratio s hould be the most relevant measurement to many applications since it is related to the quantum yield of the na notube suspension. Note that this ratio is not a quantitative measure of the fraction of individua l SWNTs and it should only be used to compare dispersion characteristics between samples. In this study, the intens ity from the (7, 6) nanotube is used and divided by the absorbance at the exci tation wavelength (662 nm). The decrease in absorbance and increase in fluorescence cause the F / A ratio to improve after interfacial trapping as shown in Table 2-1 2.2.3 Improving Separation Effectiveness In summary, the SWNT suspension resulti ng from a single-step interfacial trapping process has improved dispersion quality in comparison to homogenized and ultrasonicated suspensions as shown in Table 2-1 SWNT dispersions prepared by ultracentrifugation have higher dispersion quality, but one of the benefits of the interfacial trap ping method is that the 32

PAGE 33

process can be easily adjusted to change separation efficiency. Figure 2-5 shows the variation of fluorescence emission intensity of the most promin ent (7, 6) nanotube type in the spectra as a function of the volume ratio of toluene to the aqueous SWNT suspension ( R = V toluene / V aqueous ). As seen in Figure 2-5 (A) for an initial SWNT mass c oncentration of 0.03 mg/mL, higher fluorescence intensity is observed for R 1. The higher conductivity of the solution when R > 1 suggests that water-in-oil (w/o) emulsions are formed while oil-in-wat er (o/w) emulsions are seen for R < 1. It is also seen that the fluorescence intensity is always greater than the control for initial mass loadings of 0.03 mg/mL. Figure 2-5 (B) shows the fluorescence intensity from an initial mass loading of 0.005 mg/m L of SWNTs. As can be seen in the figure, the emission intensity tends to be lower than the control for o/w emulsions, whereas w/o emulsions show relatively constant intensity. The same trends are seen for other emission peaks and other initial mass loadings (not shown). Although the fluorescence intensity changes are relatively minor, th e changes to the absorbance spectra in Figure 2-6 (A) are significant. The absorbance spectra for all volume ratios have decreased significantly afte r interfacial trapping when compared to the control sample in Figure 2-1 The interband transition p eaks are noticed for all aque ous solutions, indicating good dispersion. The absorption decreases steadily as the volume ratio is reduced and the system shifts from w/o to o/w emulsions. The lowest absorpti on intensities are seen for o/w systems at a volume ratio of R = 0.5. Figure 2-6 (B) plots the F / A ratio as a function of the volume ratio. Higher F / A ratios are seen for o/w systems when co mpared to w/o systems indicating that a higher fraction of individual nanotubes are susp ended in o/w systems via interfacial trapping. To further improve the quality of the suspen sions, a second interfacial trapping step was introduced. The SWNT suspension was first mi xed with toluene at a volume ratio of R = 0.1 33

PAGE 34

since o/w emulsions were the most effective at removing bundled nanotubes from the aqueous phase. The aqueous phase was separated from the oil and interphase and then mixed again with toluene at a volume ratio of R = 0.1. As seen in Figure 2-7 (A), the second in terfacial trap has little effect on the fluorescence intensity. Howeve r, the absorbance spectrum shown in Figure 2-7 (B) has decreased significantl y, resulting in significant cha nges to the fraction of bundled SWNTs (see Figure A-2 (A) in the Appendix A ). The Raman aggregation peak has also shown further improvement after the second inte rfacial trapping step as shown in Figure 2-7 (C). It is important to note that changes to both Raman and absorbance spectra without changes to the fluorescence provide strong evidence that the interf acial trapping process is highly selective in the removal of bundled SWNTs fr om the aqueous phase. Table 2-1 summarizes the dispersion quality measurements for the two-step interfaci al process compared to ultracentrifugation. As seen in the table, interfacial trapping show s better dispersion than ultracentrifugation by the F / A ratio and comparable dispersion quality when ch aracterized by Raman and absorbance spectra. 2.3 Mechanism Figure 2-8 shows a conceptual diagram of the overall two-phase interfacial trapping process based on experimental observations. First, nanotubes are homogenized and ultrasonicated in a surfactant so lution resulting in a suspension that contains both individually dispersed and bundled SWNTs. An immiscible organic solvent is added to the aqueous suspension forming a two-phase system. This two-phase system is then mixed resulting in emulsions (either o/w or w/o depending on th e volume ratios) as shown previously. [ 34 ] SWNT bundles preferentially adsorb at the emulsion in terface when mixed. The emulsions then coalesce into a continuous phase with some emulsions being stabiliz ed by the SWNT bundles. Finally, phase separation into an organic phase, an interphase of stabil ized emulsions, and a transparent aqueous phase allows easy collection of the individually suspended SWNTs. 34

PAGE 35

Although bundled nanotubes have been removed from the aqueous phase at all volume ratios (see Figures 2-5 and 2-6 ), the slight decrease in emission intensity for o/w systems indicates that individual SWNTs are removed from the aqueous phase. This conclusion is also confirmed by correcting the absorbance spectra for the nonresonant background (Figure A-2 (B) in the Appendix A ). This data suggests that o/w emul sions are more effective in removing nanotubes from the aqueous suspension. The pref erential adsorption of hydrophilic silica particles in o/w emulsions was also seen by Binks and co-workers. [ 26, 64] Adjusting the hydrophilicity of the particles had significant eff ects on the contact angle and their ability to stabilize the emulsions. The individual and bundled nanotubes are coated with hydrophilic surfactants also providing bette r stabilization of o/w emulsions and, hence, removal from the aqueous phase. In contrast, individual, Gum Ar abic-coated nanotubes do not seem to stabilize w/o interfaces as effectively which results in nanotube bundles being the primary constituent of the interface. This could indicate that the system contains a small fraction of uncoated (hydrophobic) or poorly coated nanotube bundles that stabilize w/o emulsions. Selective partitioning of SWNT bundles to the interface can be understood by describing the initial driving force of a single nanotube by free energy minimization. These simple models do not account for entropic or kinetic effects but have described inte rfacial assembly of nanoparticles as small as 2.8 nm. [ 22, 26, 44, 64, 65 ] Initially, the nanotubes are dispersed in the aqueous phase as shown in Figure 2-9 (A). If both individu al and bundled SWNTs are transferred to the oilwater interface upon mixing, the nanotubes will have a contact angle measured into the aqueous phase (see Fig. 2-9 (B)). On the basis of experimental evidence [ 49] and calculations, [ 50] SWNTs are expected to orient para llel to the interface. Therefore, the decrease in interfacial area for inserting a cylindrical particle of radius ( R SWNT or R bundle ) and 35

PAGE 36

length ( L ) between the oil and water phases is given by the expression, A ow = 2RL sin as shown in Figure 2-9 (B). The position of the SWNT at th e interface depends on the contact angle and results in a loss of particle interactions with the aqueous phase as well as increased interactions with the oil phase. If the interface is assumed planar and the weight of the nanotube is ignored, the energy change upon insert ing the nanotube at the interface is sin2 360 2 2ow pw poRL RLE (2-1) where po pw and ow are the interfacial tensions at the particleoil, particlewater, and oil water interface, respectively. If E is negative as expected for hydr ophilic particles, the particle will be in a stable position at th e interface. Substituting Youngs equation, po pw = ow cos, into equation, the energy change for inser ting a nanotube at the oilwater interface is sincos 180 2 owRLE (2-2) The contact angle will be similar for indi vidual and bundled nanotubes because of their similar hydrophilicity and ow is fixed in the system. Therefore, in aqueous SWNT suspensions, the change in energy of inserting a particle at the interface depends on R and L As seen in equation, the energy is minimized when particle s with larger radius and length are at the interface. For example, it is estimated that E is approximately kT for an individual nanotube and kT for a bundle of the same length co ntaining 7 nanotubes. Note from equation (2-2) that longer SWNTs will be pr eferentially removed from the suspension. Therefore, these estimates for E are likely the minimum initial dr iving force since bundles have lengths longer than individual nanotubes. These calculations are in agreement with the experimental results presented above as well as the results of Yi et al., [ 51 ] which showed reduced emulsion stabilization for short-lengt h multiwalled nanotubes. The relatively large 36

PAGE 37

negative free energy indicates that both individual and bundled SW NTs prefer to reside at the interface. This explains why o/w emulsions showed decreases in fluorescence intensity in Figure 2-5 (B) and fewer individual SWNTs in the corrected absorbance (see Figures A-2 (A) and A-2 (B) in the Appendix A ). On the other hand, there will be a competition for the limited interfacial sites in systems with high init ial concentration, where the bundles will be preferred because of the greater driving force. 37

PAGE 38

B A Figure 2-1. A) Absorbance and B) fluorescence (ex = 662 nm) spectra for the interfacial trapping process of Gum Arabic-suspended SWNTs us ing an initial SWNT mass concentration of 0.03 mg/mL. The control spectra (black lines) are the SWNTs after homogenization and ultrasonica tion. The inset shows the in terfacial trapping process in a separatory funnel. The control sa mple is then either subjected to ultracentrifugation (green lines ) or interfacial traps (red lines). The fluorescence from specific (n, m ) types are labeled. B A Figure 2-2. A) Absorbance and B) fluoresce nce (ex = 662 nm) spectra of SWNTs at a concentration of 0.005 mg/mL. The control sp ectra (black lines) are the SWNTs after homogenization and ultrasonication. The control sample is then subjected to interfacial traps. 38

PAGE 39

B A Figure 2-3. A) Normalized Raman spectra of the RBMs (ex = 785 nm) for Gum Arabicsuspended SWNTs after homogenization and ultrasonication (black line), interfacial trapping (red line), and ultracentrifugation (green line). The suspension was prepared from an initial concentration of 0.2 mg/m L raw SWNTs. B) Steady values of the shear viscosities for the Gum Arab ic-suspended SWNTs produced after homogenization and ultr asonication (control), in terfacial trapping, and ultracentrifugation. The suspensions were prepar ed at similar concentrations to enable comparison. The final concentration of the control, interfacial trapping suspension, and ultracentrifugation samples were 0.02, 0.026, and 0.025 mg/mL, respectively. 39

PAGE 40

B A C Figure 2-4. AFM images of Gum Arabic SWNT suspensions after A) homogenization and ultrasonication and B) after bundle removal by interfacial trapping. C) Histogram of nanotube diameters measured from the AFM images. The suspension was prepared from an initial concentra tion of 0.2 mg/mL raw SWNTs. 40

PAGE 41

A B Figure 2-5. Fluorescence emission inte nsity fluctuations of the (7, 6) nanotube as a function of the volume ratio ( R ). The sample was prepared from an initial mass concentration of A) 0.03 or B) 0.005 mg/mL raw SWNTs. Excitation at 662 () and 784 nm ( ). Conductivity measurements suggests the form ation of oil-in-water (o/w) emulsions (shaded area) for R < 1 and water-in-oil emulsions (w/o) for R >1. The dashed lines represent the fluorescence spectra for the suspension after homogenization and ultrasonication B A Figure 2-6. A) Absorbance spectr a of Gum Arabic-suspended SWNTs and B) fluorescence to absorbance ratio fluctuations as a function of the volume ratio (R ). The fluorescence intensity was measured for the (7, 6) nanot ube and divided by th e absorbance at the excitation wavelength (dashed line at 662 nm ). The suspension was prepared from an initial concentration of 0.03 mg/mL raw SWNTs. 41

PAGE 42

B A C Figure 2-7. A) Fluorescence (ex = 662 nm), B) absorbance, and C) Raman RBM spectra (ex = 785 nm) of Gum Arabic-suspended SWNTs fo r one-step and two-step interfacial trapping separations compared to ultracen trifugation. The suspensi ons were prepared from an initial concentrati on of 0.2 mg/mL raw SWNTs. 42

PAGE 43

D C B A Figure 2-8. Overall process of removing SWNT bundles from aqueous suspensions via liquid liquid interfaces. Prior to in terfacial trapping, the raw SW NTs are first homogenized with a highshear mixer and ultrasonicated to coat SWNTs with surfactant (not shown), resulting in a mixture of indivi dually suspended SWNTs and SWNT bundles. A) Toluene is added to the aqueous phase and mixed to form emulsions, B) SWNT bundles are trapped at the emulsion inte rfaces, C) creaming and coalescence of emulsions separates the bundled SWNTs, a nd D) SWNT bundles are removed from the bulk aqueous fluid. 43

PAGE 44

A B Figure 2-9. Adsorption process of an individual or bundle of nanotubes at the interface of the oil and water phases. A) Prior to interfacial trapping, individual or bundled nanotubes are dispersed in the aqueous phase After mixing, the nanotubes a ssemble at the interface. B) Diagram showing the end of a nanotube at the interface where R is the radius of an individual or bundled nanotube and is the contact angle measured into the water phase. 44

PAGE 45

Table 2-1. Dispersion Quality Comparison of Aqueous SWNT Suspensions. 1 Calculation based on peak near 660 nm. aqueous suspension anaylsis control interfacial trapping ultracentrifugation two-step interfacial trapping fluorescence intensity weak strong strong strong absorbance features broad peaks blue-shifted and resolved peaks blue-shifted and resolved peaks blue-shifted and resolved peaks resonant Ratio 1 0.0118 0.0397 0.0486 0.0406 diameter distribution 2 12.8 11.9 nm 4.1 3.7 nm Raman aggregation ratio 3 0.7265 1.1868 1.8860 1.5952 F/ A Ratio 4 0.0135 0.2300 0.5234 0.6077 2 Distribution determined from AFM and SIMAGIS image analysis. 3 Calculation based on intensity from (11, 3) nanotubes and aggregation peak (~270 cm -1 ). 4 Calculation based on fluorescence intensity from (7, 6) nanotubes and absorbance at 662 nm 45

PAGE 46

CHAPTER 3 INTERFACIAL TRAPPING OF SWNT BUNDLES SUSPENDED WITH ANIONIC SURFACTANTS Single-walled carbon nanotubes (SWNTs) are one of the most widely researched nanomaterials due to their unique ph ysical and chemical properties. [ 4, 5] One of their unique properties is the emission of n ear-infrared (NIR) fluorescence, [ 2] which enables potential applications of SWNTs as op tical devices for bioimaging [ 66, 67] and biosensing. [32, 68, 69] However, most applications require suspen sions of individual SWNTs since bundled SWNTs can perturb the electronic structure of indi vidual SWNTs, quenching their fluorescence, [ 2, 9, 10] or slip past one another under stre ss, weakening composite structures. [ 41] SWNTs are typically dispersed in common solvents [ 36, 70] or in water with the aid of surfactants. [ 9, 15, 71] Ultrasonication is used to sepa rate individual SW NTs from SWNT bundles; however, the strong van der Waals attractive forces [ 8] make it difficult to isolate all SWNTs without significant damage. [ 16] Therefore, SWNT bundles still remain in the suspension, which can be removed by high-speed ultracentrifugation because of density differences between bundled and individual SWNTs. [ 9 ] However, the small mass requires high centrifugal forces (> 100,000g) and long processing times to achieve separation (~ 4 to 5 h). These requirements limit the scalability of the pr ocess. Ultracentrifugati on was the only route to remove the remaining bundles until our group recently developed a simple and novel interfacial trapping (IT) process to remove SWNT bundles. [ 34, 72] High yields of aqueous SWNT suspensions were prepared using the nonionic su rfactant gum Arabic. These SWNT suspensions were also comparable in dispersion qualit y to those prepared by ultracentrifugation. [ 72] Li and co-workers have used the IT process in fabrica ting field effect transistor (FET) devices. The ITbased FET devices showed at least an order of magnitude improvement in on/off ratios compared to those devices prepared by ultracentrifugation only. [ 42] 46

PAGE 47

The IT process for preparing aqueous SWNT su spensions is based on selective partitioning of SWNT bundles to liquid-li quid interfaces. The liquid-liquid system is mixed to form emulsions and increase interfacial area. As the SWNT bundles move to the interface, they help stabilize emulsions, known as Pickering emulsions, [ 21] which aid the separation. The driving force for the selective removal of SWNT bundles is the free en ergy changes associated with particles adsorbing at interfaces. [ 64, 72] Bundled SWNTs have larger radii and lengths, thereby taking up larger interfacial ar ea than individual SWNTs. [ 72] However, the properties of the SWNT surface also play an important role, whic h is reflected in the surface tension of the nanotubes or contact angle () of the system. The interfacial activity of particles (or emulsion stability) is related to the hydrophobicity /hydrophilicity of th e particles with = 90 yielding the strongest driving force for particles at interfaces. [ 26] In other words, highly hydrophobic or hydrophilic surfaces will not self-assemble at in terfaces. This poses problems for separating SWNT bundles coated with ionic surfactants. An ionic surfactants, such as sodium dodecyl sulfate (SDS), sodium dodecylbenzene sulfonate (SDBS), and sodium chol ate (SC), are preferred in SWNT applications [ 15, 71 ] because of the strong electrostatic repulsion, [ 73] yielding suspensions with high fluorescence intensity. In this chapter, we demonstrate that bundled SWNTs coated with th e anionic surfactant, sodium cholate (SC), can be removed from aqueous suspensions by IT. The dispersion quality of SC-suspended SWNT suspensions were improved by adjusting the initial pH of the suspension so that the surface charge on the SWNT-surf actant complex was a minimum. A two-step interfacial trapping process furt her improved the dispersion quality, providing final suspensions comparable to those produced by ultracentrifugation. 47

PAGE 48

3.1 Experimental Procedure 3.1.1 Dispersion and Interfacial Trapping SWNT suspensions were prepared with a give n initial mass (~ 8 mg) of raw SWNTs (Rice HPR 177.1) and mixed with 50 mL of an aqueous SC (Sigma-Aldrich) surfactant solution (1 wt.%) to form an initial concen tration of ~ 0.16 mg/mL. To aid dispersion, the solutions were homogenized using a high-shear IKA T-25 Ultra-Turrax mixer for about 1 1.5 h followed by ultrasonication using a Misonix S3000 for 10 min. A portion of the initial SWNT suspension was then ultracentrifuged at 20,000 rpm (Beckman Cou lter Optima L-80 K) for 5 h. The supernatant was carefully collected to rem ove the aggregated nanotubes. [ 9] The remaining portion of the SWNT suspension was added to to luene (Acros, 99%) at an oil to water volume ratio of 0.1 to form a two-phase system. In some cases, the pH of the aqueous suspension was adjusted by adding 1 M hydrogen chloride (HCl). The two-pha se system was then shaken vigorously by a vortex mixer for 30 s. Phase sepa ration started immediately but the suspensions were allowed to settle for 60 min to ensure steady state for sp ectroscopic analysis. The organic layer was then carefully removed to avoi d further emulsification. 3.1.2 Characterization Vis-NIR absorbance and NIR-fluorescence sp ectra were measured for each aqueous SWNT suspension using an Applied NanoFluor escence Nanospectrolyze r (Houston, TX) with excitation from 662 and 784 nm diode lasers. Raman Spectra were recorded by a Reinshaw Invia Bio Raman with excitation at 785 nm. The zeta-pot entials of the SWNT-surfactant complex were recorded at different pH with a Brookhaven ZetaPlus. 48

PAGE 49

3.2 Results and Discussion 3.2.1 Selective Separation of Bundles Measuring the dispersion quality of SWNT suspensions remains challenging. Spectroscopic techniques, includ ing absorbance, fluorescence, and Raman, are typically the best approaches [ 14] and were shown to be effective in evaluating the IT process based on gum Arabic (GA) suspensions. [ 72] Figure 3-1 shows spectroscopic analysis for the initial (homogenized and sonicated), ultracentrifuged, a nd IT SWNT suspensions. The Raman spectra (Fig. 3-1 (A)) show the initial suspension (curve 1) has a significan t aggregation peak, measured at ~270 cm -1 as expected for a suspension that still has a significant frac tion of bundled SWNTs. The absorbance of the initial suspension is also high (not shown). Although the fluorescence intensity of the initial SWNT suspension is relatively high (Fig. 3-1 (C)), it is clear that the dispersion quality is low once th e ratio of the fluorescence intensity to absorbance (F/A) is calculated (Fig. 3-1 (D)). [ 72] The ultracentrifuged SWNT suspension (curve 2) shows significant improvement in dispersion quality with a reduced aggregation peak (Fig. 3-1 (A)), low background absorbance (Fig. 3-1 (B)), and high F/A ratio (Fig. 3-1 (D)). Multi-step IT was previously shown to be e ffective at achieving high dispersion quality of SWNTs suspended with GA. [ 72 ] Similarly, the first IT process (one-step, curve 3) on SCSWNT suspensions showed improvement to the aggregation peak (Fig. 3-1 (A)) and absorbance (Fig. 3-1 (B)). The fluorescence intensity incr eased after the one-step IT (Fig. 3-1 (C)), similar to the previous observations w ith GA-suspended SWNTs. [ 72] This increase in intensity is attributed to a reduction in quenching once bundl es are removed from the suspension. Despite the high fluorescence intensity, th e dispersion quality remains lo wer than the ultracentrifuged SWNT suspensions (Fig. 3-1 (D)). Repeated IT steps (curves 4 5) continued to improve the dispersion quality, as seen by a lower aggregation peak and lower background absorbance. The 49

PAGE 50

significant decrease in background absorbance coupled with the rela tively constant fluorescence intensity suggests that the IT process is sel ective in removing bundled SWNTs from the aqueous suspension, as previously observed. [ 72] However, even after three consecutive IT steps, the suspension remains lower in dispersion quali ty than those suspensions prepared by ultracentrifugation. This indicates that the IT process is inefficien t at these conditions. 3.2.2 Improving Separation Effectiveness The charge on the SWNT-surfactant complex was adjusted to improve the efficiency of the IT separation process. Figure 3-2 (A) shows the zeta potential of the SWNTs in a SC suspension as a function of pH. The zeta po tential of the SWNTs shows a st rong pH dependency. The initial suspension (pH = 7.9) has a large negative char ge on the SWNTs in agreement with other reports. [ 74, 75] As the pH is adjusted to more acidi c conditions, the charge on the SWNTs goes through a sharp minimum centered at pH = 6.75. The pH-adjusted aqueous SWNT suspensions were then mixed with toluene usi ng the one-step IT process. Figure 3-2 (B) shows the F / A ratio for each SWNT suspension after IT. It is clear th at pH (surface charge) on the SWNTs has an important effect on the separation efficiency in the IT process. The most significant differences are observed for the smallest diameter SWNTs (l ower emission wavelength) where different pH could yield F / A ratios that could vary by a factor of two. The fluorescence spectra was deconvoluted into the individual ( n, m ) SWNT types. The normalized intensity of each ( n, m ) type was then plotted as a func tion of pH, as shown in Figure 3-2 (C) for the (6, 5), (8, 3), and (10, 5) SWNT types. It is clear th at the efficiency of the IT pro cess is highly correlated to the surface charge on the SWNTs. The suspensions with the best dispersion quality (highest F / A ratio) were obtained when the zeta potential was at a minimum. The dependence of dispersion quality or SWNT bundle removal efficiency on surface charge was confirmed through Rama n spectroscopy, as shown in Figure 3-2 (D). The 50

PAGE 51

aggregation ratio of the SWNT suspension (ratio of the intensity of (11, 3) SWNT to the aggregation peak) is presented in Figure 3-2 (E) as a function of pH. Once again, the best aggregation ratios match the minimum zeta-poten tial value at pH 6.75, which indicates the best dispersion quality in agreement with F / A ratio measurements. Prior studies have shown that mixing organic solvents with aqueous suspensions can alter the surfactant struct ure surrounding SWNTs. [ 76 ] This reorganization of surfactant can improve the fluorescence intensity by eliminating quenchi ng mechanisms. Before further comparisons are made between the pH-adjusted IT process and ultracentrifugation, solvent-induced changes to the spectra must be addressed. The ultracentrifuged SC-SWNT suspension was mixed with toluene and then evaporated for 24 h sim ilar to the IT processing steps. Figure 3-3 shows the fluorescence spectra of ultracentrifuged SC-SWNTs at each stage of the pr ocess. The consistent spectra throughout the process show the solvent has minimal influe nce on the surfac tant structure or the fluorescence intensity. The lack of spectral changes could be due to the strong interaction that SC has to the sidewall of the SWNTs. [ 77 ] Therefore, any changes to the spectra observed during IT can be attributed to the removal of bundled SWNTs. Figure 3-4 compares the spectral properties of SW NTs prepared by either the pH-adjusted two-step interfacial trapping proce ss (pH 6.75) or ultracentrifugation. The F / A ratio in Figure 3-4 (A) and 3-4 (B) show that SC-SWNT suspensions prepared by IT can achieve comparable dispersion quality to ultracentrifugation SWNT suspensions. The Raman spectra in Figure 3-4 (C) also show similar dispersion quality. However, the absorption spectra in Figure 4d show the concentration of individual SC-SWNTs in suspensions made by IT is almost double the concentration achieved by ultracentrifugation. This indicates that the two-step IT process has twice the yield of SWNTs in comparison to ultracentrifugation. To further illustrate the higher 51

PAGE 52

throughput, the IT suspension was diluted with 1 wt.% SC until the absorbance was nearly the same as the ultracentrifuged SWNT su spension. As can be seen in Figures 3-4 (A) 3-4 (D), the dilution has little effect on the di spersion quality. This dilution may also have added benefit to dispersion stability by adjusting the surface charge (pH) to higher values (see Fig. 3-2). Table 3-1 summarizes the dispersion qua lity measurements for the two-step IT process (pH adjusted) compared to ultracen trifugation. The table shows that the IT process has similar dispersion quality to ultracentrifugation, as m easured by multiple characterization methods. Table 3-2 summarizes the processing parameters and results for the two-step IT and ultracentrifugation approaches to preparing SC-SWN T suspensions. As seen in the table, the IT process has favorable processing times, yields, and throughput. These ch aracteristics and the simplicity of the process enable a fast and scal able approach to preparing SWNT suspensions with good dispersion quality. In addition, the IT process is more conducive to parallel processing, which will further improve throughput. 52

PAGE 53

0 0.1 0.2 0.3 0.4 0.5 160200240280320 (1) Control (2) Ultracentrifuged (3) One-step I. T. (4) Two-step I. T. (5) Three-step I. T.Normalized IntensityRaman Shift (cm-1) (a) 2 1 4 3 5 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 60080010001200AbsorbanceWavelength (nm) (b) 2 3 4 5 0 0.002 0.004 0.006 0.008 0.01 0.012 900100011001200130014001500IntensityWavelength (nm) (c) 3 4 5 2 1 0 0.01 0.02 0.03 0.04 900100011001200130014001500Intensity / AbsorbanceWavelength (nm) (d) 2 1 5 4 3 C A D B Figure 3-1. A) Radial breathing m ode (RBM) of Raman spectra, B) Vis-NIR absorption spectra, C) NIR-fluorescence spectra (662 nm excitation), and D) Fluorescence / absorbance ratio (calculated from 662 nm) of SC-suspe nded SWNTs. (1), (2), (3), (4) and (5) represent the control, ultracentrifuged, one-step interfacial trapped, two-step interfacial trapped, and three-step inte rfacial trapped SC-SWNTs, respectively. 53

PAGE 54

0 0.1 0.2 0.3 0.4 0.5 0.6 150200250300Normalized IntensityRaman Shift (cm-1) 1 2 3 4 5 6 0 0.01 0.02 0.03 0.04 900100011001200130014001500 (1) Initial (2) pH 7.09 (3) pH 6.74 (4) pH 6.42 (5) pH 6.24 (6) UltracentrifugedIntensity / AbsorbanceWavelength (nm) 6 5 1 3 2 4 -55 -50 -45 66.577.58zeta-Potential (mV)pH 0 0.2 0.4 0.6 0.8 1Normalized Intensity(8, 3) (10, 5) (6, 5) 2.5 3 3.5 4 4.5 5 5.5 66.577.58Aggregation RatiopH A B C D E Figure 3-2. A) The pH-dependent changes to the zeta-potential of SC-SWNTs. B) The effect of pH on the intensity/absorbance ratio after interfacial tra pping. The surface charge was adjusted by changing the pH of the suspensi ons prior to IT. C) The effect of surface charge on the intensity of (6,5), (8,3), and (10,5) SWNT types after interfacial trapping. All intensities are normalized to the maximum intensity. D) pH-dependent Raman spectra (784 nm excitation) of the radial breathing modes (RBMs) of SWNTs. Spectra are normalized to the G band. E) Th e effect of surface charge on the Raman aggregation ratio. 54

PAGE 55

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 900100011001200130014001500 (1) Ultracentrifuged (2) Toluene mixed (3) EvaporatedIntensityWavelength (nm) 1 2 3 Figure 3-3. The effect of solvent on the fluores cence intensity of ultracentrifuged SC-SWNTs (662 nm excitation). The ultracentrifuged SWNT s (1) are first mixed with toluene (2). The toluene was then removed by evaporation (3). 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 160180200220240260280300320Normalized IntensityRaman Shift (cm-1) 0 0.01 0.02 0.03 0.04 0.05 900100011001200130014001500 (1) Two-step IT (pH 6.75) (2) Ultracentrifuged (3) Diluted ITIntensity / Absorbance (662 nm)Wavelength (nm) 1 2 3 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 400600800100012001400AbsorbanceWavelength (nm) 1 2 3 0 0.002 0.004 0.006 0.008 0.01 900100011001200130014001500Intensity / Absorbance (784 nm)Wavelength (nm) 1 2 3 D B C A Figure 3-4. Comparison of SC-SWNT suspensions prepared by the pH-adjusted (pH 6.75) twostep IT process (1), ultracentrifugation (2), and dilution of the IT process (3) obtained from (1). The ratio of fluorescence intensity to absorbance for SWNT suspensions excited at A) 662 and B) 784 nm. C) Raman spectra of the SWNT RBMs. D) Absorption spectra of the suspensions. 55

PAGE 56

Table 3-1. Dispersion Quality Comparison of Aqueous SWNT Suspensions. aqueous suspension anaylsis control ultracentrifugation two-step interfacial trapping fluorescence intensity weak strong strong absorbance features broad peaks blue-shifted and resolved peaks blue-shifted and resolved peaks resonant ratio 1 0.015 0.070 0.057 Raman aggregation ratio 2 1.00 6.25 7.22 F/ A Ratio 3 0.0014 0.0358 0.0352 1 Calculation based on ratio of resonant and nonresonant features of the absorbance peak near 660 nm. 2 Calculation based on intensity from (11, 3) nanotubes and aggregation peak (~270 cm -1 ). 3 Calculation based on fluorescence intensity from (7, 6) nanotubes and absorbance at 662 nm. Table 3-2. Comparison of the processing parameters for the pH-adjusted two-step interfacial trapping (pH 6.75) and ultracentrifugation methods for producing high quality SWNT suspensions. Method Processing Time SWNT Yield Throughput (per liter) Interfacial Trapping < 2 h Large (4~24%) 3 ~ 10 mg/h Ultracentrifugation > 5 h Small (< 2%) < 1 mg/h 56

PAGE 57

CHAPTER 4 SWELLING THE MICELLE CORE SURROUND ING SWNTS WITH WATER-IMMISCIBLE ORGANIC SOLVENTS Single-walled carbon nanotubes (SWNTs) are tubul ar structures of carbon with several unique physical and chemical properties. [ 4, 5] Dispersing SWNTs in a solvent is required for many applications but common organic solvents o ffer insufficient solvation forces to suspend SWNTs. [ 36 ] Surfactants or polymers are often used to stabilize aqueous SWNT suspensions. The hydrophobic part of the surfactant non-covalen tly attaches to the sidewall of the nanotubes while the hydrophilic end extends into the water phase. The resulting surf actant shell creates a repulsive barrier that overcomes the strong van der Waals attractive forces needed to disperse the SWNTs. [ 8 ] The anionic surfactants sodium dodecyl sulfate (SDS) and s odium dodecylbenzene sulfonate (SDBS) are frequently used because of the high dispersion quality of the resulting SWNT suspensions. The ability to individually disperse SWNTs in solution enabled the discovery of NIR fluorescence of semiconducting nanotubes, [ 9] which correspond to the specific ( n, m ) type of each nanotube as mentioned in previous paragraph. The fluorescence emission energies of SWNTs are sensitive to the su rrounding environment with suspended SWNTs being red-shifted in comparison to SWNTs in air. [ 78-80] The extent of the energy shift is dependent on the surfactant, [ 15] protein, [66, 67, 81, 82] or polymer that encases the nanotube. [ 83, 84] The first measurements of the photoluminescence quantum yields for dispersed SWNTs were on the order of 0.1% or less. [ 9, 10] The optically excited electronic st ates of SWNTs are highly mobile, [ 85] making them sensitive to extrinsic effects that can reduce the quantum yield, including sidewall defects, [ 57, 80, 85-87] protonation, [ 85, 88-90] surfactant inhomogeneties, [ 57, 86, 87, 91] bundles, [ 57-59, 92] and nanotube ends (i.e., lengths). [ 93, 94] Indeed, single-molecule spectroscopy in both aqueous solutions [ 86] and air [ 80] have observed spots along the length of 57

PAGE 58

a nanotube, which have reduced fluorescence em ission intensity. These dark spots allowed Cognet et al. [ 85] to estimate the exciton diffusion length to be ~90 nm. When the quantum yields are calculated on single na notubes that have no dark spots, the quantum yields improve dramatically to values approaching 10%.[ 80, 85, 86] The observation of these extrinsic factors has raised questions regarding the capability of fluorescence spectroscopy to quantify relative ( n, m ) ratios. [ 80, 86] The difference in quantum yield measurements seen for bulk-scale suspensions and single-molecule analysis points to the importance in understanding how SWNT disper sion affects fluorescence intensities. For example, different quantum yields ar e observed for different surfactants. [ 91] Differences in fluorescence intensity between diffe rent surfactant suspensions al so led to the conclusion that some surfactants, such as SDS, preferably disperse small diameter SWNTs. [ 78] The ultimate objective is to learn how to comp ensate for these differences. [ 86] Recently, the elimination of small bundles was shown to improve the quantum yields of bulk SWNT suspensions to ~1 1.5%. [ 92] Understanding the importance of these extrinsic factors to th e quantum yields of SWNTs may be compounded by an incomplete understandi ng of the surfactant structure surrounding the nanotube. In general, three models of surfactant structure have been suggested: (a) hemisphere, [ 17] (b) cylindrical, [ 95] and (c) randomly adsorbed su rfactant structures surrounding the nanotubes. [ 18] However, there is no accepted stru cture for the surfactant surrounding the nanotubes. Wallace et al. [ 20] showed through coarse-grained molecular dynamics (MD) calculations that the surfactant structure around SWNTs was con centration dependent. At low concentrations, the surfactant molecules tend to orient themselves along the length of the nanotube. However, the surfactant extends out into the aqueous phase at higher concentrations. 58

PAGE 59

These simulations combined with the previous ly observed extrinsic factors suggest that processing conditions may play an important ro le in the photophysical properties of SWNT suspensions. In this chapter, we describe the swelling of the hydrophobic core of the micelle surrounding the SWNTs. Mixing aqueous suspensi ons of SWNTs with various immiscible organic solvents results in solvatochromic shifts indicat ive of a coating surrounding the nanotubes. The solvatochromic shifts and change s in the fluorescence intensity are dependent on the surfactant coating the nanot ube, allowing the spectral change s to be used to probe the surfactant structure around SWNTs. While the so lvatochromic shifts are reversible once the solvent is removed, the fluorescence intensity depends on the surfactant-solvent combination used to swell the micelle. For SDS-coated SWNTs, the largest di ameter nanotubes show significant increases in fluorescence intensity after some organic solvents are removed. In addition, these SWNTs are better protected from the fluorescence quenching effect of acid. The spectral changes suggest that th e organic solvent is capable of reorganizing the surfactant surrounding the SWNT. 4.1 Experimental Procedure 4.1.1 Dispersion SWNT suspensions (Rice HPR 145.1) were prepared in both 1 wt% SDS and SDBS solutions followed by ultracentrifugation. Aqueous nanotube suspensions were prepared by mixing 20 mg of raw SWNTs with 200 mL of an aqueous SDS or SDBS surfactant solution (1 wt. %). High-shear homogenization (IKA T-25 U ltra-Turrax) for 1.5 2.0 h and ultrasonication (Misonix S3000) for 10 min were used to aid di spersion. After ultrasonic ation, the mixture was ultracentrifuged at 20,000 rpm for 5 h (Beckman Coulter Optima L-80 K). Immiscible solvents, such as o-dichlorobenzene (ODCB), were a dded to each SWNT suspension (solvent: water 59

PAGE 60

volume ratio of 0.5) and mixed. The mixture was shaken vigorously for 30 s with a vortex stirrer. After shaking, a white emulsion phase imme diately started to phase separate. The characterization of aqueous SWNT suspensions was done after waiting for 1.5 2 h to reach steady state. The excess organic solvent was th en carefully removed fr om the aqueous SWNT suspension to prevent further emulsification. SW NT suspensions prepared in pure ODCB were ultrasonicated and then filtered thro ugh coarse filter paper to rem ove the large visible aggregates in the suspension prior to characterization. [ 9] 4.1.2 Characterization The aqueous phase was characterized by vis-NI R absorbance and NIR-fluorescence spectra using an Applied NanoFluorescence Nanospectro lyzer (Houston, TX) with excitation from 662 and 784 nm diode lasers. Raman spectra were reco rded with a Renishaw Invia Bio Raman with excitation from a 785 nm diode laser. 4.2 Results Figure 4-1 shows the NIR fluorescence spectra for the initial aqueous SWNT suspension and the aqueous suspension after mixing with ODCB. In both the SDBSand SDS-SWNT suspensions, the peaks have red-shifted by 0.01 0.1 eV (9 17 nm) indicating a change to the environment surrounding the nanotubes. The fluorescence intensity has also significantly decreased in both suspensions. In SDS suspensi ons, the fluorescence emission from the largest diameter SWNTs has almost completely disappeared. Nanotubes have limited solubility in pure ODCB without us ing surfactants, [ 36] allowing direct comparison to aqueous susp ensions mixed with ODCB. Figure 4-2 compares the normalized fluorescence spectra of ODCB-suspe nded nanotubes to aqueous SDSand SDBSsuspended nanotubes after mixing with ODCB. Once the SDBSand SDS-coated aqueous SWNT suspensions are mixed with ODCB, the su spensions show similar emission spectra to 60

PAGE 61

SWNTs dispersed in pure ODCB. The smallest diameter nanotubes (larger emission energy) have very good agreement in peak position wi th nanotubes in ODCB, especially for SDBScoated SWNTs. The similarity in the spectra implies that ODCB molecules are in close proximity to the nanotubes. The organic solvent was removed from the su rfactant-coated SWNT suspensions by room temperature evaporation for 24 h. Figure 4-3 (A) shows the fluorescence spectra for SDS-coated SWNT suspensions during evaporation of ODCB. Th e time dependent changes to the intensity of the (7, 6) and (8, 3) nano tubes are shown in Figure 4-3 (B). The smallest diameter nanotubes (e.g., (8, 3) SWNTs) recover their fluorescence in tensity faster than the larger diameter nanotubes (e.g., (7, 6) SWNTs). All nanotube fluorescence intensities have reached equilibrium after ~20 h. The spectra for both SDSand SDBS-suspended SWNTs return to their original peak positions and have in tensities close to their initial values (compare Fig. 4-1 (B) for SDSsuspended SWNTs, discussed further below). Spectral changes were also observed for a queous SWNT suspensions mixed with other organic solvents. Figure 4-4 plots the fluorescence spectra of SDBSand SDS-suspended SWNTs mixed with various organic solvents. Once again, the solvatochromic shifts and changes to fluorescence emission intensity were dependent on the surfactant used to suspend the SWNTs. For each solvent, the SDBS-suspended SWNTs in Figures 4-4 (A) and 4-4 (C) showed similar changes to both the solvatochromic shif ts and fluorescence intensity for each ( n, m ) SWNT type. SDBS-SWNT mixtures with hexane showed a slight increase in fluorescence emission and blueshifts. On the other hand, mixtures of benzen e with SDBS-SWNTs showed red-shifts and both slight increases and decreases in fluorescence intensity. Finally, chloro form mixtures with SDBS-coated SWNTs had similar decreases in inte nsity and red-shifts as the ODCB mixtures. In 61

PAGE 62

contrast to the SDBS-coated SWNT su spensions, the SDS-SWNTs in Figures 4-4 (B) and 4-4 (D) showed changes to the fluorescence inte nsity, which was dependent on the SWNT ( n, m ) type. The emission spectra for SDS-SWNTs red-shif t and significantly decrea se in intensity for hexane mixtures while benzene mixtures have small blue-shifts of the emission and significant increases in fluorescence intensity. The respons e of SDS-SWNTs mixed with chloroform was diameter dependent and varied slightly between different suspensions. In general, the small diameter nanotubes (e.g., (8, 3) SWNTs) have minor changes to both emission energy and intensity while large diameter nanotubes show no sh ifts but large decreases in intensity. Similar to mixtures with ODCB, the intensity of the largest diameters (e.g., (9, 5) SWNTs) almost completely disappeared. Table 4-1 summarizes the fluorescence emission of SWNT suspensions after being mixed with each solvent. In general, a ll systems that red-shift show d ecreases in fluorescence intensity. Figure 4-4 also shows that the largest diameter SW NTs have the most significant decreases in fluorescence emission intensity. The spectral shif ts for SDBS-coated SWNTs closely mirror the differences in dielectric constant and polarity of the organic solvent. However, there is no apparent trend of the spectral changes of the SD S-coated SWNT suspensions with the solvent. The largest decreases in fluorescence intensity occu rred when mixed with the solvent that has the highest dielectric constant (ODCB) and the lowest dielectric consta nt (hexane). The polarity also seems to have no influence on the fluorescence inte nsity since the two nonpolar species showed the ability to either decrease (hexane) or incr ease (benzene) the intensity after mixing with the aqueous SWNT suspensions. The spectra for both SDBSand SDS-suspended SWNTs exposed to immiscible solvents return to their original peak positions after the solvent is evaporated fr om the aqueous suspension 62

PAGE 63

and have intensities close to the surfactant only SWNT susp ensions, as shown in Figure 4-5 This is somewhat surprising for SDS-suspended nanotube s since the peaks had n early disappeared in Figures 4-4 (B) and 4-4 (D) after mixing with ODCB, chloro form, and hexane. For example, the fluorescence spectra for SWNTs mixed with ODCB in Figures 4-4 (B) and 4-4 (D) shows the fluorescence of only the (8, 3), (7, 5) and (6, 5) SWNT types. After the evaporation of ODCB, the fluorescence emission from all SWNT ( n, m ) types return, as shown in Figures 4-5 (B) and 45 (D). There are some minor differences in th e intensity of SDBS-coated SWNTs after exposure to the organic solvents, which may be attributed to removal of nanotubes or impurities at the interface. [ 34, 72] On the other hand, substantial increase s to the fluorescence intensity for SDSsuspended SWNTs are observed after ODCB and chlo roform have evaporated. For example, the large diameter SWNTs in Figures 4-5 (B) and 4-5 (D) have considerable increases in fluorescence emission intensity after the organic solvent is removed. These changes in fluorescence in tensity occur without any diff erences to the relative ratios between ( n, m ) types in either the absorbance or Raman spectra. The absorbance spectra after the solvent is removed from the aqueous suspension in Figure 4-6 (A) show that, in general, the spectra are similar to the initial SWNT su spension. Although the SDS-SWNT suspensions exposed to ODCB and chloroform have minor ch anges to the absorbance baseline, there are no changes to the ratio of peak heights. The d ecrease in absorbance observed with chloroform would indicate further proof that impurities were removed from the system. [ 34, 72 ] Most importantly, there is no broadening or red-shifting of the peaks for any of the spectra. The radial breathing modes (RBMs) of the Raman spectra ar e dependent on the diameter of the nanotubes and can also indicate changes to the ( n, m ) types. As seen in Figure 4-6 (B), the intensity of the Raman RBMs for each peak is nearly identical to that observed for the initial SDS-SWNT 63

PAGE 64

suspension. Note that the (10, 5) SWNT has the largest intensity increase in Figure 4-5 d. In addition, there are no changes to the so-called aggregation peak (~270 cm -1 ) after the organic solvents are removed. These results indicate that the suspension has no significant changes to the relative concentration of nanotube ( n m ) types or bundling, which could cause the changes in fluorescence emission seen in Figures 4-5 (B) and 4-5 (D). [ 34, 59, 86, 91, 96] 4.3 Discussion 4.3.1 Swelling Effects The solvatochromic shift in fluorescence em ission spectra of SWNTs mixed with organic solvents indicates a change to the environment surrounding the nanotubes. When compared to SWNTs suspended in pure organic solvents, these spectra show nearly identical peak positions (see Figure 4-2 ) suggesting that organic solvents ar e forming a layer or a shell around the sidewall. Since the fluorescence, absorbance, and Ra man spectra return to the initial values after the solvent has evaporated, the surfactant must still be present. Othe rwise, there would be substantial bundling of the nanot ubes resulting in broadened and red-shifted absorbance as well as quenching of the fluorescence emission. [ 9, 86, 91 ] The reversibility of both the spectral shifts and intensities is good evidence that the presence of the organic solvent molecules rather than aggregation or surfactant removal is responsible for the observed changes. Therefore, the data suggests that the organic solvent is swelli ng the hydrophobic region of the surfactant micelle surrounding the SWNTs, similar to the swelling of micelle cores in block copolymers, [ 97] as shown in Figure 4-7 (A). This conclusion is also supported by the consistent trend of the solvatochromic shift of SDBS-coated SWNT su spensions with solvent polarity (see Figures 4-4 (A), 4-4 (C) and Table 4-1), indicating that the envi ronment around the nanotubes is systematically changing when mi xed with different solvents. 64

PAGE 65

The spectral changes observed for SDS-coat ed SWNT suspensions when mixed with organic solvents are more complex. In hexane mixtures, the fluorescence intensity of SWNTs increases significantly in Figures 4-4 (B) and 4-4 (D). On the other hand, fluorescence emission decreases considerably for other solvents. The pr esence of the organic solvents can affect the fluorescence intensity; [ 98] however, the solvatochromic shifts and intensity changes should tend to follow polarity changes as observed with SDBS-SWNT suspensions. The effect of each solvent should also yield consistent changes to the SWNT emission spectra. In other words, if a solvent (e.g, ODCB) is inducing change s to the emission spectra in Figures 4-4 (B) and 4-4 (D), similar changes to the emission sp ectra should be observed in the spectra in Figures 4a and 4c. The lack of a consistent trend for SDS-coated SWNTs (see Figure 4-4 (B), 4-4 (D) and Table 41) and the different response for each surfactant suggests that another phenomenon is responsible for the changes in SDS-coated SWNT suspensions. Researchers have observed increased fluorescence intensity due to Frster energy transfer between individual SWNTs. [ 57-59, 79 ] For example, the emission of large diameter SWNTs (smaller band gap energy) was enhanced dur ing nanotube bundling by excitation and energy transfer from the larger band gap, smaller diameter SWNTs. [ 58] If energy transfer is responsible for the increased fluorescence emission of large di ameter SWNTs after solv ent evaporation, then a decrease in fluorescence intensity would be expected for the smaller diameter SWNTs. However, the increases in nanotube emission intensity observed afte r chloroform and ODCB exposure occur without a simultaneous decrease in intensity from other SWNTs. The lack of any significant changes to the Raman ag gregation peak also seems to ru le out Frster energy transfer as the cause for increased fluorescence emission of some ( n, m ) types. The fact that a new peak corresponding to the (9, 7) SWNT appear s in the emission spectra in Figure 4-5 (D) strongly 65

PAGE 66

suggests that this species is pres ent in the initial susp ension but is not fluorescing. This behavior often is indicative of exciton energy transfer [ 57-59, 79 ] but this SWNT type is already observed in the spectra for SDBS-suspended SWNTs in Figure 4-5 (C) and the Raman spectra in Figure 46 (B). In fact, the spectra in Figures 4-5 (B) and 4-5 (D) after exposure to chloroform are strikingly similar in shape and relative intensity wh en compared to all of the spectra in Figures 45 (A) and 4-5 (C). SWNTs are also susceptible to doping when the solvent is in close proximity to the surface. If solvent doping were responsible for the observed intensity changes after evaporation, then both SDSand SDBS-coated SWNTs should have similar changes. As shown in Figure 4-1 both systems have a solvatochromic shift that indicates that ODCB is surrounding the SWNTs. However, only the SDS-SWNT suspensions show improvements to the fluorescence intensity in Figure 4-5 In addition, the nanotube ( n, m ) types which show decrease s in intensity when the solvent is present are the same na notube types that have increased emission after exposure to the solvents. For example, the fluorescence emission from the (7, 6), (12, 1), (11, 3), and (10, 5) SDS-SWNT suspensions in Figures 4-4 (B) and 4-4 (D) are completely quenched but then show the largest increases in intensit y after evaporation in Figures 4-5 (B) and 4-5 (D). Without changes to the chemical structure of the solven t during evaporation, it is unlikely that solvent doping can cause this behavior. Recently, resear chers showed that decomposition products of ODCB generated during ultrasoni cation could dope SWNTs. [ 99] However, the authors observed that mixing with solvents, similar to the approa ch in this study, does not cause decomposition or doping. To determine if photochemistry could initiate ODCB decomposition and doping during evaporation, SWNTs were also processed under natural light and dark conditions. The fluorescence spectra of the light and dark suspen sions were identical (not shown) eliminating 66

PAGE 67

decomposition-induced changes. Finally, Raman sp ectra after solvent evaporation show the G and G band do not have any shifts or width changes associated with doping. [ 99, 100 ] The G band in Figure 4-6 (C) is partially obscured by the background fluores cence but clearly shows that there are no shifts after solvent evaporation. [ 99, 100] Similarly, the G band in Figure 4-6 d has no major changes to the shifts or width after evaporation. It does appear that ODCB may have a slight red-shift (~1 cm -1 ); however, this would indicate a small amount of n-type doping rather than p-type doping observed previously. [ 99 ] The lack of significant changes to the G, G and RBM modes of the Raman spectra suggest that solvent doping is no t responsible for the changes to fluorescence emission intensity observed in Figures 4-5 (B) and 4-5 (D). [ 99-101] One plausible explanation is that the fluores cence of the large diameter SWNTs in SDSsuspended SWNTs was quenched in the initial susp ension. It is generally accepted that exposure of the nanotube to protons in the aq ueous phase quenches the exciton. [ 85, 88-91 ] The surfactant shell surrounding the nanotube, therefore, provides the need ed protection to prevent nonradiative recombination of the exciton. Okazaki et al. [ 78] observed different ratios of the fluorescence emission intensities of SWNTs suspende d in SDS to SWNTs in air leading them to conclude that SDS did not susp end larger diameter nanotubes. This conclusion is supported by the fact that smaller diameter SWNTs ha ve stronger binding energies to SDS. [ 102 ] Researchers have observed that larger diameter SWNTs co ated with SDS have a higher sensitivity to quenching. [ 91] Indeed, the initial SDS-suspensions have lower fluorescence intensities for larger diameter SWNTs in comparison to the initial SDBS-suspensions, as seen in Figures 4-4 (C) and 4-4 (D). Therefore, these inte nsity differences could mean that the largest diameter nanotubes have either (1) inadequa te surfactant shells to aid disp ersion and are truly not present 67

PAGE 68

or (2) incomplete or non-uniform surfactant stru ctures surrounding the nanotubes, which quench the emission fluorescence. 4.3.2 Possible Quenching Mechanism Wallace and Sansom [ 20 ] showed through MD simulations that surfactant coverage was non-uniform and the thickness vari ed along the length of the nanotube. In other words, the surfactant permeability to protons from the water could potentially be altered as the surfactant concentration is changed, affecting fluorescence intensity. Others have also speculated similar holes in the surfactant layer to he lp explain fluorescence behavior. [ 89, 91] If the permeability of the surfactant structure were altered, then these potential quenching spots could be removed, resulting in increased emission intensity. These al terations could occur by introducing solvents to protect the nanotube or by reor ganization or redistribution of the surfactant surrounding SWNTs. These changes to the SDS surfactant structure shoul d be most sensitive to the largest diameter SWNTs. To test if solvents can impr ove the ability of the surfactan t to protect the SWNT from quenching protons, the pH of the aqueous phase was adjusted to control the concentration of protons (fluorescence quenchers) while the organic solvent was still present and after the solvent is evaporated. As shown in Figures 4-7 (B) and 4-7 (C), the fluorescence intensity from the (7, 6) and (10, 5) SWNT types in a pure SDS-SWNT su spension decrease by two orders of magnitude at acidic pH similar to previous observations. [89, 91 ] However, the aqueous SWNT suspensions with the organic solvents forming a shell around the nanotubes have different trends from the pure SDS-SWNTs. The aqueous suspension mixe d with benzene shows that the SWNTs are more resistant to pH changes. This resistance suggests that benzene offers another protective layer around the nanotube, providi ng further support for a swelle d micelle state surrounding the SWNTs. On the other hand, suspensions mixed w ith ODCB showed more susceptibility to pH 68

PAGE 69

quenching effects across the entire pH range. Only at pH values above 10, does the fluorescence intensity start to recover. These results suggest that the surfactant stru cture has become more permeable to quenching by protons or water, [ 89, 91 ] indicating that the nanotube surfactant structure has been altered or reorganized. Once ag ain, the trends of all SWNT suspensions after solvent evaporation are similar to th at of pure SDS as shown in Figures 4-7 (D) and 4-7 (E), which indicates that the surfactan t structure has been recovered to a state similar to the initial state. However, there are some noticeable diffe rences in the pH response. For the (7, 6) nanotube, ODCB provides some improvement to quenching in pH values near neutrality while benzene also shows improvements with slightly be tter fluorescence intensity at acidic pH values. In contrast, the larger diameter (10, 5) nanotube in Figure 4-7 (E) shows significant improvement in fluorescence emission intensity. SDS-SWNT suspensions exposed to benzene show significant improvement at pH values between 5 and 10. ODCB exposed nanotube suspensions show even better protection to the quenching e ffects of acid. This organic solvent improves the fluorescence intensity across the entire pH range but most notably for the acidic pH values. The physisorption of surfactant molecule s on nanotubes is a dynamic process, continuously being interchanged with free surfacta nt in solution. The intr oduction of an organic phase may alter this process. Indeed, surfactan t molecules have slower desorption kinetics in emulsions than in micelles. [ 103] The organic phase may also enable increased mobility of the surfactant around the nanotubes. Depe nding on the interactions and thei r strengths, the surfactant reorganization may result in benefi cial or detrimental effects to the fluorescence emission. These differences in surfactant-solvent-SWNT interactions helps explain the difference in spectra seen for SDS-suspensions in Figure 4-4 While benzene appears to swell the micelle, the other solvents open holes, enabling quenching from the aqueous phase. Upon evaporation of the 69

PAGE 70

solvent, the surfactant structur e can also be altered, which can improve the resistance to quenching and the fluorescence emission intensity. Table 4-2 summarizes these increases in intensity, which is also directly related to in creases in their quantum yield since absorbance changes are relatively minor. The in tensity increases are most significant for the largest diameter nanotubes. Chloroform tends to improve the fl uorescence emission intensity more than ODCB. The largest emission increases are observed for the (11, 3), (10, 5), and (9, 7) SWNT types, which have a 107.1%, 179.3%, and 177.6% in crease in intensity, respectively. 70

PAGE 71

A B Figure 4-1. Comparison of NIR fluorescence spectra (ex = 662 nm) of surfactant-coated SWNT suspensions before ( 1 and 3) and after mixing with ODCB ( 2 and 4). A) SDBSsuspended SWNTs and B) SDS-suspended SWNTs. A B Figure 4-2. Comparison of norma lized NIR fluorescence spectra (ex = 662 nm) of surfactantcoated SWNT suspensions to SWNTs dispersed in only ODCB ( 1). A) SDBSSWNTs ( 2) and B) SDS-SWNTs ( 3) after mixing with ODCB. 71

PAGE 72

A B Figure 4-3. A) NIR fluorescence spectra (ex = 662 nm) of SDS-coated SWNT suspensions mixed with ODCB during solvent evaporation. B) Time-dependent recovery of (7, 6) and (8, 3) peak intensities of SDS-SWNTs mixed with ODCB. 72

PAGE 73

A C D B Figure 4-4. NIR fluorescence spectra of SDBSa nd SDS-coated SWNT suspensions after being mixed with the immiscible organic solvents: hexane ( 2), benzene (3), ODCB ( 4), or chloroform (5). The initial surfactant-SWNT suspension ( 1 ) is plotted for comparison. Excitation from 662 and 784 nm lasers. 73

PAGE 74

A C B D Figure 4-5. NIR fluorescence spectra of SDBSand SDS-coated SWNT suspensions after evaporation of the immiscible organic solvents: hexane ( 2 ), benzene (3 ), ODCB ( 4), or chloroform ( 5). The surfactant-SWNT suspension exposed to air for 24 h ( 1) is plotted for comparison. Note that both SDSand SDBS-coated SWNT suspensions showed minor changes to the emission sp ectra after being op en to air for 24 h. Excitation from 662 and 784 nm lasers. 74

PAGE 75

A C B D Figure 4-6. A) Absorbance and (bd) Raman spectra (ex = 785 nm) of SDS-coated SWNT suspensions after evaporation of the organic solvent: hexane ( 2), benzene ( 3), ODCB ( 4), or chloroform (5). The surfactant-SWNT suspension exposed to air for 24 h ( 1) is plotted for comparison. Th e radial breathing modes (RBMs) for specific ( n, m ) SWNT types are shown in B) while C) a nd D) show the Raman spectrum of the G and G band of the SDS-coated SWNTs susp ensions, respectively. The dashed lines indicate the positions of the surfactant suspensions prior to mixing with organic solvents. All spectra in C) and D) are offset for clarity. 75

PAGE 76

10-410-310-210-1 (1) SDS only (2) benzene (3) ODCB4681012Intensity of (7,6) SWNTpH 1 2 3 10-410-310-210-14681012Intensity of (7,6) SWNTpH 1 2 3 10-410-310-24681012Intensity of (10,5) SWNTpH 1 2 3 10-410-310-24681012Intensity of (10,5) SWNTpH 1 2 3Mixed with organic solventOrganic Solvent Evaporated(b) (c) (d) (e) D B A C E Figure 4-7. A) Swelling of the hydrophobic core of the micelle surrounding SWNTs. B)-E) The effect of pH on the fluorescence intensity of the (7, 6) SWNT (ex = 662 nm) and (10, 5) SWNT (ex = 784 nm) in SDS-suspensions after being mixed with the immiscible organic solvents benzene and ODCB and after evaporation of the solvent. 76

PAGE 77

Table 4-1. Spectral changes of surfactant-co ated SWNTs mixed with organic solvents. SDBS-SWNTs SDS-SWNTs Organic Solvent Dielectric Constant a Dipole Moment (D) a Fluorescence Intensity Change Solvatochromic Shift Change Fluorescence Intensity Change Solvatochromic Shift Change Hexane 1.89 0 small increase blue large decrease red Benzene 2.28 0 varied red large increase slight blue Chloroform 4.81 1.04 decrease red varied varied ODCB 10.12 2.50 decrease red large decrease red water 80.1 1.85 a Values taken from the CRC Handbook of Chemistry and Physics Table 4-2. Fluorescence intensity changes for various SWNT (n, m ) types after exposure to ODCB or chloroform. Intensity Increase (%) SWNT Diameter (nm) ODCB chloroform (7,6) a 0.895 8.3 6.0 (9,4) b 0.916 20.6 10.1 (8,6) a 0.966 24.9 42.4 (9,5) a 0.976 27.9 57.5 (12,1) b 0.995 46.0 75.9 (11,3) b 1.014 56.8 107.1 (10,5) b 1.050 72.4 179.3 (9,7) b 1.103 34.6 177.6 a Excitation from 662 nm laser b Excitation from 784 nm laser 77

PAGE 78

CHAPTER 5 SOLVATOCHROMIC EFFECTS OF SWNTS IN NONPOLAR MICROENVIRONMENTS Single-walled carbon nanotubes (SWNTs) have unique properties suitable for a wide variety of applications. [ 4, 5] However, characterizing the pr operties of SWNTs has been difficult because of the ensemble of nanotubes present in a suspension. [ 104] Further complicating the characterization of SWNT prope rties is the inability to suspend SWNTs in known dielectric environments, such as organic solvents. [ 36] Dispersing SWNTs with the aid of surfactants has significantly advanced SWNT characterization. For example, the unique photoluminescence (PL) emission of se miconducting SWNTs can identify the ( n, m ) types in a suspension. [ 2] However, an incomplete understanding of the surface coverage and structure of surfactants surrounding SWNTs makes it difficult to assess the dielectric constant of the media surrounding SWNTs. [ 105 ] The unknown environment surrounding SWNTs poses problems to characterizing and separating SWNTs through di electrophoresis, and using SWNTs in sensing applications. [ 27, 76, 89, 91] The polarizability of individual SWNTs is impor tant to many applications as well as the photophysical and photochemical response of SWNTs. The polarizability is highly anisotropic with the longitudinal component ( || ) at least an order of magnitude larger than the transverse component ( ). Therefore, the longitudinal polarizabil ity is the dominant contribution and the dipole moment is oriented along the nanotube ax is. Several models have been developed to calculate the longitudinal polarizability. These si mulations suggest the longitudinal polarizability is a function of the nanotube radius and ba nd gap energy of SWNTs but disagree on the relationship. Researchers have f ound that the longitudinal polarizab ility is proportional to radius over band gap squared, [ 106] inverse band gap, [ 107] inverse band gap squared, [ 108] and radius squared. [ 109] 78

PAGE 79

While there are several theoretical studies on polarizability, there are few experimental studies. [ 110, 111 ] Measuring the photophysical properties of molecules, such as solvatochromic shifts, can provide information about the exc ited states of the molecules and even their polarizability. Choi and Strano used solvatochr omic shifts of SWNTs to determine that longitudinal polarizability varies with inverse nanotube radius times inverse band gap squared. [ 110] However, these measurements were based on relatively few systems including surfactantsuspended SWNTs, which have unknown diel ectric environments around nanotubes. The theoretical and experimental studies descri bed above show that the exact form of the longitudinal polarizability remains unknown. Expe rimental measurement of solvatochromic shifts in a variety of solvents remains a good a pproach to study polari zability, provided that SWNTs can be suspended in several solvents of known dielectric constant. Our group has observed that mixing a suspension containing so dium dodecylbenzenesulfonate (SDBS)-coated SWNTs with immiscible organic solvents indu ces solvatochromic shifts dependent on the solvent. [ 76] The fluorescence spectra of SDBS-SWNT suspensions mixed with odichlorobenzene (ODCB) showed identical shif ts to the PL from SWNTs suspended in only ODCB (i.e., no surfactant or water). [ 36] The similarity in the peak positions indicates that the hydrophobic core of the surfactant forms an em ulsion-like environment of ODCB around the nanotube. In this chapter, we describe using th ese new microenvironments around SWNTs to investigate the photophysical propert ies of SWNTs in a variety of solvents and mixtures. The solvatochromic shifts follow the expected behavior from a polarizable solute in a polarizable solvent. The fluorescence intensity is shown to be very sensitive to small concentrations of polar 79

PAGE 80

components. The peak positions of known solvent environments are then used to describe the dielectric media in surfacta nt-coated SWNT suspensions. 5.1 Experimental Procedure 5.1.1 Dispersion All experiments used SDBS since this surfact ant did not demonstrat e any changes to the surfactant structure upon mixing with organic solvents, as observed for SDS-SWNT suspensions [ 76]. SWNT suspensions (Rice HPR 145.1) were pr epared in 1 wt% SDBS solutions followed by ultracentrifugation. Aqueous suspensions of na notubes were prepared by mixing 20 mg of raw SWNTs with 200 mL of a sodium dodecylbenzene sulfonate (SDBS) solution (1 wt.%). Highshear homogenization (IKA T-25 Ultra-Turrax) for 1.5 2 h and ultrasonication (Misonix S3000) for 10 min were used to aid disper sion. After ultrasonication, the mixture was ultracentrifuged at 20,000 rpm for 5 h (Beckman Coulter Optima L-80 K). Immiscible solvents were added to each SWNT suspension (solvent: water volume ratio of 0.5) and mixed. The dielectric constant, refractive index, dipole moment, and inducti on polarization for each solvent is given in Table 5-1 The mixture was shaken vigorously fo r 30 s with a vortex stirrer. A white emulsion phase immediately started to phase sepa rate after shaking. After waiting for 1.5 2 h to reach steady state, the excess organic solvent was then carefully removed from the aqueous SWNT suspension to prevent further emulsi fication. The mixtures of hexane and odichlorobenzene (ODCB) were prepared by mixing di fferent molar ratios of each solvent prior to their interaction with aqueous SWNT suspensions The same solvent: water volume ratio (0.5) and procedures were used to mix, separate and characterize the suspension. 5.1.2 Characterization The aqueous phase was characterized by vis-NI R absorbance and NIR-fluorescence spectra using an Applied NanoFluorescence Nanospectro lyzer (Houston, TX) with excitation from 662 80

PAGE 81

and 784 nm diode lasers. Although the organic solv ents have some absorbance bands in the NIR region, their affect on the spectral properties were determined to be minor. This conclusion is supported by the fact that no changes to the ab sorbance spectra of SWNTs are observed, which is consistent with a very small volume of solvent in the system. Likewise, the presence of solvent was concluded to have minimal effect s on the fluorescence emission intensity. 5.2 Results and Discussion 5.2.1 Solvatochromic Shifts Figures 5-1 (A) and (B) show peak shif ts of the PL spectra for th e (7, 6) and (10, 5) SWNT types when mixed with a variety of nonpolar solv ents having different dielectric constants, refractive index, and polarity. As seen in the figu res, systematic changes are observed in both the intensity and peak position. These changes show that the emission ener gy red-shifts and the intensity decreases as the solven t polarity is increa sed from hexane to ODCB. In addition, the peak widths get broader in high polarity solvents, such as 3, 4dichlorotoluene and ODCB. The absorbance spectra for both the E 11 and E 22 transitions of the (7, 6) SWNT shown in Figures 5-1 (C) and (D) also red-shift as the solvent polarity is increased. The E 22 transitions are generally weaker than the E 11 peak shifts, in agreemen t with prior observations. [ 110 ] The peak shifts in the spectra of Figure 5-1 indicate that SWNTs are experiencing different environments when mixed with nonpolar organic solvents. SWNTs have no net dipole moment but are highly polarizable. The presence of di fferent solvent environments around the SWNTs results in different reaction fields, whic h affect the absorbance spectra (Figures 5-1 (C) and (D)). The excited state of SWNTs afte r photon absorption forces the so lvent structure to rearrange around the excited state SWNT dipole. The change in solvent orientation establishes slightly different reaction fields around the SWNTs, re sulting in solvatochromic shifts of the fluorescence emission of SWNTs (Figures 5-1 (A) and (B)) that differ from those observed in 81

PAGE 82

absorbance. The observed spectral changes in Figure 5-1 suggest the measured solvatochromic shifts are due to the formation of a solv ent microenvironment encapsulating the SWNTs. Different ratios of hexane and ODCB were prepared and mixed with the aqueous SWNT suspension to verify that a microenvironment en capsulating the SWNTs forms. The polarity and dielectric constant of hexane and ODCB are the lowest and highest, respectively, of all solvents used in Figure 5-1 The fluorescence spectra for e ach mixture is shown in Figure 5-2 (A) for the (10, 5) SWNT. Pure hexane shows the highest fluorescence intensity and largest emission energy. As the fraction of ODCB in the mixture increases, the fluorescence intensity steadily decreases while the peak position red-shifts. Thes e trends continue until the spectra converge to that observed for pure ODCB. In ideal non-ass ociating solvent mixtures, the solvatochromic shift should follow a simple mixing rule, yielding a linear trend of the solvatochromic shift as a function of mole fraction. [ 112] Figure 5-2 (B) shows that the p eak shift for multiple ( n, m ) types are indeed linear, demonstrating that systema tic changes to the so lvent environment around SWNTs is occurring. The PL intensity of each SWNT type, however, shows non-linear behavior. The fluorescence intensity quickly falls as the fr action of ODCB is increased. These intensity decreases are dependent on the SWNT diamet er. For example, the fluorescence emission intensity of small diameter SWNTs (e.g. (8, 3) types) decrease by approximately 60% while the fluorescence of large diameters (e.g. (10, 5) types) decrease by 80%. These results also show that SWNT fluorescence emission is very sensitive to even small amounts of a more polar solvent (the intensity decreases by approximately 50% at x ODCB 0.2). This sensitivity could have significant implications in understanding the fl uorescence emission inte nsity from aqueous suspensions, where minor amounts of water in contact with th e SWNT sidewall may 82

PAGE 83

significantly affect fluorescence in tensity. Indeed, researchers ha ve found that better surfactant layers result in higher fluorescence intensity. [ 76, 89] 5.2.2 SWNT-Solvent Interaction The solvents used for these experiments are all considered to be nonpolar because of the relatively low dielectric constants. [ 113] Therefore, the solvatochromic shifts for each solvent are primarily influenced by the polarizabilit ypolarizability interactions given by [ 98, 112]: solventair 11 solvent solvent 11 air 11 11f a CEEE 3 (5-1) where C solvent is a fluctuation parameter associated with SWNT dispersion forces, 11 is the change in polarizability of SWNTs between the ground and excited states, is a shape factor for the SWNT, is a parameter associated with the loca tion of the SWNT dipo le in the volume a 3 and f is the solvent induction polarization described by the Onsager polarity functions, f ( 2 ) = 2( 2 1)/(2 2 + 1). [ 98, 112] Figure 5-3 shows the measured PL peak shifts from the data in Figure 5-1 relative to air for multiple ( n, m ) types as a function of f ( 2 ). The solvatochromic shifts ( E 11 ) are larger for the small diameter (lar ge band gap) SWNTs. The solvatochromic shifts also tend to be linear as expected from equation (5-1). Ho wever, the solvatochromic shifts start to exhibit non-linear behavior at high f ( 2 ). The general solvent effect models used here ignore higher order terms, such as dipole moments induced in the solvent molecules by the SWNTs and vice versa. [ 98, 112] These effects are expected to yield a curved relationship to the solvatochromic shift plots shown in Figure 5-3 [ 112] The largest deviations are for those solvents that possess a small dipole. 83

PAGE 84

The E ii transitions for each ( n, m ) SWNT type from either the fluorescence or absorbance spectra can be used to estimate the SWNT polar izability or its functi onal form; however, the fluorescence data is used because the ( n, m ) peaks are better resolved. Previously, Choi and Strano [ 110] concluded that the difference in SWNT polarizability used in equation (5-1) is primarily determined by the longitudinal polarizability of the exciton (i.e., 11 11,|| ). Several research groups have described the lo ngitudinal polarizability of each SWNT ( n, m ) type by a function of the form, where k is a constant, R is the SWNT radius, E b 11 a 11EkR ||, 11 is the band gap as measured in air, and a and b are integers. [ 106-109] As described above, there is no consensus on the relationship (i.e., constants a and b) for the polarizability. If the volume associated with the solvent reaction field is assu med to have a radius equivalent to the SWNT radius, then equation (5-1) becomes: fERDfEkR C Eb 11 a solvent b 11 a solvent 11 3 3 (5-2) For simplicity, the constants C solvent k and are combined into one constant D solvent Since equation (5-2) is valid for all solvents, the measured solvatochromic shifts in Figure 5-3 can then be compared to this equation to determine the global variables, a and b, and the solvent specific variables, D solvent All fluorescence spectra in Figure 5-1 were deconvoluted into their respective ( n, m ) peaks. Details of the deconvolution and fitting pro cedure are given in the supporting information. Briefly, the confidence level of each peak in the deconvoluted spectra was evaluated to assess its relative importance in regression analysis. This a pproach is chosen because of the disparity in peak intensity and overlap. For example, the fluo rescence peak associated with the (7, 6) SWNT type typically has high intensity and little overlap with other ( n, m ) SWNT types while the (12, 84

PAGE 85

2) SWNT peak often has low intensity and si gnificant overlap. Therefore, more confidence should be assigned to peak parameters and solv atochromic effects for the (7, 6) SWNT type because of its intensity and sepa ration from other peaks. However, the fluorescence spectra of other SWNT types still contains important in formation. To account for the varying confidence levels in each deconvoluted peak, weighting fact ors were assigned based on the relative intensity and amount of overlap. A constr ained nonlinear optimization model was then formulated using the generalized reduced gradient (GRG) method. [ 114] An objective function () was chosen to be the squared sum of residuals mu ltiplied by the weighting factors ( w i ) where E 11 are the measured shifts relative to air or the calculated values from equation (5-1 ). These residuals were summed over all ( n, m ) types and solvents to obtain the final objective function: solvents all all 2 , ,minmn, mn calculated 11 mn measured 11 mn Overlap mn IntensityE Eww (5-3) The objective function (5-3) was first solved for each solven t independently. The fluorescence emission data from all solvent systems was then included in the objective function to obtain the parameters a, b, and D solvent The regression analysis shows that the global solution and 11 of 16 single-solvent optimization solutions yield a = and b = These results deviate from prior experimental and computation resu lts, which obtained values of a ranging from 2 to and b from 0 to [ 106-109] However, note that those solutions that deviate from the other results tend to be the most polar solvents. The changes in intensity and peak position for these systems induce more error in peak assignment. On the other hand, some of these systems give a and b constants comparable to those obtained before. [ 107, 110] The known solvatochromic shifts of SWNTs in solvents can be used to probe unknown environments. [ 98] This approach may be particularly useful in understanding the structure of 85

PAGE 86

surfactants surrounding SWNTs, which still remains unknown. [ 20, 76, 105] Figure 5-4 shows the peak position of fluorescence emission from (7, 6) SWNTs surrounded by the surfactants SC, SDBS, SDS, and CTAB relative to the positions in different solvent microenvironments. The peak positions for SCand SDBS-coated SWNTs ar e similar to those measured in alkanes and other low dielectric media. These results suggest that SC and SDBS molecules are oriented such that the hydrocarbon backbone is in contact with the nanotube si dewall and the anionic groups are extended away from the SWNT. On the othe r hand, the peak positions of SDS and CTAB show considerable red-shifts and have envir onments that are similar to the relatively high dielectric media of 3, 4-dichlo rotoluene and ODCB. These two surfactants have long chain alkanes as the backbone (hexadecane and dodecane) and these systems would be expected to have similar peak positions as the alkane so lvents and SDBS. These deviations suggest the surfactant structures for SDSand CTAB-coated SWNTs differ from the SCand SDBS-based systems. The red-shifts of the fluorescence spectra could indicate that the anionic groups are in close proximity to the nanotube sidewall [ 105] or that the surfactant structure does not adequately cover the nanotube, [ 76, 105] allowing significant interactions with water. 86

PAGE 87

1.051.1 1.15 0.96 1 1.08 1.11 1.8 1.95Intensity or Absorbance (a.u.)Energy (eV)hexane heptane cyclohexane carbon tetrachloride p-xylene benzene toluene 2,6-dichlorotoluene chloroform 1-chlorohexane 1-chlorobutane 1,6-dichlorohexane 3,4-dichlorotoluene o-dichlorobenzene(a)(b)(c)(d)3-heptanol 2-heptanol D C B A Figure 5-1. PL spectra ( E 11 emission) for the A) (7, 6) a nd B) (10, 5) SWNT types and C) E 11 and D) E 22 absorbance for the (7, 6) SWNT type measured in microenvironments of various nonpolar solvents. 00 51 (8,3) (7,6) (10,5)0.02 0.05 0.08xODCB E11 (eV) 00 51 0 0.2 0.4 0.6 0.8 1xODCBNormalized Intensity 0.005 0.01 0.015 0.02 0.025 0.03 0.95 0.97 0.99 1.01 1.03Intensity (a.u.)Energy (eV) 1 2 3 4 5 6(a) (b) (c) A B C Figure 5-2. A) The PL emission spectra for the (10, 5) SWNT in a hexane/ODCB mixture at different mole fract ions of ODCB: ( 1) x ODCB = 0.0; ( 2) x ODCB = 0.2; ( 3) x ODCB = 0.4; ( 4 ) x ODCB = 0.6; ( 5) x ODCB = 0.8; ( 6) x ODCB = 1.0. The dashed line is the position of the E 11 emission energy in air. The B) peak shif t (relative to air not the initial SDBS suspension) and C) intensity changes for multiple ( n, m ) types at different mole fractions of ODCB in a hexa ne/ODCB mixture. The dashed line in C) are guides to show the trend. 87

PAGE 88

0 0.02 0.04 0.06 0.08 0.1 0.360.380.40.420.440.460.480.5 (6,5) (8,3) (8,6) (9,7) E11 (eV)f (2)hexane heptane 1-chlorobutane 1-chlorohexane cyclohexane chloroform 1,6-dichlorohexane carbon tetrachloride benzene toluene p-xylene 3,4-dichlorotoluene o-dichlorobenzene 2,6-dichlorotoluene 3-heptanol 2-heptanol Figure 5-3. Solvatochromic shifts of various ( n, m ) SWNT types in nonpol ar solvents with different solvent induction polarization, f ( 2 ). 1.088 1.092 1.096 1.100 1.104Energy (eV)hexane heptane carbon tetrachloride benzene p-xylene chloroform 2,6-dichlorotoluene 3,4-dichlorotoluene o-dichlorobenzene SC SDBS SDS CTAB Figure 5-4. Comparison of the peak positions for th e (7, 6) SWNT type suspended in surfactants and suspended in sodium dodecylbenzenesulfonate (SDBS) but surrounded by various nonpolar solvents. Sodium cholat e (SC), sodium dodecylbenzenesulfonate (SDBS), sodium dodecyl sulfate (SDS), and cetyl trimethylammonium bromide (CTAB). 88

PAGE 89

Table 5-1 Solvent Parameters Solvents Dielectric Constant a Refractive Index a Dipole Moment Induction Polarization f ( 2 ) hexane 1.89 1.37 0.00 a 0.369 heptane 1.92 1.39 0.00 a 0.383 cyclohexane 2.02 1.43 0.00 a 0.411 carbon tetrachloride 2.23 1.46 0.00 a 0.430 p-xylene 2.27 1.50 0.00 a 0.455 benzene 2.28 1.50 0.00 a 0.455 toluene 2.39 1.50 0.38 a 0.455 2, 6-dichlorotoluene 3.36 1.55 0.83 b 0.483 chloroform 4.81 1.45 1.04 a 0.424 1-chlorohexane 6.10 1.42 1.94 a 0.404 3-heptanol 7.07 1.42 1.71 a 0.404 1-chlorobutane 7.28 1.40 2.05 a 0.390 1, 6-dichlorohexane 8.60 1.46 2.03 c 0.430 3, 4-dichlorotoluene 9.39 1.55 3.00 a 0.483 2-heptanol 9.72 1.42 1.71 a 0.404 o-dichlorobenzene 10.12 1.55 2.50 a 0.483 a Values taken from CRC Handbook of Chemistry and Physics [115] b Values taken from Calderbank et al. [ 116] c Values taken from Vaughan et al. [117] 89

PAGE 90

CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS This dissertation covers three research topics that are important to the application of SWNTs. The conclusions associat ed with the removal of SWNT bundles are presented in Section 6.1. The conclusions associated with the nonpolar microenvironment around SWNTs are presented in Section 6.2. The conclusions associated with the solvatochromism of SWNTs are presented in Section 6.3. The recommendations for further study are presented in Section 6.4. 6.1 Removing SWNT Bundles via Interfacial Trapping SWNT bundles can be removed from an a queous suspension of individual SWNTs by interfacial trapping. This techni que offers a simple route to achieve large-scale production of aqueous SWNT suspensions. The significant de crease in absorbance intensity and the Raman aggregation peak combined with increases in fluorescence intensity suggests that nanotube bundles are selectively removed from the aque ous suspension. Although the quality of SWNT dispersions of a single interfaci al trapping step is lower than that of the established ultracentrifugation method, different processing c onditions and multiple extraction steps improve the dispersion quality, making it comparable to ultracentrifugation. A simple model is developed to describe the changes in free energy and helps explain why SWNT bundles preferentially exist at the interface, yielding effective separations. The removal of SWNT bundles th rough the IT process is based on selective partitioning at liquid-liquid interfaces. However, the ionic grou ps on typical surfactants used to disperse SWNTs poses problems for trapping SWNT bundles at the interface. Efficient trapping of SWNT bundles at interfaces is achieved by ad justing the surface charge. Zeta potential measurements show that a minimum surface charge occurs at pH 6.75, which correlates to the most efficient IT separations. A pH-adjusted tw o-step IT process is shown to yield SWNT 90

PAGE 91

suspensions comparable in dispersion quality to suspensions prepared by ultracentrifugation. Most importantly, the IT process is a simple and fast approach with double the yield of suspended SWNT material when compared to the conventional ultracentrifugation method. Further, the IT process can be easily s caled-up to provide s ubstantial throughput. 6.2 Nonpolar Microenvironments Around SWNTs The hydrophobic core of the surfactant structur e surrounding SWNTs can be swelled with immiscible organic solvents. Thes e swelled states could be either a continuous or discontinuous organic layer that results in significant changes to the fluorescence and absorbance spectra. The spectral shifts are reversible disappearing once the solvent is removed. However, some surfactant-solvent systems show permanent increases to fluorescence em ission intensity, which cannot be attributed to changes in concentra tion, dispersion, or doping. These permanent changes to emission intensity are attributed to surf actant reorganization, which helps eliminate fluorescence quenching from the aqueous phase. Thes e results may have important implications in processing SWNT suspensions and analyzi ng the fluorescence of SWNT suspensions. For example, the exposure of SWNTs to small amounts of organic solvents may give the illusion that separation has occurred as seen in Figures 4-1 (B), 4-4 (B), and 4-4 (D). Finally, the changes in fluorescence attributed to surfactan t reorganization suggest that ch anges to processing conditions may yield improved fluorescence intensity (qua ntum yields) of bulk SWNT suspensions. 6.3 Solvatochromism of SWNTs We also demonstrate that multiple solvents form microenvironments around SWNTs with characteristic solvatochromic sh ifts. The shifts scale well with the solvent induction polarization, f ( 2 ), as expected for interactions of a polarizable SWNT with a polarizable solvent. In addition, solvent mixtures follow the expected trends and also show the sensitivit y of SWNT fluorescence to slightly polar solvents. A constrained non linear optimization model was used to study the 91

PAGE 92

polarizability changes with each solvent micr oenvironment. The results yield a longitudinal polarizability of the form 11,|| 1/( R 2 E 11 3 ). 6.4 Recommendations Future work should focus on expanding the comb ination of surfactants and solvents that can produce high-quality dispersions of individual SWNTs by interf acial trapping. This database of viable surfactants and solvents will be beneficial to both industry and academia since the interfacial trapping method is scal able and economical. These studies may also lead to a better understanding of the factors that determine the formation of solvent microenvironments around SWNTs and how the surfactant reor ganizes upon solvent evaporation. Another possible future work should be th e sensing devices based on the unique NIRfluorescence properties of SWNTs. Our work has shown that the fluorescence can be improved after the treatment of immiscible organic solvents. Therefore, this improvement can be utilized in biosensing and bioimaging fields since the fluorescence are widely applied in these research areas. The surfactant structure can be reorgani zed to protect SWNTs fr om extrinsic quenching mechanism after liquid-liquid in teractions to increase the qua ntum yields of SWNTs. The increased quantum yields of nanotubes should help the efficiency and accuracy of the imaging and sensing in medical treatment. 92

PAGE 93

APPENDIX A SUPPORTING INFORMATION FOR INTERFACIAL TRAPPING OF SWNT BUNDLES A.1 Optical Linearity In order to interpret concentration change s in the spectra, the linear regions of the absorbance and fluorescence spectra were de termined. A concentrated Gum Arabic SWNT suspension (1 mg/mL) was homogenized and ultr asonicated and then diluted with 1 wt% Gum Arabic solution. At each dilution step, the abso rbance and fluorescence spectra were recorded. Figure A-1 (A) and Figure A-1 (B) show the absorbance at 763 nm and the intensity (by 662 nm excitation) of the most prominent peak corresponding to the (7, 6) SWNT t ype with the variation of the SWNT concentration, respectively. The initial SWNT concentration strongly affected the absorbance and fluorescence values obtained. These effects could possibly be due to changes in equilibrium concentrations for individual and bundled SWNTs. [ 16 ] Regardless of the initial concentration, the fluorescence spectra showed a linear relation to concentrations of ~ 0.03 mg/m L. The linear response of the absorbance was dependent on the initial concen tration. However, all absorbance spectra was linear to at least 0.1 mg/mL. All spectral data pr esented in the manuscript are within the linear regions. A.2 Correction of Absorbance Spectra of SWNTs In some cases, the absorbance spectra were corrected by subtracting the non-resonant background to obtain the spectra of individual SWNTs. The background was determined by the following equation: b bgk A (A-1) where k and b are the empirical fit parameters. [ 118, 119] 93

PAGE 94

Figure A-2 (A) represents the corrected spectra of SWNTs after sing leand two-step interfacial trapping compared to ultracentrifugation. The spectra show that interfacial trapping does remove some individual SWNTs from the suspension but favors the removal of bundled SWNTs. By comparing the overall absorbance change between the singleand two-step interfacial trappi ng process (Figure 2-7 (B)) and the absorbance ch ange of individual SWNTs (Figure A-2 (A)), it is estimated that only 15% of the total absorbance change in Figure 2-7 (B) is due to removal of individua lly suspended SWNTs. In addition, the final concentration of individual SWNTs in the two-step and ultracentrifuged suspensions are similar. Figure A-2 (B) shows the corrected spectra of SWNTs after interfacial trapping using different oil-to-aqueous suspension volume ratios ( R ). The overall concen tration of individual SWNTs is higher for water-in-oil (w/o) emulsion systems ( R > 1) than oil-in-water emulsions (R < 1). However, the fraction of individual SWNTs in w/o systems is lower than o/w systems as shown in Figure 2-6 (B). 94

PAGE 95

A B Figure A-1. A) Absorbance at 763 nm the (7, 6) SWNTs type as a function of nanotube concentration. Initial SWNT concentrati on of 1 mg/mL B) Fluorescence emission intensity the (7, 6) SWNT type (E x. = 662 nm) as a function of nanotube concentration. Initial SWNT concentration of 1 mg/mL A B Figure A-2. A) Corrected absorbance of ultracentr ifuged, single and two-st ep interfacial trapped SWNT suspensions after subtracting th e non-resonant background. B) Corrected absorbance of interfacial trapped SWNT suspensions at different oil-to-aqueous volume ratios after subtracting the non-resonant background. 95

PAGE 96

APPENDIX B SUPPORTING INFORMATION FOR SOLVATOCHROMISM OF SWNTS B.1 Measured solvatochromic shifts Figures B-1 (A) and B-1 (B) show the fluorescence emission spectra of SWNTs excited with a 662 and 784 nm laser, respectively, in various nonpolar solvent microenvironments. All features of the spectra show solvatochromic shifts as well as intensity changes with each solvent microenvironment. Figure B-1 (C) is the absorbance spectra of SWNTs in the same solvent microenvironments. The spectra of SWNTs shows similar solvatochromic shifts; however, the E (0.85 1.4 eV) transitions of the SWNTs are more sensitive to the solvents (larger shifts) than the E (1.4 2.1 eV) transitions. Figure B-1 (D) shows the 11 22 fluorescence emission spectra of SWNTs in a hexane/ODCB mixture at different mole fractions. The spectra show smooth changes to both the pe ak position and intensity. B.2 Deconvolution of the spectra SWNT fluorescence spectra contains information on the E 11 interband transitions for each ( n, m ) SWNT type. All fluorescence spectra were deconvoluted into their respective bands for each ( n, m ) type using the Applied Na nospectrolyzer software (Hous ton, TX). The deconvolution routine uses Voigt profiles for each ( n, m ) peak. Changes to the position ( E 11 ) and width of each ( n, m ) peak were limited to 0.1% and 3%, respec tively, for each iteration to prevent the misidentification of peaks. After several iterati ons, the mean standard deviation (MSD) between the simulated and experimental data point s was smaller than 0.005 (MSD < 0.005). Figure B-2 shows a typical simulated curve (red) compared to the experimental curve (blue) as well as the deconvoluted peaks corresponding to each ( n, m ) type. The simulated curve shows a good fit to the experimental curve. The peaks for each ( n, m ) component are shown under the simulated curve. The peak position of every (n, m ) type of SWNTs after dec onvolution is used in the 96

PAGE 97

nonlinear optimization model described below. Note that the attached solvent characterization sheets show the deconvolution for all fluorescence spectra. B.3 Nonlinear optimization model The solvatochromic shifts of the interband transitions of SWNTs in each nonpolar solvent is given by the polarizabilitypol arizability interactions [ 98, 112]: solventair 11 solvent solvent 11 air 11 11f a CEEE 3 (B-1) where C solvent is a fluctuation parameter associated with SWNT dispersion forces, 11 is the change in polarizability of SWNTs between the ground and excited states, is a shape factor for the SWNT, is a parameter associated with the loca tion of the SWNT dipo le in the volume a 3 and f is the solvent induction polarization given by f ( 2 ) = 2( 2 1)/(2 2 + 1). The difference in SWNT polarizability, 11 used in equation (B-1) is as sumed to be the longitudinal polarizability of the exciton (i.e., 11 11,|| ) [ 110]. The longitudinal polarizability of each SWNT (n, m ) type is assumed to be a function of the form, where k is a constant, R is the SWNT radius, E b 11 a 11EkR ||, 11 is the band gap as measured in air, and a and b are integers. [106-109] The reaction field volume is assumed to have a ra dius equivalent to the SWNT radius and the constants C solvent k and are combined into one constant D solvent to yield: fERDEb 11 a solvent 11 3 (B-2) All fluorescence spectra were deco nvoluted into their respective ( n, m ) peaks as described above. The E 11 transitions for each ( n, m ) SWNT type were recorded from each spectrum in each solvent and subtracted from the E 11 in air to give the solvatochromic shift ( E 11 ). The optical transitions in air were based on the equation [ 110 ]: 97

PAGE 98

2 3 213cos 1241d A dAA E11 (B-3) where A 1 and A 2 are 61.1 nm and 1113.6. A 3 depends on the chirality of the nanotubes expressed as mod (( nm ), 3) = j For E 11 transitions, A 3 is .077 and 0.032 eV nm 2 for j equal to 1 and 2, respectively. As shown in Figure B-2, there can be considerable overlap in the deconvoluted fluorescence spectra for each ( n, m ) SWNT type. For example, the (7, 6) SWNT type shown in Figure B-2 (A) has good intensity and little overlap with other ( n, m ) SWNT types. On the other hand, the (12, 2) SWNT type in Figure B-1 (B) shows considerable overlap with other ( n, m ) types and low intensity. Obviousl y, more confidence should be assi gned to those peaks that are well resolved and give intense fluorescence emissi on, such as the (7, 6) SWNT type. To account for the varying confidence levels in each deconv oluted peak, weighting factors were assigned based on the relative intensity ( w Intensity ) and amount of overlap ( w Overlap ) in each solvent. This technique is often used in non linear optimization to account for the relative importance of data. [ 114, 120] As shown in Table B-2, both the intensity and overla p were broken up into three categories and assigned valu es of 1, 0.5, or 0.25. Figure B-2 shows some examples of those peaks associated with each category. These two we ighting factors were multiplied together to get the weighting factors listed in Table B-1 for all SWNT types in all solvents. The general trend shown in the table is that the ( 7, 6) and (7, 5) spectra have the highest confidence while the (9, 4) and (12, 1) have the lowest in each solvent. A nonlinear optimization model was then form ulated based on the fluorescence spectra with the objective function () equal to the squared sum of re siduals multiplied by the weighting factors (w i ): 98

PAGE 99

solvents all all 2 , ,minmn, mn calculated 11 mn measured 11 mn Overlap mn IntensityE Eww (B-4) where E 11 are the measured values from Figure 5-5 subtracted from equation (B-4), or the calculated values from equation (B-2). These residuals were summed over all ( n, m ) types and solvents. The objective function was minimized us ing the generalized reduced gradient (GRG) method [ 114 ]. The values of all constants were cons trained to values between and 10 to aid optimization. The global parameters a and b had the additional constraints of being integers. Central differences were used to calculate the partia l derivatives of the objective function. The search direction was determined at each ite ration by using the conjuga te method. The objective function was considered to converge once the rela tive change between iterations was less than 10 -7 A minimum of four different initial starting points for the pa rameters were used to obtain the global minimum rather than local minima. In addition to the solution for all solvents, each solvent was solved independently to obtain the parameters a, b, and D solvent Table B-3 shows the results from all optimization calculations. Some optimization models had minima that were weak functions of the parameters a and b. In other words, the values of a and b could be changed by 1 without affecting the value of the objective function. For example, the optimiza tion of only heptane gave solutions of a = b = and a = b = However, one of these solutions coincided with the global optimization solution. Therefore, th is data set was listed in Table B-3. Nearly all optimization solutions yielded solutions of a = and b = As seen in Table B-3, a global solution with all weighting factors set equal to 1 gave a different solution. Howeve r, as described above, this solution is discarded because of the difference in data quality for each ( n, m ) SWNT type. 99

PAGE 100

A B C D Figure B-1. PL emission spectra of SWNTs excite d at A) 662 and B) 784 nm and C) absorbance spectra of SWNTs in nonpolar microenvir onments. D) PL emission spectra (Ex. = 784 nm) of SWNTs in various hexane/ ODCB mixture microenvironments. All spectra in a-c are offset for clarity and arranged by dielectric constant. 100

PAGE 101

101 Figure B-2. Deconvolution of PL emission spectra for excitation at A) 662 and B) 784 nm. Weighting factors were assigned to each p eak based on the amount of overlap with other peaks and the intensity. Those peaks highlighted in A) show examples of the three weighting factors associ ated with overlap while B) shows those associated with intensity. The green, blue, and purple curv es were assigned values of 1, 0.5, and 0.25, respectively. Table B-1. Criterion for assigning th e weighting factors for both in tensity and amount of overlap with other peaks. Weighting Factor Area Overlap Intensity (662 nm excitation) Intensity (784 nm excitation) 1 < 50% > 0.4 > 0.08 0.5 between 50 and 100% between 0.2 and 0.4 between 0.04 and 0.08 0.25 100% < 0.2 < 0.04 Table B-2. Combined weighting factors ( w = w Intensity w Overlap ) assigned for each ( n ,m ) type as a function of solvent. solvent (6,5) (8,3) (7,5) (7,6) (10,2) (9,4) (11,1) (8 ,6) (9,5) (12,1) (1 1,3) (1 2,2) (1 0,5) (9,7) hexane 0.5 0.5 1 1 0.5 0.25 0.25 0.25 0.5 0.25 1 0.25 1 1 heptane 0.5 1 1 1 0.5 0.25 0.25 0.25 0.5 0.5 1 0. 25 1 1 cyclohexane 0.5 1 1 1 0.5 0.25 0.25 0.25 0.5 0.5 1 0. 25 1 1 carbon tetrachloride 0.5 0.5 1 1 0.5 0.25 0.25 0.25 0.5 0.25 1 0.25 1 1 p-xylene 0.5 0. 5 1 1 0.5 0.0625 0.25 0.25 0. 5 0.125 1 0.25 1 0.5 benzene 0.5 0.5 1 1 0.5 0.125 0.25 0.5 0.5 0.125 1 0. 25 1 1 toluene 0.5 0.5 1 1 0.5 0.125 0.25 0.5 0.5 0.125 1 0. 25 1 0.5 2,6-dichlorotoluene 0.5 0. 5 0.5 1 0.5 0.125 0.0625 0.5 0.0625 0.0625 1 0.25 0.25 0.5 chloroform 0.5 0.5 1 1 0.5 0.125 0.25 0.5 0.125 0.125 0. 5 0.25 0. 25 1 1-chlorohexane 0.5 0. 5 1 1 0.5 0.125 0.25 0.5 0.125 0.125 1 0.25 0.5 1 3-heptanol 0.5 0.5 1 1 0.5 0.25 0.25 0.25 0.5 0.25 1 0.25 1 1 1-chlorobutane 0.5 0. 5 1 1 0.5 0.125 0.25 0.5 0.125 0.125 1 0.25 0.5 1 1,6-dichlorohexane 0.5 0.5 1 1 0.5 0.125 0. 25 0.5 0.125 0.0625 0. 5 0.25 0.25 1 3,4-dichlorotoluene 0.5 0. 5 1 1 0.5 0.0625 0.25 0.5 0. 25 0.0625 0.25 0.125 0.0625 2-heptanol 0.5 0.5 1 1 0.5 0.25 0.25 0.25 0.5 0.25 1 0.25 1 1 o-dichlorobenzene 0.5 0.5 1 1 0.5 0.125 0. 25 0.5 0.125 0.0625 0.25 0. 125 0. 25 average 0.50 0.56 0.97 1.00 0.50 0.16 0. 24 0.39 0.34 0. 19 0. 84 0. 23 0. 73 0. 84 A B

PAGE 102

102 Table B-3. Optimization results for all fitted parameters when solved as individua l systems and collectively. 1 D is given in meV nm 5 eV 3 2 Minimum was weak function and al so gave a possible solution of a = and b = The listed solution was chosen since it was identical to the overall optimization parameter set for all solvents. 3 Minimum was weak function and al so gave a possible solution of a = and b = The listed solution was chosen since it was identical to the overall optimization parameter set for all solvents. 4 Optimization solution with all weighting factors set to 1. Th is solution is not recommended given the uncertainty in the deconvolution. solvent Parameter 1 All solvents hexane heptane 2 cyclohexane 2 carbon tetrachloride p-xylene 3 benzene toluene 2,6-dichlorotoluene chloroform 1-chlorohexane 3-heptanol 1-chlorobutane 1,6-dichlorohexane 3 3,4-dichlrotoluene 2-heptanol o-dichlorobenzene All solvents ( w i = 1) 4 a -2 -2 -2 -2 -2 -2 -2 -2 0 -2 -2 -3 -2 -2 0 -3 -1 -3 b -3 -3 -3 -3 -3 -3 -3 -3 -1 -3 -3 -4 -3 -3 -1 -4 -2 -4 D hexane 4.1 4.1 2.2 D heptane 3.9 3.9 2.1 D cyclohexane 3.6 3.6 1.9 D carbon tetrachloride 3.6 3.6 1.9 D p-xylene 3.6 3.6 1.9 D benzene 3.6 3.6 1.9 D toluene 3.6 3.6 1.9 D 2,6-dichlorotoluene 3.6 13.3 1.8 D chloroform 3.9 3.9 2.0 D 1-chlorohexane 4.0 4.0 2.1 D 3-heptanol 3.8 2.0 2.0 D 1-chlorobutane 4.1 4.1 2.2 D 1,6-dichlorohexae 3.9 3.9 2.0 D 3,4-dichlorotoluene 3.9 14.4 1.9 D 2-heptanol 3.8 2.0 2.0 D o-dichlorobenzene 3.9 7.5 2.0

PAGE 103

LIST OF REFERENCES [1] P. G. Collins, P. Avouris, Sci. Am. 2000, 62. [2] S. M. Bachilo, M. S. Strano, C. Kittrell, R. H. Hauge, R. E. Smalley, R. B. Weisman, Science 2002 298, 2361. [3] M. S. Dresselhaus, G. Dresselhaus, A. Jori o, A. G. Souza Filho, M. A. Pimenta, R. Saito, Acc. Chem. Res. 2002, 35, 1070. [4] M. S. Dresselhaus, G. Dresselhaus, P. C. Eklund, Science of Fullerenes and Carbon Nanotubes, Academic Press, San Diego, CA 1996 [5] R. Saito, G. Dresselhaus, M. S. Dresselhaus, Physical Properties of Carbon Nanotubes Imperial College Press, London 1998. [6] T. W. Odom, J. L. Huang, P. Kim, C. M. Lieber, Nature 1998, 391, 62. [7] J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, H. Dai, Science 2000, 287, 622. [8] L. A. Girifalco, M. Hodak, R. S. Lee, Phys. Rev. B 2000, 62, 13104. [9] M. O'Connell, S. M. Bachilo, C. Huffman, V. C. Moore, M. S. Strano, E. Haroz, K. L. Rialon, P. Boul, W. H. Noon, C. Kittrell, J. Ma, R. H. Hauge, R. B. Weisman, R. Smalley, Science 2002 297, 593. [10] M. Jones, C. Engtrakul, W. K. Metzger, R. J. Ellingson, A. J. Nozik, M. J. Heben, G. Rumbles, Phys. Rev. B 2005, 71, 115426. [11] B. J. Landi, H. J. Ruf, J. J. Worman, R. P. Raffaelle, J. Phys. Chem. B 2004, 108, 17089. [12] A. M. Rao, E. Richter, S. Bandow, B. Chase, P. C. Eklund, K. A. Williams, S. Fang, K. R. Subbaswamy, M. Menon, A. Thess, R. E. Sma lley, G. Dresselhaus, M. S. Dresselhaus, Science 1997 275, 187. [13] D. A. Heller, P. W. Barone, J. P. Swanson, R. M. Mayrhofer, M. S. Strano, J. Phys. Chem. B 2004, 108, 6905. [14] J. A. Fagan, B. J. Landi, I. Mandelbaum, J. R. Simpson, V. Bajpai, B. J. Bauer, K. Migler, A. R. Hight Walker, R. Raffaelle, E. K. Hobbie, J. Phys. Chem. B 2006, 110 23801. [15] V. C. Moore, M. S. Strano, E. H. Haroz, R. H. Hauge, R. E. Smalley, Nano Lett. 2003, 3, 1379. [16] M. S. Strano, V. C. Moore, M. K. Miller, M. J. Allen, E. H. Haroz, C. Kittrell, R. H. Hauge, R. E. Smalley, J. Nanosci. Nanotech. 2003, 3, 81. [17] M. F. Islam, E. Rojas, D. M. Bergey, A. T. Johnson, A. G. Yodh, Nano Lett. 2003, 3, 269. 103

PAGE 104

[18] K. Yurekli, C. A. Mitchell, R. J. Krishnamoorti, J. Am. Chem. Soc. 2004, 126, 9902. [19] R. Bandyopadhyaya, E. Nativ-Rot h, O. Regev, R. Yerushalmi-Rozen, Nano Lett. 2002 2, 25. [20] E. J. Wallace, M. S. P. Sansom, Nano Lett. 2007, 7, 1923. [21] S. U. Pickering, J. Chem. Soc, Trans. 1907, 91, 2001. [22] Y. Lin, H. Skaff, T. Emrick, A. D. Dinsmore, T. P. Russell, Science 2003, 299, 226. [23] J. He, Q. Zhang, S. Gupta, T. Emrick, T. R. Russell, P. Thiyagarajan, Small 2007, 3, 1214. [24] P. Jauregi, M. A. Hoeben, R. G. J. M. va n der Lans, G. Kwant, L. A. M. van der Wielen, Ind. Eng. Chem. Res. 2001, 40, 5815. [25] H. F. Lu, B. Fei, J. H. Xi n, R. H. Wang, L. Li, W. C. Guan, Carbon 2007, 45, 936. [26] B. P. Binks, J. H. Clint, Langmuir 2002, 18, 1270. [27] R. Krupke, F. Hennrich, H. von Lohneysen, M. M. Kappes, Science 2003 301, 344. [28] Z. H. Chen, X. Du, M. H. Du, C. D. Rancken, H. P. Cheng, A. G. Rinzler, Nano Lett. 2003, 3, 1245. [29] M. Zheng, A. Jagota, M. S. Strano, A. P. Sant os, P. Barone, S. G. Chou, B. A. Diner, M. S. Dresselhaus, R. S. McLean, G. B. Onoa, G. G. Samsonidze, E. D. Semke, M. Usrey, D. J. Walls, Science 2003, 302, 1545. [30] M. S. Arnold, A. A. Green, J. F. Hulvat, S. I. Stupp, M. C. Hersam, Nature Nanotech. 2006, 1, 60. [31] K. J. Ziegler, Trends Biotechnol. 2005 23, 440. [32] P. W. Barone, S. Baik, D. A. Heller, M. S. Strano, Nature Mater. 2005, 4, 86. [33] Z. Yao, C. L. Kane, C. Dekker, Phys. Rev. Lett. 2000, 84, 2941. [34] R. K. Wang, R. D. Reeves, K. J. Ziegler, J. Am. Chem. Soc. 2007, 129, 15124. [35] R. Shvartzman-Cohen, E. Nativ-Roth, E. Baskaran, Y. Levi-Kalisman, I. Szleifer, R. Yerushalmi-Rozen, J. Am. Chem. Soc. 2004, 126, 14850. [36] J. L. Bahr, E. T. Mickelson, M. J. Bronikowski, R. E. Smalley, J. M. Tour, Chem. Commun. 2001, 193. [37] T. Rueckes, K. Kim, E. Joselevich, G. Y. Tseng, C.-L. Cheung, C. M. Lieber, Science 2000, 289, 94. 104

PAGE 105

[38] S. Banerjee, T. Hemraj-Benny, S. S. Wong, Adv. Mater. 2005 17, 17. [39] A. Hirsch, Angew. Chem. Int. Ed. 2002, 41, 1853. [40] V. N. Khabashesku, W. E. Billups, J. L. Margrave, Accounts Chem. Res. 2002, 35, 1087. [41] P. M. Ajayan, J. M. Tour, Nature 2007, 447, 1066. [42] C. W. Lee, C.-H. Weng, L. Wei, Y. Chen, M. B. Chan-Park, C.-H. Tsai, K.-C. Leou, C. H. P. Poa, M. Glerup, J. Wang, L.-J. Li, J. Phys. Chem. C 2008, 112, 12089. [43] P. Jauregi, M. A. Hoeben, R. G. J. M. va n der Lans, G. Kwant, L. A. M. van der Wielen, Biotechnol. Bioeng. 2002 78, 355. [44] M. A. Hoeben, R. G. J. M. van der Lans, L. A. M. van der Wielen, G. Kwant, AIChE J. 2004, 50, 1156. [45] H. Wang, E. K. Hobbie, Langmuir 2003, 19, 3091. [46] E. K. Hobbie, B. J. Bauer, J. Stephens M. L. Becker, P. McGuiggan, S. D. Hudson, H. Wang, Langmuir 2005, 21, 10284. [47] K. J. Ziegler, D. J. Schmidt, U. Rauwald, K. N. Shah, E. L. Flor, R. H. Hauge, R. E. Smalley, Nano Lett. 2005 5, 2355. [48] P. Asuri, S. S. Karajanagi, J. S. Dordick, R. S. Kane, J. Am. Chem. Soc. 2006, 128, 1046. [49] A. Menner, R. Verdejo, M. Shaffer, A. Bismarck, Langmuir 2007, 23, 2398. [50] M. i. h. Panhuis, V. N. Paunov, Chem. Commun. 2005, 1726. [51] H. Yi, H. Song, X. Chen, Langmuir 2007, 23, 3199. [52] K. J. Ziegler, U. Rauwald, Z. Gu, F. Li ang, W. E. Billups, R. H. Hauge, R. E. Smalley, J. Nanosci. Nanotech 2007 7, 2917. [53] E. I. Yaseen, T. J. Herald, F. M. Aramouni, S. Alavi, Food Research International 2005, 38, 111. [54] C. Sanchez, D. Renard, P. Robert, C. Schmitt, J. Lefebvre, Food Hydrocolloids 2002 16, 257. [55] A. N. G. Parra-Vasquez, I. Stepanek, V. A. Davis, V. C. Moore, E. H. Haroz, J. Shaver, R. H. Hauge, R. E. Smalley, M. Pasquali, Macromolecules 2007, 40, 4043. [56] Y. Tan, D. E. Resasco, J. Phys. Chem. B 2005, 109, 14454. [57] H. Qian, C. Georgi, N. Anderson, A. A. Green, M. C. Hersam, L. Novotny, A. Hartschuh, Nano Lett. 2008, 8, 1363. 105

PAGE 106

[58] P. H. Tan, A. G. Rozhin, T. Hasan, P. Hu, V. Scardaci, W. I. Milne, A. C. Ferrari, Phys. Rev. Lett. 2007, 99, 137402. [59] O. N. M. Torrens, D. E.; Zheng, M.; Kikkawa, J. M., Nano Lett. 2006, 6, 2864. [60] R. A. Graff, J. P. Swanson, P. W. Ba rone, S. Baik, D. A. Heller, M. S. Strano, Adv. Mater. 2005, 17, 980. [61] M. J. O'Connell, S. Sivaram, S. K. Doorn, Phys. Rev. B 2004, 69, 235415. [62] W. Zhou, M. F. Islam, H. Wang, D. L. H o, A. G. Yodh, K. I. Winey, J. E. Fischer, Chem. Phys. Lett. 2004, 384, 185. [63] H. Wang, W. Zhou, D. L. Ho, K. I. Winey, J. E. Fischer, C. J. Glinka, E. K. Hobbie, Nano Lett. 2004, 4 1789. [64] B. P. Binks, S. O. Lumsdon, Langmuir 2000, 16, 8622. [65] V. N. Manoharan, Solid State Commun. 2006, 139, 557. [66] P. Cherukuri, S. M. Bachil o, S. H. Litovsky, R. B. Weisman, J. Am. Chem. Soc. 2004, 126, 15638. [67] T. K. Leeuw, R. M. Reith, R. A. Simonette M. E. Harden, P. Cherukuri, D. A. Tsyboulski, K. M. Beckingham, R. B. Weisman, Nano Lett. 2007, 7, 2650. [68] P. W. Barone, R. S. Parker, M. S. Strano, Anal. Chem. 2005, 77, 7556. [69] D. A. Heller, H. Jin, B. M. Martinez, D. Pa tel, B. M. Miller, T. K. Yeung, P. V. Jena, C. Hobartner, T. Ha, S. K. Silverman, M. S. Strano, Nature Nanotech. 2009 4, 114. [70] S. D. Bergin, V. Nicolosi, P. V. Streich, S. Giordani, Z. Y. Sun, A. H. Windle, P. Ryan, N. P. Niraj, Z. T. Wang, L. Carpenter, W. J. Bl au, J. J. Boland, J. P. Hamilton, J. N. Coleman, Adv. Mater. 2008, 20, 1876. [71] R. Haggenmueller, S. S. Rahatekar, J. A. Fagan, J. H. Chun, M. L. Becker, R. R. Naik, T. Krauss, L. Carlson, J. F. Kadla, P. C. Trulove D. F. Fox, H. C. DeLong, Z. C. Fang, S. O. Kelley, J. W. Gilman, Langmuir 2008, 24, 5070. [72] R. K. Wang, H.-O. Park, W.-C. Chen, C. Silv era-Batista, R. D. Reeves, J. E. Butler, K. J. Ziegler, J. Am. Chem. Soc. 2008, 130, 14721. [73] L. Vaisman, H. D. Wagner, G. Marom, Adv. Coll. Inter. Sci. 2006, 128-130, 37. [74] B. White, S. Banerjee, S. O'Brien, N. J. Turro, I. P. Herman, J. Phys. Chem. C 2007, 111, 13684. [75] Z. Sun, V. Nicolosi, D. Rickard, S. D. Bergin, D. Aherne, J. N. Coleman, J. Phys. Chem. C 2008, 112, 10692. 106

PAGE 107

[76] R. K. Wang, W.-C. Chen, D. K. Campos, K. J. Ziegler, J. Am. Chem. Soc. 2008, 130, 16330. [77] A. Ishibashi, N. Nakashima, Chem. Eur. J. 2006, 12, 7595. [78] T. Okazaki, T. Saito, K. Matsuura, S. Ohshima, M. Yumura, S. Iijima, Nano Lett. 2005, 5, 2618. [79] T. Kato, R. Hatakeyama, J. Am. Chem. Soc. 2008, 130, 8101. [80] J. Lefebvre, D. G. Austing, J. Bond, P. Finnie, Nano Lett. 2006 6, 1603. [81] R. J. Chen, Y. G. Zha n, D. W. Wang, H. J. Dai, J. Am. Chem. Soc. 2001, 123, 3838. [82] P. Cherukuri, C. J. Gannon, T. K. Leeuw, H. K. Schmidt, R. E. Smalley, S. A. Curley, B. Weisman, P. Natl. Acad. Sci. USA 2006, 103, 18882. [83] Y. Kang, T. A. Taton, J. Am. Chem. Soc. 2003, 125, 5650. [84] Y. Dror, W. Pyckhout-Hintzen, Y. Cohen, Macromolecules 2005, 38, 7828. [85] L. Cognet, D. A. Tsyboulski, J.-D. R. Roch a, C. D. Doyle, J. M. Tour, R. B. Weisman, Science 2007 316, 1465. [86] D. A. Tsyboulski, J. R. Rocha, S. M. Bachilo, L. Cognet, R. B. Weisman, Nano Lett. 2007, 7, 3080. [87] T. Gokus, A. Hartschuh, H. Harutyunyan, M. Allegrini, F. Hennrich, M. Kappes, A. A. Green, M. C. Hersam, P. T. Araujo, A. Jorio, Appl. Phys. Lett. 2008, 92, 153116. [88] M. S. Strano, C. B. Huffma n, V. C. Moore, M. J. O'Connell, E. H. Haroz, J. Hubbard, M. Miller, K. Rialon, C. Kittrell, S. Ramesh, R. H. Hauge, R. E. Smalley, J. Phys. Chem. B 2003, 107, 6979. [89] J. G. Duque, L. Cognet, A. N. G. Parra-Vas quez, N. Nicholas, H. K. Schmidt, M. Pasquali, J. Am. Chem. Soc. 2008, 130, 2626. [90] G. Dukovic, B. E. White, Z. Y. Zhou, F. Wang, S. Jockusch, M. L. Steigerwald, T. F. Heinz, R. A. Friesner, N. J. Turro, L. E. Brus, J. Am. Chem. Soc. 2004, 126, 15269. [91] J. L. Blackburn, T. J. McDonald, W. K. Me tzger, C. Engtrakul, G. Rumbles, M. J. Heben, Nano Lett. 2008, 8, 1047. [92] J. Crochet, M. Clemens, T. Hertel, J. Am. Chem. Soc. 2007, 129, 8058. [93] D. A. Heller, R. M. Mayrhofer, S. Baik, Y. V. Grinkova, M. L. Usrey, M. S. Strano, J. Am. Chem. Soc 2004, 126, 14567. 107

PAGE 108

[94] J. A. Fagan, J. R. Simpson, B. J. Bauer, S. H. De Paoli Lacerda, M. L. Becker, J. Chun, K. B. Migler, A. R. Hightwalker, E. K. Hobbie, J. Am. Chem. Soc. 2007, 129, 10607. [95] M. J. O'Connell, P. Boul, L. M. Ericson, C. Huffman, Y. H. Wang, E. Haroz, C. Kuper, J. Tour, K. D. Ausman, R. E. Smalley, Chem. Phys. Lett. 2001, 342, 265. [96] F. S. Wang, M. Y.; Huang, L.; Huang, X. M. H.; Wu, Y.; Kim, J.; Hone, J.; O'Brien, S.; Brus, L. E.; Heinz, T. F., Phys. Rev. Lett. 2006, 96 167401. [97] T. Nose, N. Numasawa, Comput. Theor. Polm. Sci. 2001, 11, 167. [98] P. Suppan, N. Ghoneim, Solvatochromism The Royal Society of Chemistry, Cambridge, UK 1997. [99] K. R. Moonoosawmy, P. Kruse, J. Am. Chem. Soc. 2008, 130, 13417. [100] M. S. Dresselhaus, P. C. Eklund, Adv. Phys. 2000, 49, 705. [101] W. Zhou, J. Vavro, N. M. Nemes, J. E. Fisc her, F. Borondics, K. Kamaras, D. B. Tanner, Phys. Rev. B 2005, 71, 205423. [102] T. J. McDonald, C. Engtrakul, M. Jones, G. Rumbles, M. J. Heben, J. Phys. Chem. B. 2006, 110, 25339. [103] E. A. G. Aniansson, S. N. Wall, M. Al mgren, H. Hoffmann, I. Ki elmann, W. Ulbricht, R. Zana, J. Lang, C. Tondre, J. Phys. Chem. 1976, 80, 905. [104] S. Arepalli, P. Nikolaev, O. Goreli k, V. G. Hadjiev, W. Holmes, L. Yowell, Carbon 2004, 42, 1783. [105] N. R. Tummala, A. Striolo, ACS Nano 2009, 3, 595. [106] L. X. Benedict, S. G. Louie, M. L. Cohen, Phys. Rev. B 1995, 52, 8541. [107] E. N. Brothers, G. E. Scuseria, K. N. Kudin, J. Phys. Chem. B 2006, 110, 12860. [108] B. Kozinsky, N. Marzari, Phys. Rev. Lett. 2006, 96 166801. [109] G. Y. Guo, K. C. Chu, D. S. Wang, C. G. Duan, Comput. Mater. Sci. 2004, 30, 269. [110] J. H. Choi, M. S. Strano, Appl. Phys. Lett. 2007, 90, 223114+. [111] M. S. Brown, J. W. Shan, C. Lin, F. M. Zimmermann, Appl. Phys. Lett. 2007, 90, 203108. [112] P. Suppan, J. Photoch. Photobio. A 1990, 50, 293. [113] T. H. Lowry, K. S. Richardson, Benjamin-Cummings Publishing Company, 1987. [114] S. Smith, L. Lasdon, J. Comput. 1992, 4, 1. 108

PAGE 109

[115] D. R. Lide, CRC Handbook of Chemistry and Physics. 80th Edition CRC Press. [116] K. E. Calderbank, R. J. W. Le Fvre, R. K. Pierens, J. Chem. Soc. 1969, 968. [117] W. E. Vaughan, National Acad emy of Sciences, Washington, DC 1975. [118] A. G. Ryabenko, T. V. Dorofeeva, G. I. Zvereva, Carbon 2004, 42, 1523. [119] N. Nair, M. L. Usrey, W. J. Kim, R. D. Braatz, M. S. Strano, Anal. Chem. 2006, 78, 7689. [120] K. J. Ziegler, J. Chlistunoff, L. Lasdon, K. P. Johnston, Comput.Chem. 1999, 27, 421. 109

PAGE 110

BIOGRAPHICAL SKETCH Randy Kai-Wei Wang grew up in Taipei, Taiwa n. He received his Bachelor of Science degree from the Department of Chemical Engineering at the National Taiwan University in Taiwan in June 2003. He started his military service in July 2003 and he was discharged in February 2005. He began his graduate studies at the University of Florida in August 2005 and joined Professor Kirk J. Zieg lers nanoscale material research group to pursue a Doctor of Philosophy degree. He graduated in the summ er of 2009 after spending four years being educated in chemical engin eering and material science. 110