<%BANNER%>

Manipulation of the Microenvironment Surrounding Single Wall Carbon Nanotubes and Its Effect on Photoluminescence and Se...

Permanent Link: http://ufdc.ufl.edu/UFE0042883/00001

Material Information

Title: Manipulation of the Microenvironment Surrounding Single Wall Carbon Nanotubes and Its Effect on Photoluminescence and Separation Processes
Physical Description: 1 online resource (188 p.)
Language: english
Creator: SILVERA BATISTA,CARLOS A
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: AGAROSE -- CHROMATOGRAPHY -- OPTICAL -- SANS -- SHEAR -- SOLVATOCHROMISM
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: The photoluminescence from single wall carbon nanotubes (SWNTs) has been studied after manipulating the immediate environment surrounding SWNTs. First, the effect of shearing on the assembly of sodium dodecyl sulfate (SDS) on the surface of SWNTs is probed. Shearing SWNTs coated with sodium SDS in microfluidic channels significantly increases the photoluminescence (PL) intensity and dispersion stability. These improvements are attributed to surfactant reorganization rather than disaggregation of SWNTs bundles or shear-induced alignment. The results highlight potential opportunities to eliminate discrepancies in the PL intensity of different suspensions and further improve the PL of SWNTs by tailoring the surfactant structure around SWNTs. Second, SWNTs are encapsulated with microenvironments of nonpolar solvents, providing a new method to measure the photophysical properties of nanotubes in environments with known properties. Photoluminescence (PL) and absorbance spectra of SWNTs show solvatochromic shifts in 16 nonpolar solvents, which are proportional to the solvent induction polarization. The solvatochromic shifts of SWNTs were used to determine the relationship between the longitudinal polarizability, band gap and radius. Elution chromatography through columns packed with agarose beads has been used to separate metallic from semiconducting SWNTs. Prior studies have attributed the separation to either selective adsorption or size-exclusion (due to selective aggregation) of semiconducting SWNTs. Initial SWNT suspensions with different aggregation states were prepared to test these competing theories. The selective adsorption of nanotubes on the agarose matrix is confirmed by modifying the surfactant structure around the SWNTs without changing the aggregation state of the suspension. In addition, salt-modifiers and solvent-modifiers allow systematic changes to the surfactant aggregation number, orientation, and sidewall coverage. The retention characteristics from these modified SWNT suspensions suggest that surfactant orientation rather the exposed regions on the surface of the nanotubes is the dominant factor in the adsorption process. Small-angle neutron scattering (SANS) is used to study the formation of nonpolar solvent microenvironments on the SWNT surface. It was determined that the solubilization of the nonpolar solvents by the SDS molecules assembled on the SWNT surface do not change the aggregation state of the dispersed nanotubes. In addition, it was possible to obtain the characteristic lengths that describe the solvent microenvironments around SWNTs.
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 CARLOS A SILVERA BATISTA.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Ziegler, Kirk.
Local: Co-adviser: Butler, Jason E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042883:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042883/00001

Material Information

Title: Manipulation of the Microenvironment Surrounding Single Wall Carbon Nanotubes and Its Effect on Photoluminescence and Separation Processes
Physical Description: 1 online resource (188 p.)
Language: english
Creator: SILVERA BATISTA,CARLOS A
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: AGAROSE -- CHROMATOGRAPHY -- OPTICAL -- SANS -- SHEAR -- SOLVATOCHROMISM
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: The photoluminescence from single wall carbon nanotubes (SWNTs) has been studied after manipulating the immediate environment surrounding SWNTs. First, the effect of shearing on the assembly of sodium dodecyl sulfate (SDS) on the surface of SWNTs is probed. Shearing SWNTs coated with sodium SDS in microfluidic channels significantly increases the photoluminescence (PL) intensity and dispersion stability. These improvements are attributed to surfactant reorganization rather than disaggregation of SWNTs bundles or shear-induced alignment. The results highlight potential opportunities to eliminate discrepancies in the PL intensity of different suspensions and further improve the PL of SWNTs by tailoring the surfactant structure around SWNTs. Second, SWNTs are encapsulated with microenvironments of nonpolar solvents, providing a new method to measure the photophysical properties of nanotubes in environments with known properties. Photoluminescence (PL) and absorbance spectra of SWNTs show solvatochromic shifts in 16 nonpolar solvents, which are proportional to the solvent induction polarization. The solvatochromic shifts of SWNTs were used to determine the relationship between the longitudinal polarizability, band gap and radius. Elution chromatography through columns packed with agarose beads has been used to separate metallic from semiconducting SWNTs. Prior studies have attributed the separation to either selective adsorption or size-exclusion (due to selective aggregation) of semiconducting SWNTs. Initial SWNT suspensions with different aggregation states were prepared to test these competing theories. The selective adsorption of nanotubes on the agarose matrix is confirmed by modifying the surfactant structure around the SWNTs without changing the aggregation state of the suspension. In addition, salt-modifiers and solvent-modifiers allow systematic changes to the surfactant aggregation number, orientation, and sidewall coverage. The retention characteristics from these modified SWNT suspensions suggest that surfactant orientation rather the exposed regions on the surface of the nanotubes is the dominant factor in the adsorption process. Small-angle neutron scattering (SANS) is used to study the formation of nonpolar solvent microenvironments on the SWNT surface. It was determined that the solubilization of the nonpolar solvents by the SDS molecules assembled on the SWNT surface do not change the aggregation state of the dispersed nanotubes. In addition, it was possible to obtain the characteristic lengths that describe the solvent microenvironments around SWNTs.
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 CARLOS A SILVERA BATISTA.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Ziegler, Kirk.
Local: Co-adviser: Butler, Jason E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042883:00001


This item has the following downloads:


Full Text

PAGE 1

MANIPULATIONOFTHEMICROENVIRONMENTSURROUNDINGSINGLEWALL CARBONNANOTUBESANDITSEFFECTONPHOTOLUMINESCENCEAND SEPARATIONPROCESSES By CARLOSA.SILVERABATISTA ADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOL OFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENT OFTHEREQUIREMENTSFORTHEDEGREEOF DOCTOROFPHILOSOPHY UNIVERSITYOFFLORIDA 2011

PAGE 2

2011CarlosA.SilveraBatista 2

PAGE 3

Whollydedicatedto: mywife,JulianneD.Vernon,forherlove,compassion,encouragementand understanding; mybrother,JonathanA.SilveraBatista,restinpeace,whosememoryhasbeena motivetobecomeabetterhumanbeingeveryday; mymom,LibiaBatista,forherloveandcarewhileraisingmeandmybrother; mymaternalgrandparentsfortheirimmenseloveandabnegation; myfather,CarlosA.SilveraMejia,forhisingenioushustlingtogivemethe essentialopportunitiestopursuemydesiredpathinlife. 3

PAGE 4

ACKNOWLEDGMENTS IthanktheLouisStokesAllianceforMinorityParticipationLSAMPandProf.Ilona KretzschmarattheCityUniversityofNewYorkforshowingmethewaytograduate school.IgratefullyacknowledgethenancialsupportattheUniversityofFloridathrough thealumnifellowshipandtheSouthEastAllianceforGraduateEducationandthe ProfessoriateSEAGEP.IthankProf.YiiderTsengforaccesstotheresourcesinhis labandtheRichardSmalleyInstituteatRiceUniversityforsupplyingSWNTs.Ialso thankDr.JoontaekParkforhishelpwiththeBrowniandynamicscode.Iamthankful forthecontinuousguidance,supportandpatiencefrommyadvisors,KirkJ.Ziegler andJasonE.Butler.Iamlargelyindebtedwithmyundergraduatecollaborators,John Blazeck,PhilipWeinbergandStevenM.McLeod.Finally,ImustsaythatIsurvivedthe bumpyroadthatgraduateschoolisthankstothededicatedeffortofmywife. 4

PAGE 5

TABLEOFCONTENTS page ACKNOWLEDGMENTS..................................4 LISTOFTABLES......................................9 LISTOFFIGURES.....................................11 ABSTRACT.........................................15 CHAPTER 1INTRODUCTION...................................17 1.1StructureandPropertiesofSingleWallCarbonNanotubesSWNTs...18 1.2Applications...................................20 1.2.1Field-EffectTransistorsFETsandOptoelectronics.........20 1.2.2EnergyConversion...........................21 1.3SeparatingSWNTs...............................22 1.4OtherApplications...............................26 1.4.1OpticalBiosensors...........................26 1.4.2PolymerComposites..........................28 1.5SuspendingSWNTs..............................29 1.6CharacterizingSWNTsthroughPhotoluminescencePLSpectroscopy.32 1.7Organization..................................33 2LONG-TERMIMPROVEMENTSTOPHOTOLUMINESCENCEANDDISPERSION STABILITYBYFLOWINGSWNTSUSPENSIONSTHROUGHMICROFLUIDIC CHANNELS......................................36 2.1Introduction...................................36 2.2Methods.....................................38 2.2.1Dispersion................................38 2.2.2MicrochannelFabrication.......................38 2.2.3ShearingofSWNTsuspensions...................39 2.2.4Dielectrophoresis............................39 2.2.5Characterization............................39 2.3Results.....................................40 2.4Discussion...................................42 2.4.1Shear-InducedAlignment.......................42 2.4.2DisaggregationofSWNTBundles...................43 2.4.3FluidFlowSegregationorRemovalofImpurities..........46 2.4.4SurfactantReorganization.......................46 2.4.5ImprovedDispersion..........................49 2.4.6DiscrepanciesamongSuspensions..................51 2.4.7ShearingEffectsonotherSWNT-SurfactantSystems........53 5

PAGE 6

2.5Conclusions...................................54 3SOLVATOCHROMICSHIFTSOFSINGLEWALLCARBONNANOTUBESIN NONPOLARMICROENVIRONMENTS......................55 3.1Introduction...................................55 3.2Methods.....................................57 3.2.1Reagents................................57 3.2.2AqueousSWNTSuspensions.....................57 3.2.3SolventMicroenvironmentsaroundSWNTs.............57 3.2.4SWNTCharacterization........................58 3.2.5DeconvolutionofPLSpectra......................59 3.3ResultsandDiscussion............................59 3.3.1SolvatochromicShiftsofSWNTSpectrainVariousSolvents....59 3.3.2CharacterizingSolvatochromicShiftsofSWNTs...........63 3.3.3NonlinearOptimizationModelforApproximatingtheLongitudinal Polarizability...............................67 3.4Conclusions...................................72 4SWELLINGTHEHYDROPHOBICCOREOFSURFACTANTSUSPENDED SINGLEWALLCARBONNANOTUBES:ASANSSTUDY............75 4.1Introduction...................................75 4.2Methods.....................................77 4.2.1Reagents................................77 4.2.2AqueousSWNTSuspensions.....................78 4.2.3MixingSWNTsuspensionswithinsolubleorganicsolvents.....78 4.2.4SWNTCharacterization........................78 4.2.5DataAnalysis..............................79 4.3Results.....................................81 4.4Discussion...................................83 4.4.1H-SDSSolution.............................83 4.4.2SWNTSuspension...........................85 4.4.3EffectsofSolventSolubilizationonSDSMicelles..........87 4.4.4EffectsofSolventSolubilizationonSWNTSuspension.......88 4.4.5CharacterizingSolventMicro-StructuresontheSWNTSidewall..91 4.5Conclusions...................................94 5AMECHANISTICSTUDYOFTHESELECTIVERETENTIONOFSDS-SUSPENDED SINGLEWALLCARBONNANOTUBESONAGAROSEGELS.........96 5.1Introduction...................................96 5.2Methods.....................................97 5.2.1Reagents................................97 5.2.2AqueousSWNTSuspensions.....................98 5.2.3ColumnExperiments..........................98 5.2.4DensityGradientUltracentrifugation.................99 6

PAGE 7

5.2.5SWNTCharacterization........................99 5.3ResultsandDiscussion............................99 5.3.1RetentionBehaviorofas-PreparedSDS-SWNTSuspensions...99 5.3.2ProbingSize-ExclusionasaMechanismforSelectiveRetention ofs-SWNTs...............................101 5.3.3ProbingSelectiveAdsorptionasaMechanismforSelectiveRetention ofs-SWNTs...............................103 5.3.4EffectofSurfactantStructureontheSelectiveInteractionofs-SWNTs withAgarose..............................105 5.3.5ChangestoRetentionCharacteristicsbyAlteringtheSurfactant StructurethroughElectrolyteTuning.................106 5.3.6ChangestoRetentionCharacteristicsbyAlteringSurfactantStructure throughSolventSwelling........................109 5.3.7EffectofExposedNanotubeSurfaceonRetentionCharacteristics.112 5.3.8SummarizingtheEffectofSurfactantStructureontheSelective Interactionofs-SWNTswithAgarose.................114 5.4Conclusions...................................116 6BROWNIANDYNAMICSSIMULATIONSOFSWNTSEPARATIONINMICROFLUIDIC CHANNELS......................................117 6.1Introduction...................................117 6.2DescriptionofDEPSeparationDevice....................119 6.3FabricationofDevices.............................120 6.4ModelDescription...............................122 6.4.1EquationsofMotion..........................123 6.4.2EvaluationoftheDielectrophoreticForce...............126 6.4.3NumericalIntegrationoftheGoverningEquations..........128 6.4.4PropertiesofSWNTsandtheSuspendingMedium.........129 6.4.5EvaluatingtheDevicePerformance..................131 6.4.6NumericalCalculationofFlowandElectricFields..........132 6.4.7PerformingtheCalculations......................133 6.5ResultsandDiscussion............................136 6.5.1DevicePerformance..........................136 6.5.2ExperimentalValidation........................138 6.6Conclusions...................................141 7CONCLUSIONSANDFUTUREDIRECTIONS..................142 7.1Conclusions...................................142 7.2FutureDirections................................143 APPENDIX AQUANTUMYIELDMEASUREMENTSANDTHEEFFECTOFBATHSONICATION ONPHOTOLUMINESCENCEAFTERULTRACENTRIFUGATION.......145 7

PAGE 8

A.1FlowParameters................................145 A.2SWNTPLasaFunctionofTime.......................145 A.3QuantumYieldCalculation..........................148 BSUPPORTINGINFORMATIONFORCHAPTER3:DECONVOLUTIONOF THESPECTRAANDSOLVENTCHARACTERIZATIONSHEET........152 B.1DeconvolutionoftheSpectra.........................152 B.2SolventCharacterizationSheets.......................152 CSUPPORTINGINFORMATIONFORCHAPTER5:SETUP,DATAANALYSIS ANDADDITIONALDATA..............................161 C.1ExperimentalSetup..............................161 C.2DataAnalysis..................................162 C.3SWNTSuspensionsatDifferentAggregationStates............162 C.4GumArabic-ModiedSWNTSuspension..................164 C.5Salt-ModiedSWNTSuspension.......................164 C.6Solvent-ModiedSWNTSuspension.....................167 C.7SWNTsSuspendedwithDeoxyribonucleicAcidDNA...........168 REFERENCES.......................................171 BIOGRAPHICALSKETCH................................188 8

PAGE 9

LISTOFTABLES Table page 2-1Quantumyields%forselected n m speciesinshearedsuspensionswith PLintensitiesthatwereinitiallygoodandinitiallypoor...............53 3-1PropertiesofsolventsusedtoformmicroenvironmentsaroundSWNTs.....58 3-2Criterionforassigningtheweightingfactorsforbothintensityandamountof overlapwithotherpeaks...............................67 3-3Combinedweightingfactors w = w Intensity w Overlap assignedforeach n m type asafunctionofsolvent................................69 3-4Optimizationresultsforallttedparameters....................73 3-5Optimizationresultsforallttedparameters....................74 4-1Scatteringlengthdensitiesofreagents.......................78 4-2ParametersusedtotthescatteringfromtheH-SDSsolution..........84 4-3ParametersusedtotthescatteringfromtheH-SDSsolutionaftertreatment..88 4-4ParametersusedtotthescatteringfromtheD-SDS-coatedSWNTstreated withbenzeneandODCB...............................93 5-1EffectthatthemodiershadonPLintensityandretentionbehavior.......107 6-1Operationalregimeofthedevice...........................138 A-1SummaryoftheSWNTsuspensionowparameterswhileowingthroughthe microuidicchannels.................................145 A-2Relativequantumyields%forselected n m speciesinshearedsuspensions thatwereinitiallypoorandinitiallygoodi.e.,suspensionsAandBinFigure 2-5..........................................151 B-1Hexane.........................................153 B-2Heptane........................................154 B-3Cyclohexane......................................155 B-4CarbonTetrachloride.................................156 B-5p-Xylene........................................157 B-6Benzene........................................158 B-7Toluene.........................................159 9

PAGE 10

B-82,6-Dichlorotoluene..................................160 10

PAGE 11

LISTOFFIGURES Figure page 1-1Diagramofmainideas................................18 1-2Graphenesheet....................................19 1-3FETdevicesfromSWNTs..............................20 1-4OrganicphotovoltaicsfromSWNTs.........................22 1-5IEXchromatographyofDNA-SWNT.........................24 1-6SchematicrepresentationoftheDGUprocess...................25 1-7Retentionofs-SWNTsontheagarosegel.....................27 1-8Glucosesensor....................................28 1-9RepresentativeimagesfortheassemblyofSDSon,6a,d,,18b,c, and,30SWNTs..................................31 1-10Densityofstatesforas-SWNT............................33 1-11Opticaltransitionenergiesasafunctionofdiameterfors-SWNTs........34 2-1PLemissionspectrafromSWNTsthatwereshearedinarheometerorby owingthroughthemicrouidicchannels......................41 2-2Probingtheaggregationstateofthesuspensionbeforeandaftershearing throughabsorbanceandRamanspectroscopyanddielectrophoresis......44 2-3pH-dependentPLintensityofthea,7andb,3SWNTspecies.....48 2-4Time-dependentPLintensityofa,3andb,5SWNTspecies......50 2-5PLspectrafromgoodandpoorsuspensionsaftershearing..........52 3-1PLemissionandabsorbancespectraofSWNTsinnonpolarmicroenvironments.60 3-2PLspectra E 11 emissionforthea,6andb,5SWNTspecies.....61 3-3SchematicofthenonpolarsolventmicroenvironmentsformedaroundSWNTs.62 3-4Solvatochromicshiftsofvarious n m SWNTtypesinnonpolarsolventsasa functionofadielectricconstantandbsolventinductionpolarization.....63 3-5ThePLemissionaintensityandbnormalizedintensityforselected n m SWNT typesasafunctionofdielectricconstant......................65 3-6SolvatochromicshiftsofSWNTsinchloroformmicroenvironmentsrelativetoair.71 11

PAGE 12

4-1SANSscatteringintensityfromSWNTssuspendedin1wt%H-SDSandthe respectiveSDSsolution................................82 4-2ScatteringprolefromH-SDSsolutionandSWNTsuspensionswelledwith abenzeneandbODCB..............................82 4-3ScatteringintensityfromSWNTsuspensionsafterbenzeneandODCBhave beenevaporated....................................83 4-4FittothescatteringfromtheH-SDSsolution....................84 4-5SWNTcontributiontoscatteringaftersubtractingthecontributionfromthe H-SDSsolution....................................86 4-6Cartoondepictingthelociofsolubilizationfororganiccompounds........87 4-7Scatteringfromabenzene-andbODCB-swelledSWNTsuspensions....89 4-8Cartoondepictingtheinitialandswelledsurfactantmicelles............91 4-9IntensityfromD-SDS-suspendedSWNTs,D-SDSswelledwithbenzeneand SWNTsuspensionswelledwithbenzene......................92 4-10CartoondepictingthesurfactantassemblyontheSWNTsidewall........94 5-1GeneralretentionbehaviorofSWNTssuspendedin1wt%SDS.........100 5-2Retentionbehaviorof1wt%SDS-SWNTsuspensionspreparedwithdifferent aggregationstates...................................102 5-3RetentionbehaviorofSDS-SWNTsuspensionsmodiedwith0.1wt%ofGA..104 5-4EffectofsurfactantstructureontheinteractionpotentialbetweentheSWNT andtheagarosesubstrate..............................106 5-5Retentionbehaviorofsalt-modiedSWNTsuspensions..............108 5-6AlteringthesurfactantstructurearoundSWNTswithODCBandCCl 4 ......110 5-7RetentionbehaviorofODCB-andCCl 4 -modiedSWNTsuspensions......111 5-8RetentionbehaviorofSWNTssuspendedwithDNA................114 6-1IllustrationoftheDEPseparationdevice......................119 6-2IllustrationoftheDEPseparationdeviceintroducedbyShinetal.........121 6-3WeibulldistributionofSWNTlength.........................131 6-4Cartoonshowingtheproceduretoperforminterpolation..............134 6-5Unitcellsforowandelectriceldcalculations...................135 12

PAGE 13

6-6Effectofthemagnitudeoftheappliedvoltageonthedeviceperformance....137 6-7Effectofthemeanvelocityonthedeviceperformance...............138 6-8Effectofthechannellengthnumberofelectrodesonthedeviceperformance.139 6-9NIRimageoffocusedSWNTs............................140 A-1PLintensityofa,5andb,5SWNTspeciesinaninitialandsheared SWNT-SDSsamplesatdifferentpHvalues.....................146 A-2PLintensityofa,7andb,5SWNTspeciesinaninitialandsheared SDS-SWNTsamplesatdifferenttimesaftershearingthesuspension......147 A-3E 22 regionofabsorbancespectrafortwoshearedsuspensions.4 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 thatinitiallyshowedgoodandpoorPLintensities.................148 A-4EmissionspectraofanSDS-SWNTsuspensionwithpoorPLemissionEx. = 784 nmmeasuredatdifferenttimesasthesuspensionisultrasonicated.......149 B-1DeconvolutionofPLemissionspectra........................152 B-2Hexane =1.89 =1.37 D =0 f 2 =0.369 .................153 B-3Heptane =1.92 =1.39 D =0 f 2 =0.383 ................154 B-4Cyclohexane =2.02 =1.43 D =0 f 2 =0.411 ..............155 B-5CarbonTetrachloride =2.23 =1.4 D =0 f 2 =0.369 ..........156 B-6p-Xylene =2.27 =1.50 D =0 f 2 =0.455 ................157 B-7Benzene =2.28 =1.50 D =0 f 2 =0.369 .................158 B-8Toluene =2.39 =1.50 D =0.38 f 2 =0.455 ...............159 B-92,6-Dichlorotoluene =3.36 =1.55 D =0 f 2 =0.483 ..........160 C-1Schematicofthesetupusedforthecolumnexperiments.............161 C-2Normalizedvis-NIRabsorbanceof1wt%SDS-SWNTsuspensionsultracentrifuged fordifferenttimes...................................163 C-3Retentionbehaviorof1wt%SDS-SWNTsuspensionspreparedatdifferent aggregationstates...................................163 C-4PLemissionspectrafrom1wt%SDS-SWNTsuspensionstitratedwithdifferent amountsofGA.....................................164 C-5RetentionbehaviorofSDS-SWNTsuspensionsmodiedwith0.1wt%ofGA..165 C-6AlteringthesurfactantstructurearoundSWNTswithsalt.............166 13

PAGE 14

C-7Retentionbehaviorofsalt-modiedSWNTsuspensions..............167 C-8ElutionofSDS-SWNTswith1wt%SDS-NaClsolutiongradient.........168 C-9Schematicoftheprocessbywhichthesolventswelledsamplesareprepared.169 C-10RetentionbehaviorofODCB-modiedSWNTsuspensions............169 C-11RetentionbehaviorofCCl 4 -modiedSWNTsuspensions.............170 C-12ComparisonoftheabsorbanceaandphotoluminescencebspectraofSWNT suspendedwithSDSorDNA.............................170 14

PAGE 15

AbstractofDissertationPresentedtotheGraduateSchool oftheUniversityofFloridainPartialFulllmentofthe RequirementsfortheDegreeofDoctorofPhilosophy MANIPULATIONOFTHEMICROENVIRONMENTSURROUNDINGSINGLEWALL CARBONNANOTUBESANDITSEFFECTONPHOTOLUMINESCENCEAND SEPARATIONPROCESSES By CarlosA.SilveraBatista May2011 Chair:KirkJ.Ziegler Major:ChemicalEngineering ThephotoluminescencefromsinglewallcarbonnanotubesSWNTshasbeen studiedaftermanipulatingtheimmediateenvironmentsurroundingSWNTs.First,the effectofshearingontheassemblyofsodiumdodecylsulfateSDSonthesurface ofSWNTsisprobed.ShearingSWNTscoatedwithsodiumSDSinmicrouidic channelssignicantlyincreasesthephotoluminescencePLintensityanddispersion stability.Theseimprovementsareattributedtosurfactantreorganizationratherthan disaggregationofSWNTsbundlesorshear-inducedalignment.Theresultshighlight potentialopportunitiestoeliminatediscrepanciesinthePLintensityofdifferent suspensionsandfurtherimprovethePLofSWNTsbytailoringthesurfactantstructure aroundSWNTs.Second,SWNTsareencapsulatedwithmicroenvironmentsofnonpolar solvents,providinganewmethodtomeasurethephotophysicalpropertiesofnanotubes inenvironmentswithknownproperties.PhotoluminescencePLandabsorbance spectraofSWNTsshowsolvatochromicshiftsin16nonpolarsolvents,whichare proportionaltothesolventinductionpolarization.ThesolvatochromicshiftsofSWNTs wereusedtodeterminetherelationshipbetweenthelongitudinalpolarizability,bandgap andradius, 11, k / 1 = R 2 E 3 11 Elutionchromatographythroughcolumnspackedwithagarosebeadshasbeen usedtoseparatemetallicfromsemiconductingSWNTs.Priorstudieshaveattributedthe 15

PAGE 16

separationtoeitherselectiveadsorptionorsize-exclusionduetoselectiveaggregation ofsemiconductingSWNTs.InitialSWNTsuspensionswithdifferentaggregation stateswerepreparedtotestthesecompetingtheories.Theselectiveadsorptionof nanotubesontheagarosematrixisconrmedbymodifyingthesurfactantstructure aroundtheSWNTswithoutchangingtheaggregationstateofthesuspension.In addition,salt-modiersandsolvent-modiersallowsystematicchangestothesurfactant aggregationnumber,orientation,andsidewallcoverage.Theretentioncharacteristics fromthesemodiedSWNTsuspensionssuggestthatsurfactantorientationratherthe exposedregionsonthesurfaceofthenanotubesisthedominantfactorintheadsorption process. Small-angleneutronscatteringSANSisusedtostudytheformationofnonpolar solventmicroenvironmentsontheSWNTsurface.Itwasdeterminedthatthesolubilization ofthenonpolarsolventsbytheSDSmoleculesassembledontheSWNTsurfacedonot changetheaggregationstateofthedispersednanotubes.Inaddition,itwaspossible toobtainthecharacteristiclengthsthatdescribethesolventmicroenvironmentsaround SWNTs. 16

PAGE 17

CHAPTER1 INTRODUCTION Becauseoftheirexcellentmechanical,electrical,thermalandopticalproperties, singlewallcarbonnanotubesSWNTshavebeenproposedandusedformany applications.However,insteadofbeingasinglematerial,SWNTsareafamilyof materials.Consequently,theirelectricalandopticalpropertiesvaryfromonespecies toanother.TofullytakeadvantageofSWNTpropertiesinthedifferentapplications describedinSection1.2,samplesofnanotubesdisplayingmonodispersityinone orseveralproperties,i.e.bandgap,diameterorlength,areneeded.Post-synthesis separationmethodsareusuallyemployedtoachievethatobjective.Themostefcient separationmethodstailortheSWNTinterfacetoamplifydifferencesindensityor strengthofinteractionwithchromatographicmedia,asdescribedinSection1.3.Some applicationsrequireinsertingSWNTsintodifferentmedia.However,asexplained inSection1.5,SWNTsinteractstronglywitheachother,andthesurfaceofthetube mustbetailoredtocreatefavorableinteractionswiththedesiredmedia.Inapplications suchassensing,theSWNTsurfacemustbemodiedtoincreasethesensitivitytothe presenceoftargetanalytes. Fornanotubestobecomestandardmaterialsinindustry,reliableandversatile methodsofcharacterizationareneeded.Someofthemostfrequentlyusedcharacterization methodsareabsorbanceandphotoluminescencePLspectroscopy.However, spectranotonlydependontheinherentSWNTpropertiesbutalsoontheimmediate environmentsurroundingtheSWNTs.Therefore,anunderstandingofenvironmental effectsisessential.AsshowninFigure1-1,theuseofSWNTsinmanytechnological applicationsdependsontheadvancementintheareasofmetrologyandseparations, whicharelinkedtosomedegree.Allthreeareasoutlinedabovedependtoacertain extentontheabilitytocontrolandcharacterizetheinterfaceofSWNTs.Theobjective 17

PAGE 18

Figure1-1.Diagramshowingtherelationamongthemainideasofthiswork. ofthisworkistostudytheenvironmentaleffectsonthePLresponseandseparationof SWNTs. 1.1StructureandPropertiesofSingleWallCarbonNanotubesSWNTs SWNTsarestablethree-dimensionalstructuresofcarbon.ThewallsofSWNTs consistofalargenumberofcarbonatomsthatarearrangedinhexagonalorhoneycomb units.Duetogeometricalconstraints,theendsofSWNTsmustbecappedduring synthesis.Thecaps,contrarytothemaincylindricalsection,consistofcarbonatoms arrangedinacombinationofpentagonalandhexagonalunits.SWNTscanhavevery largeaspectratios,withdiametersrangingfrom0.7to2nmandlengthsofuptoa fewmicrons.InordertounderstandthestructureofSWNTs,theycanbeconceptually formedbyrollinga2-dimensionalDsheetofgraphenewiththestructureatedges matchedtoformaseamlesscylinder,asdepictedinFigure1-2.Rollingthesheetin differentdirectionswillgiveadifferentorganizationofthehexagonalunitsalongthe mainaxisofthetube.Thedifferentrollingdirectionsanddiametersoftheresultant cylindercanbewrittenintermsofthevectorsdescribingtheprimitivelatticeofgraphene 18

PAGE 19

Figure1-2.SWNTscanbeviewedasagraphenelayerwrappedintoatubularstructure. Thevectorconnectingtwopointsontheplanecanbedescribedbythe indices n and m andtwocoordinates n and m 1 Theresultantvectorwith n m coordinatesiscalledthe chiralvector. Thevarietyofstructuresthatformduetothedifferenceinchiralvectorsgiveriseto SWNTswithawidevarietyofinterestingphysicalandchemicalproperties.Ithasbeen shownthatcarbonnanotubesareverystrongmaterials;thespecicstrengthofSWNTs canbeashighas56timesthatofsteel. 2 SWNTshavealsobeenshowntohaveavery highthermalconductivity,3000W/Km,alongthemainaxisbutamoderateconductivity inthedirectionperpendiculartothemainaxis. 3,4 However,themostinterestingfeature ofSWNTsisthattheycanbehaveeitherasametallicorsemiconductingmaterial.The electricalpropertieshavebeenshowntodependonthediameterandchiralvectorof thetubes.ExperimentsandtheoryrevealthatSWNTshaving n m equaltoaninteger multipleof3aremetallicwhiletherestofthetubesaresemiconductors.Ithasbeen reportedthatmetallicSWNTsm-SWNTscanconductelectricityballistically, 5 which meanselectronsaretransportedwithnoscatteringeventsatroomtemperature.In 19

PAGE 20

Figure1-3.aCartoondepictingaFETdevicefabricatedfromaSWNT.bCartoon representingalightemitterwhereholesandelectronsareinjectedintothe SWNT,whilecshowstheconceptsbehindalightdetector.Figurewas takenfromAvourisetal. 6 addition,evenatthediffusiveconductionregime,SWNTscanhaveveryhighelectron mobilities. 6,7 Theoriginofthispropertyisthe1DnatureofSWNTsandthestrong covalentbondbetweenthecarbonatoms.Hence,onlyforwardandbackwardtransport ofthechargecarriersisallowed.Inaddition,duetotheirsmoothsurface,scatteringby surfacedefectsandroughnessisminimal. 1.2Applications 1.2.1Field-EffectTransistorsFETsandOptoelectronics Recently,therehasbeenanaggressivesearchformaterialsthatallowthe continuousscalingdownorminiaturizationofelectronicdevices.Carbonnanotubes havebeenproposedaspotentialcandidatesduetotheirsuperbintrinsicelectronicand thermalproperties.The1Dnatureandstrongcarbonbondsallowelectronictransport withnegligiblescattering,sosmallerdevicescanbebuilt. 8 EarlyworkusesSWNTs instandardsilicon-baseddevices,likeeld-effecttransistorsFETsbyreplacingthe conductivechannelwithanindividuals-SWNT,asshowninFigure1-3.In1998,Tans andcollaboratorsreportedthefabricationofaFETfromas-SWNT. 9 Recently,the fabricationofananotuberadiowasreported. 10 ItisalsopossibletoinjectholesandelectronsdirectlyintoSWNTswherelightis emittedaftertherecombinationofthechargecarriers,asshowninFigure1-3.The 20

PAGE 21

lightemittedfromSWNTsislocalizedandhasanarrowpeakwidth. 6,11 Figure1-3 showsthattheinverseprocessisalsopossibleandchargecarrierscanbewithdrawn fromSWNTsafterexcitation,constitutingthebasisforphoto-detectors.However, achievingscalabilityandreproducibilityofdevicesmadeofindividualSWNTshasbeen difcult.Themainreasonsaretheneedforpreciseassemblyofeachcomponentand thepresenceofbothm-ands-SWNTspecies.AmorerealisticapproachtoSWNT electronicsmaybetheuseofthin-lmdevices.Thin-lmtransistorsTFTarebased oneasilyfabricatednetworksofSWNTs.Theyareofinterestbecausetheycanbe fabricatedatroomtemperature,canbescaled-upto,forexample,wafersizecircuitry andcanbedepositedonexiblesubstrates,suchasplastics.Theabilitytomakelarge arraysofdevicesonexiblesubstratesisofsignicantimportanceforapplicationsthat requirelight-weight,shockresistanceorexibleformats. 12 Thepresenceofm-and s-SWNTspeciesisalsoproblematicfortheperformanceofTFTs.Ithasbeenshown thattheperformanceofTFTmadeofhighlyenrichedSWNTsamplesdisplaysuperior performance. 13 Inaddition,bychangingthepurityofhighlyenricheds-SWNTsamples, themobilityandon/offratioofmacroelectronicsystemscanbetuned. 14 1.2.2EnergyConversion FacingclimatechangedemandswaystoreducetheemissionsofCO 2 .Asa consequence,thereisanongoingefforttodevelopmethodstoharvestenergyfrom cleanandrenewablesources,suchassolarenergy.OrganicphotovoltaicsOPVs areofinterestbecausetheycanbefabricatedatlowcostandcanbeexibleand versatile.InanOPV,anelectronisexcitedintotheconductionbandwhileaholeis createdinthevalenceband.Duetogeometricalrestrictions,theelectronandthehole areboundthroughCoulombicinteractionscreatinganexciton.Creatingthefreecharge carriersrequiresthattheexcitonisseparatedordissociatedattheinterfacebetweenthe electrondonorandelectronacceptormaterial.Then,freecarriersmustbeconductedto anexternalcircuit. 15 ThebiggestchallengestoconstructOPVsarelowcarriermobility, 21

PAGE 22

Figure1-4.CartoondepictinghowSWNT-basedOPVswork.Thesunlightexcitesthe polymerandtheexcitonisdissociatedattheinterfacebetweenthepolymer andtheSWNT.FigurewastakenfromFergusonetal. 17 highexcitonbindingenergyandthedegradationofthematerials.TheOPVsystemwith thehighestefciencyiscomposedofpoly-hexylthiopheneP3HT,whichworksas theelectrondonor,andaphaseofC 60 fullerene. 16 SWNTswereproposedasnatural candidatestoreplaceC 60 becausetheyarechemicallystableanddisplayhighintrinsic andnetworkelectronmobility.OPVsystemsusingSWNTsastheelectrondonor materialhavebeenbuilt,asshownschematicallyinFigure1-4,buttheefciencyhas beenlow. 17 ItissuspectedthatthepoorperformanceoftheSWNT/polymersystemsis duetothepoordispersionofSWNTsinthepolymermatrixandthepresenceofmetallic andsemiconductingSWNTs.Indeed,devicesformedbywellisolatedandlongSWNTs showedefciencies50to100timeshigher. 18 Ontheotherhand,anenhancementinthe performancewasachievedbyusingsampleshighlyenrichedins-SWNTs. 16 1.3SeparatingSWNTs TheapplicationsdescribedaboveclearlyrequireSWNTsuspensionswitha highdegreeofmonodispersityinlength,diameterandtype.Insearchforsolutionsto thisproblem,researchershavedevisedseveralmethodstoselectivelyproduceand sortSWNTsbyelectronictype,bandgapanddiameter.Themethodscanbedivided 22

PAGE 23

intothreegeneralcategoriesasselectivegrowth,chemicalseparationsandphysical separations.Inselectivegrowth,theideaistocontrolthesizeofthecatalystparticles thatworkasseedsduringthesynthesisprocessinordertocontrolthediameterof SWNTs. 19 Diameterdifferencesbetweentwometallicandsemiconductorspeciescan beaslowas0.03,soitisdifculttoattainsuchlevelofcontrolduringthesynthesis process.ChemicalmethodswhichmakeuseofselectivereactionswithSWNThave beenproposed.Themetallicspecicdiazoniumsaltreactions 20,21 andsaltinduced aggregationhavebeenimportantdevelopments.However,theneedforsedimentation andcentrifugationinthesemethodsproducepoorresponsetimesandselectivity. Therefore,post-synthesisseparationprocessesareneededtoobtainspecictypes ofSWNTs.Physicalseparationmethodshaveshownhighlevelsofselectivity.Using dielectrophoresis,itispossibletoproducedepositsofSWNTsthatwere80%metallic. 22 ImprovingontheworkofKrupkeetal.,Kimetal.wereabletoimprovethepurityof thelmsdepositedbydielectrophoresisattainingSWNTlmsnearly100%metallicin nature. 23 However,onlymicroscopicamountsofSWNTshavebeendepositedthrough dielectrophoreticdeposition 100pg. Zhengetal. 24 wereabletoseparatem-ands-SWNTsbyperformingion-exchange IEXchromatographyofSWNTssuspendedwiththeaidofDNAstrandspolyGT, asshowninFigure1-5a.Whatallowsthediscriminationbetweenthosespeciesis theabilityofm-SWNTstocreateimagechargesandscreenthechargesonthe DNAstrandsduetotheirhigherpolarizability. 25 Hence,whenpassingthroughan ion-exchangecolumn,themetallicspecieseluterstbecausetheirinteractionwiththe chromatographymediaisweaker. TheinteractionsofDNAwithSWNTsandthestructureofthecomplexdepend ontheDNAsequenceandtheSWNTstructure.RecentexperimentsbyZhengand co-workershavetakenadvantageofthatfact.AfteranextensivesearchofDNA sequences,itwaspossibletondstrandsthatcouldformstructuresonagivenSWNT 23

PAGE 24

Figure1-5.aSchematicsoftheinteractionbetweenthepolyGT-SWNTnanotubes andtheIEXresin. 24 bStructureofshortDNAstrandsassembledon SWNTs.ADNAbarrelisformedthroughhydrogenbondsbetweenadjacent DNAstrands. 26 cAbsorbancespectraofelutedfractionsmainlycontaining asingletypeof n m species.Figureswereadaptedfromreferences24and 26. speciesandminimizeitsinteractionwiththeIEXmedia.AsFigure1-5cshows,itwas possibletoobtainfractionswithupto90%purityofasingle n m type. 26 Atthemoment, theyieldandscaleofthemethodarelimitedbytheirreversibleadsorptionofSWNTsto theIEXmediaandthehighpriceofthematerials. Anotherpowerfulmethodisdensity-gradientultracentrifugationDGU.Themethod consistsincentrifugingsurfactant-suspendedSWNTsinadensitygradientmedium. Duringcentrifugation,SWNTsmovesthroughthemediauntiltheirdensityismatched, 24

PAGE 25

Figure1-6.aSchematicrepresentationoftheDGUprocess.Agoodseparation demandsindividualizedSWNTsandadifferenceindensityforthe surfactant-SWNTcomplexes.bImageofthetubeafterultracentrifugation. Agoodseparationisevidentwhenbandsofdifferentcolorsareobserved. cTheextentofseparationisprobedbycollectingtheabsorbancespectra ofthedifferentfractions.FigurewastakenfromArnoldetal. 27 asshowninFigure1-6. 27 Separationoccurswhenthedensitymatchingpointsare differentforthe n m speciesthatcomposethesample.Theversatilityofthemethod stemsfromtheabilitytotunethesurfactantshellbydifferentmeans.Thedifference indensityisaresultofhowthesurfactantassemblesontheSWNTsasdeterminedby thesurfactantstructure,theSWNTtypeanddiameter.First,SWNTscanbeseparated bydiameterusingsodiumcholateSCorsodiumdeoxycholateSDCbecausethe diameterofthetubedeterminesthesurfacedensityofthesurfactantasshownin Figure1-6b-c.Ontheotherhand,thecompetitivebindingofdifferentsurfactants,i.e., sodiumdodecylsulfateSDSandSC,canbeusedtoseparateSWNTsintometallic andsemiconductingspecies.Infact,ithasbeenshownthatSDSpreferentiallyadsorbs onthemetallicspecies,whichcanpotentiallybeduetotheirhigherpolarizability.It hasbeenarguedthatm-SWNTshavetheabilitytoformimagecharges,whichpartially screenedthechargesontheSDSmolecules. 28,29 Asaconsequence,moreSDS moleculescanabsorbonthem-SWNTspecies,whichdecreasestheirdensitywhen 25

PAGE 26

comparedtos-SWNT.Ithasbeendemonstratedthatthebiascanbeampliedthrough furtherscreeningthechargesonSDSmoleculesbyincreasingtheionicstrengthofthe suspension.Theoreticalcalculationspredictthatthedensityofthesurfactant-SWNT complexcanalsobeaffectedbytheabilityofsurfactantmoleculestogoinsidethe SWNT. 30 Separationbyhandednesshasalsobeenachievedbyusingnon-linear insteadofthemostcommonlyusedlineardensitygradients. 31 Inaddition,byrunning theseparationintransientstate,itwaspossibletoseparateSWNTsbylength. 32 The drawbackofDGUistheneedforhighcentrifugalforces,000-200,000 gandlong processingtimes. AnotherinterestingmethodrecentlydevelopedbyKatauraandcollaboratorsis basedontheselectiveinteractionofSDS-coatedSWNTsandagarosegels,asshownin Figure1-7. 33 Separationoccursbecausesometubes,especiallys-SWNTs,showstrong interactionswithagarosegels.Althoughthequalityofseparationwiththismethodis notashighaswiththeothertwomethods,therehavebeentwoimportantresults.First, ithasbeenpossibletoenrichSWNTsbytype,withfractionscontainingasmuchas 90%s-SWNTs.Furthermore,ithasbeenpossibletoenrichs-SWNTsintosuspensions withfewerchiralities.Theadvantageofthistechniqueisthatitisamenabletoscaling. Beadedagaroseiscommonlyusedasagelltrationmediaandtheseparationof SWNTscanbeperformedcontinuouslyinachromatographycolumn. 1.4OtherApplications 1.4.1OpticalBiosensors Thelargesurfacearea,mechanicalstrength,andtheexcellentchemicaland thermalstabilityofSWNTsmakethemsuitableforsensingapplications.Detection andanalysisofindividualSWNTscanbeperformedusinguorescencemicroscopy. Ingeneral,todetectandanalyzesinglemoleculeswithuorescencemicroscopy, itisdesirablethattheuorescentmoleculeuorophoreunderstudyhaveahigh quantumyield,ashortuorescencelifetimeandhighphotostability.Thequantumyield 26

PAGE 27

Figure1-7.a-cTheseparationofm-froms-SWNTscanbeperformedbysqueezinga pieceofagarosegelimpregnatedwiththeSWNTsuspension.dThe methodhasbeenusedtoseparateSWNTsproducedthroughtheHiPcoand laserablationmethodswhichdifferintheirdiameterdistribution.Cartoon depictingtheretentionofs-SWNTsontheagarosegel.Figurewasadapted fromTanakaetal. 33 determineshowmuchtheuorophorefavorsradiativetonon-radiativeprocessesasa meanofrelaxation.Fluorescencelifetimeisthecharacteristictimeofanabsorption-emission cycle,andphotostabilityreferstothenumberofabsorption-emissioncyclesthat auorophorecanundergobeforedecomposingorphotobleaching.Comparedto uorophoresregularlyusedinsinglemoleculestudies,SWNTshavefavorableand unfavorableproperties.Fluorophores,suchasRhodamine6G,haveaquantumyield around90%,whileithasbeenestimatedthatSWNTshavemaximumquantumyieldsof theorderof8%. 34 However,studieshaveshownthatSWNTsdonotphotobleachwhile mostuorophoresphotobleachafteranitenumberofcyclesoftheorderof10 5 35,36 Inaddition,theuorescencelifetimeofSWNTsisestimatedtobeoftheorderof100 ps,whichisashorttimecomparedto10nsofotheruorophores.Theadsorptionof analytesontheSWNTsurfacealsomodulatestheemissionenergyandintensity.The twomostimportantadvantagesofSWNTsarethattheyemitintheNIRpartofthe spectrumandarehighlyphotostablenegligiblephotobleaching.NIRimagingisvery 27

PAGE 28

Figure1-8.DiagramshowinghowaSWNTglucosesensorworks.Figurewastaken fromBaroneetal. 37 convenientinthatmostimpurities,solvents,and,mostimportantly,biologicalmediado notuoresceinthatregionofthespectrum,thusincreasingthesignal-to-noiseratio. Consequently,itispossibletoperformdetectionofanalytesinbiologicalmediafor prolongedperiodsoftime.ThepioneeringworkofStranoandcollaboratorshastaken advantageofSWNTpropertiesinsensingapplications.Theyengineeredthesurface ofSWNTtodetect -D-glucose,asshowninFigure1-8. 37 Analytes,suchasHg,were detectedusingDNA-coatedSWNTssinceHginducesatransitionoftheDNAstrand fromtheBtotheZconguration. 38 Thattransformation,inconsequence,modulatedthe dielectricenvironmentaroundtheSWNTsaffectingtheemissionenergy.Theresolution ofthesensorshasbeenpushedtothesinglemoleculelevelbyanalyzingthequenching eventsonindividualfragmentsofSWNTs. 39 ThisallowedthestudyofH 2 O 2 cellular signalingwithasinglemoleculeresolution.Inaddition,itwaspossibletodiscriminate thespatialoriginoftheH 2 O 2 quenchingmolecule. 40 Also,multipleanalytescanbe detectedbyusingthemodulationoftheemissionenergyandintensityoftwoSWNTs. 41 1.4.2PolymerComposites SWNTsarecurrentlyusedinpolymercompositematerials,whichtakeadvantage ofthesynergisticinteractionsofSWNTsandthehostpolymer.Theideaistoreinforce 28

PAGE 29

thepropertiesofthepolymermatrixwiththeSWNTs.Itisexpectedthatsubstantial enhancementscanbeachievedwithlowloadingofSWNTs.However,polymer compositesloadedwithSWNTsdonotalwaysperformasexpected.Thebiggest challengeistodisperseSWNTsintothepolymermatrixasindividuals.Also,controlling thenatureoftheinterfaceisofutterimportance.Arecentstudyrevealedthatalthough individualtubeshavehighthermalconductivity,oncetheyareinabundleorinagiven polymermatrix,theirsurfaceresistivityincreases. 42 Thestudyalsorevealedthatinorder toenhancethethermalconductivityoftheSWNTcompositeitwasnecessarytowrap SWNTswithcoatingsthathavesimilarphononspectrums.Thereasonisthatalthough SWNTarehighlyconductivethemismatchattheinterfacegeneratesaconsiderable amountofscattering,henceloweringthesurfaceconductivity. 1.5SuspendingSWNTs DispersionandseparationofSWNTsareinherentlylinkedsincenoseparation bytypecanoccurwithoutadequatedispersionofindividualSWNTs.Thestrongvan derWaalsinteractionbetweentubesmakestheprocessofdispersionorsolubilization difcult.SWNTsprefertointeractwiththemselvesratherthanothermediaduetostrong vanderWaalsinteractionsamongthemselves.Toovercomethisissue,oneapproachis toselectsolventswhereSWNThavealargeenthalpygainbysolubilizinginthemedia. TheworkofColemanshowsthat N -methyl-2-pyrrolidoneNMPcanbeusedforthat purpose. 43,44 However,thelevelofexfoliationandtheamountthatcanbesolubilized couldbeimproved.AnotherapproachistochangethenatureoftheSWNTsurface sothattheyhavemorefavorableinteractionswiththemediabycovalentlyattaching functionalgroupstotheSWNTsurface.However,thisapproachisnotidealsincethe electronicpropertiesoftheSWNTareperturbedintheprocess,whichalsoremovesthe possibilitytocharacterizeSWNTsthroughPLspectroscopy. 45 Themostcommonlyusedapproachistonon-covalentlyfunctionalizethesurface ofSWNTsusingcompoundsthatinteractfavorablywithboththesuspendingmedium 29

PAGE 30

andtheSWNTs.Themostcommonlyusedsuspendingagentsaresurfactantsand bio-aswellassyntheticpolymers.Themethodusuallyrequiresaninputofenergy intothesystemfromsonicationtoovercomethestrongvanderWaalsinteractions betweenthetubes.Oncethetubesareseparated,thesuspendingagentattaches totheSWNTsidewalltoavoidaggregation.Inmostapplications,itisnecessaryto separatetheremainingaggregatedorbundledSWNTsfromtheindividualSWNTs. Whileultracentrifugationhasbecomethestandard,itispossibletotakeadvantage ofthedifferentsurfaceactivityofbundledandindividualnanotubestoseparatethem atwater/oilinterfaces. 46,47 Theprocessingconditionsareofutterimportancesinceit hasbeenshownthatsonicationcutstheSWNTsandintroducesdefectsontheSWNT surface. 48 Ionicnegativelychargedaswellcationicpositivelychargedsurfactantshave beenused.ThemostcommonlyusedsurfactantsaresodiumdodecylsulfateSDS, sodiumdodecylbenzenesulfonateSDBS,andbilesaltssuchasSCandSDC.Ithas beendeterminedthatthemostimportantparametertoensureoptimaldispersionis thetotalconcentrationofsurfactantratherthantheratioofsurfactanttoSWNTs.Ithas alsobeenshownthatsmallerSWNTsareeasiertosuspendbysomesurfactants. 49 Forexample,SDSwasfoundtosuspendHiPcoSWNTs 1nmeffectivelybutnot largerdiametertubes,suchasthoseproducedthroughthearcdischargeorlaser ablationmethods. 50 SDContheotherhandcansuspendsmallaswellaslargediameter SWNTs. Itwasoutlinedabovehowtheassemblyofsuspendingagentsdeterminesthe extentofSWNTseparationbydifferentmethods.Toimprovethecurrentseparation methodsandperhapsdevelopnewones,itisessentialthattheassemblyofthe suspendingagentsonSWNTsisbetterunderstood.Theassemblyofsurfactantson SWNTsislessunderstoodthantheassemblyofDNAonSWNTs.Itisalsobecoming increasinglyclearthatSWNT-surfactantandsurfactant-surfactantinteractionsare 30

PAGE 31

Figure1-9.RepresentativeimagesfortheassemblyofSDSon,6a,d,,18b,c, and,30SWNTsc,f.Theinsetsshowthenumberdensityofheadand tailgroupsasafunctionofthedistancefromtheSWNTsurface.TheSDS surfacedensitywaseithera-chigh.8moleculespernm 2 andd-flow .0moleculespernm 2 .Thehydrophobictailsarerepresentedbythegreen lines,whilethesulfurandoxygenatomsarerepresentedbytheredand yellowspheres.NoticethepartsoftheSWNTthatareexposedatlow surfacedensityforthe,6species.FigurewastakenfromXuetal. 57 importantindeningbothdispersionandPLproperties. 51,52,28,53 Threemodelsof surfactantassemblyonSWNTsidewallshavebeensuggested:hemispherical, 54 hemicylindrical, 55 andassembliescharacterizedbylackoflongrangeorder. 56 However, arecentsimulationstudysuggeststhelattermodelmaybethemostappropriate. 53 This studyshowsthatsurfactantmoleculesdonothavelong-rangeorderaroundSWNTs andlargeportionsofthenanotubesidewallmaybeexposedtowater.Adifferentstudy alsosuggeststhattheassemblyofSDSonaSWNTsurfacedependsonthesurface densityofthesurfactantmolecules. 57 Itwasshownthatallthreesurfactantassemblies arepossibleonSWNTs,andtheycanbeachievedbyvaryingthesurfacedensityor aggregationnumberofthesurfactant.Figure1-9showssnapshotsfromthesimulation results.Atahighsurfacedensityof2.8moleculespernm 2 ,SDSmoleculesassemble ascylindricalmicelleonthesurfaceof,6nanotubes.AsthediameterofSWNT increases,theSDSmoleculesassembleintohemimicellesasshownforthe,18and 31

PAGE 32

,30speciesinFigure1-9c-b.AsthesurfacedensityofSDSdecreases,thecoverage ofSDSmoleculesonthe,6speciesisincompleteleavingpartsoftheSWNT sidewallexposed.Interestingly,thesurfactantheadgroupsareclosetothesurface ofthe,18and,30species,formingauniformmonolayer.Thisisanimportant factsincethedifferentstructurescouldhavedifferenteffectsinseparationprocesses. Althoughtherehavenotbeenstudiescomparingtheassembliesonsemiconductor andmetallicnanotubes,ithasbeensuggestedthatthemetallicspeciesmighthavea largersurfactantdensityontheSWNTsurface.Hence,itispossiblethatm-SWNThave differentsurfactantcongurationontheSWNTsurface.Insummary,akeyreoccurring featurefromthesimulationsishowthesurfacedensitycaninuencetheassemblyofthe suspendingagentontheSWNTsurface. 1.6CharacterizingSWNTsthroughPhotoluminescencePLSpectroscopy Duetotheirquasi-onedimensionality,SWNTsdisplaysharpvanHovepeaksin theirdensityofelectronicstates,asshowninFigure1-10.Transitionsfromonepeak inthevalencebandtoanotherintheconductionbandorviceversadeterminethe opticalpropertiesofSWNTs.However,thesefeaturesareonlyobservedforisolated SWNTssincecontactwithotherspeciescanperturbthesestates.Therefore,theability tosuspendandisolateSWNTsusingsurfactantswasabreakthroughthatenabled theuseofabsorbanceandPLspectroscopytoquicklyidentifys-SWNTsfromabulk sample.ThePLfromas-SWNTatenergy E 11 isobservedafterlightabsorptionoccurs atenergy E 22 .ItispossibletoidentifySWNTsinanbulksamplebecausethebandgap ofSWNTsisaninversefunctionofthediameter.Fluorescencemeasurementsalong withRamanmeasurementsallowedtheobservedspectralfeaturestobeassignedto specic n m species.Measurementson33differentnanotubeswereusedbyBachilo andWeismantocreateafunctionthatallowspredictingahundred n m species,that rangefrom0.48to2.0nmindiameter.Theexperimentalvaluesandthepredictionsare showninFigure1-11. 32

PAGE 33

Figure1-10.Densityofstatesforas-SWNT. E 11 isTheenergydifferencebetweenthe rstpeaksinthevalenceandconductionbands. E 22 istheenergy differencebetweenthesecondpeaksinthevalenceandconductionbands. 1.7Organization ThefactthattheassemblyofsurfactantsontheSWNTsdependsonthesurface densityisofsignicantimportance.Thisofferstheopportunitytomanipulatethe surfactantstructuretochangetheassemblyandenhance,forexample,thePLquantum yieldQYfromSWNTsandimprovetheefciencyofseparations.Changesinthe surfacedensitycanbeinducedbycontrollingthehydrophobicvolumeandpolarareaof thesurfactant.Therstcanbeaccomplishedbyusingchargescreening.Thesecond canbeaccomplishedbycapitalizingontheabilityofsurfactantaggregatestosolubilize nonpolarorganiccompounds. 51 InChapter2,IexploretheeffectofSWNTprocessing,suchasowthrough microchannelsonthesurfactantstructureandPLemissionofSWNTs.Thehighshear environmentinthechannelresultsinsignicantincreasesinPLintensityafterowing throughthechannel.Thesechangespersistformonthsandcannotbeattributed toshear-inducedalignment,disaggregationoruidowsegregation.Instead,itis 33

PAGE 34

Figure1-11.Opticaltransitionenergiesasafunctionofdiameterfors-SWNTs.Solid symbolsareexperimentaldata;opensquaresandcirclesarepredictionsof E 11 and E 22 ,respectively,fromempiricalttingfunctions.Figurewastaken fromWeismanandBachilo. 58 suggestedthatshearingaSWNTsuspensionhelpsannealthesurfactantshellaround thenanotube.Thealteredsurfactantstructureenhancesthestabilityofthesuspension andincreasesthePLintensity,especiallyforthelargestdiameterSWNTs.Ialso demonstratethatthisprocesscaneliminatediscrepanciesinthePLintensityofdifferent suspensions. InChapter3,microenvironmentsofnonpolarsolventsareusedtosystematically studyhowSWNTPLvariesasthepolarityofthemediumchanges.Thechangesin emissionenergyfollowtheexpectedbehaviorfromapolarizablesoluteinapolarizable solvent.ThePLintensityisshowntobeverysensitivetopolarsolvents. Chapter4presentsastudyofthestructuralfeaturesofthenonpolarsolvent microenvironmentreferredtoinChapter3.Small-angleneutronscatteringSANSwas 34

PAGE 35

usedtoaddressedquestionsaboutthesizeofthesolventdomains,theiruniformityand distributionandtheeffectofsolventswellingontheaggregationstateofthenanotubes. Chapter5showshowprocessesthatmodifythesurfactantassemblyonSWNTs wereusedtoexplorethemechanismofchromatographicseparationofSWNTsin agarosecolumns.Thetwopotentialmechanisms,size-exclusionandselective adsorption,areprobed.Theresultsconrmthatselectiveadsorptionisresponsible fortheseparationofs-fromm-SWNTs.Finally,thefeaturesofthesurfactantshellthat enableadsorptionareinvestigatedaswellasthenatureofanactiveadsorptionsiteon theSWNTsurface. Chapter6describeshowdielectrophoresisandmicroudidicsystemscanbeused toseparateSWNTsbytype.Browniandynamicssimulationswereperformedtoassess theperformanceofamicrouidicdevicetoseparateamixtureofm-ands-SWNTs.It isshownthatbychoosingtherightchannelgeometryandoperationalparameters,itis possibletoobtainfractionsthatareeitherhighlyenrichedinm-SWNTsors-SWNTs. Experimentalresultsaimedatvalidatingthepredictionsarealsoshown. 35

PAGE 36

CHAPTER2 LONG-TERMIMPROVEMENTSTOPHOTOLUMINESCENCEANDDISPERSION STABILITYBYFLOWINGSWNTSUSPENSIONSTHROUGHMICROFLUIDIC CHANNELS 2.1Introduction SincethediscoveryofSWNTs,researchershaveenvisionedmanyapplications thattakeadvantageoftheirastoundingphysicalproperties. 59 However,dispersingand separatingas-grownSWNTsremainsasignicantimpedimenttotheiruseinmost applications. 60 DispersionandseparationofSWNTsareinherentlylinkedsinceno separationbytypecanoccurwithoutadequatedispersionofindividualSWNTs.A signicantmilestonewasthedispersionofSWNTsinaqueoussuspensionswiththeaid ofsurfactants. 55 MostworkhasfocusedontheuseofsodiumdodecylsulfateSDS, sodiumdodecylbenzenesulfonateSDBS,andsodiumcholateSC,althoughmultiple surfactantshavebeenused. 61,62,63 ThedispersionofSWNTsenables n m typecharacterizationthroughabsorbance andphotoluminescencePL. 64,65 ThePLemissionisoftenusedtoquantifyseparations ofsemiconductingSWNTs. 66,67,68 However,theeffectofthesurfactantandsurrounding environmentisstillnotwellunderstood. 69,34,51 Forexample,differentintensities andquantumyieldsareobtainedwhensuspendedbydifferentsurfactantsorby differentresearchgroups. 62,63,54 Thesedifferencesinintensityhavebeenattributed todifferencesindispersion 70 orextrinsicenvironmentaleffects.Theopticallyexcited electronicstatesofSWNTsarehighlymobile, 71 makingthemsensitivetoextrinsic effectsthatreducePLintensity,includingthestateofaggregation, 72,73,74,75 polarizability ofthesurroundingenvironment, 51,76 pHofthesuspension, 71,52,77 surfacereactions, 71,78,69,79 sidewalldefects, 71,34,73,80,81 surfactantinhomogeneties, 34,73,80,82 ,electricelds 83 and thelengthofSWNTs. 84,48 Single-moleculestudiesofSWNTshavebeenimportantinidentifyingand characterizingtheeffectsofextrinsicfactorsonPL.FluctuationsinPLemissionintensity 36

PAGE 37

alongthelengthofnanotubeshasbeenobservedinbothaqueoussolutions 34 andair. 81 ThesedarkspotsallowedCognetetal. 71 toestimatetheexcitondiffusionlengthtobe approximately90nm.InSWNTsthathaveuniformPLintensity,quantumyieldsQYs areestimatedtobe78%. 34,81 Thesemeasurementssurpasspreviousensemble estimatesofQY.11%. 55,82,85 Therefore,thepresenceofextrinsicfactorsmayhelp explainthediscrepancybetweenbulkandsingle-moleculemeasurementsofSWNTQY. ItisbecomingincreasinglyclearthatSWNT-surfactantandsurfactant-surfactant interactionsareimportantindeningbothdispersionandPLproperties. 51,52,28,53 ThreemodelsofsurfactantassemblyonSWNTsidewallshavebeensuggested: hemispherical, 54 hemicylindrical, 55 andassembliescharacterizedbylackoflong rangeorder. 56 However,arecentsimulationstudysuggeststhelattermodelmaybe themostappropriate. 53 Thisstudyshowsthatsurfactantmoleculesdonotpresent long-rangeorderaroundSWNTsandlargeportionsofthenanotubesidewallmay beexposedtowater.ThishasimportantimplicationsonthePLemissionintensity sinceseveralresearchershavedemonstratedthequenchingeffectsofprotonsor water, 71,77,82 indicatingthatPLemissionofSWNTsmaydependonbothextrinsic factorsandprocessesthatalterthesurfactantstructure. Researchershavedeterminedthatprocessingconditionshaveasignicantimpact onthePLemission. 86 Theultimateobjectiveistocompensateforthesedifferences. 62,34 Differencesinsurfactantstructure,ashighlightedbyrecentsimulations 87,27 may helpexplainthedifferencesinPLemissionobservedbetweendifferentsurfactants anddifferentresearchgroupsforthesamesurfactant. 62,63,54 Someresearchers alreadyhavedevelopedprocessestocontrolthesurfactantstructurearoundSWNTs, includinginsitupolymerizationofpolyvinylpyrrolidonePVP, 52 electrolytetuning 28 andcombiningsurfactantsfordensitygradientcentrifugation. 27 Ourgrouprecently demonstratedthatorganicsolventscanalterthesurfactantshellaroundSWNTs. 51 Here IexploretheeffectofSWNTprocessing,suchasowthroughmicrochannels, 88,89,90 37

PAGE 38

onthesurfactantstructureandPLemissionofSWNTs.Thehighshearenvironment inthechannelresultsinsignicantincreasesinPLintensityafterowingthroughthe channel.Thesechangespersistformonthsandcannotbeattributedtoshear-induced alignment 91 ordisaggregation.Photoluminescence,absorbance,andRamanspectroscopy suggestthatshearingaSWNTsuspensionhelpsannealthesurfactantshellaroundthe nanotube.Thealteredsurfactantstructureenhancesthestabilityofthesuspension andimprovesthePLintensity,especiallyforthelargestdiameterSWNTs.Wealso demonstratethatthisprocesscaneliminatediscrepanciesinthePLintensityofdifferent suspensions. 2.2Methods 2.2.1Dispersion Nanotubesuspensionswerepreparedwith60mgofSWNTsRiceHPR162.3 andmixedwith200mLofa1wt%SDSaqueoussolution.SDSwaspurchasedfrom Sigma-Aldrichandusedasreceived.Thesuspensionwasthenmixedwithahigh-shear homogenizerIKAT-25Ultra-Turraxat13,000rpmfor2handultrasonicatedina cuphornMisonixS3000at130Wfor10min.Subsequently,thesuspensionwas ultracentrifugedat20,000rpmBeckmanCoulterOptimaLKfor5htoremove nanotubebundles. 2.2.2MicrochannelFabrication Microuidicchannelswerefabricatedusingsoftlithography 92 withpolydimethylsiloxane PDMSasthestructuralmaterial.Briey,a4inchsiliconwaferwaspatternedwiththe negativephotoresistSU-82015Microchemusingaprintedtransparencyasthe shadowmask.A1/10ratiomixtureofPDMScuringagentandbaseSylgard184;Dow Corningwaspouredoverthepatternedsiliconwaferandcuredbyheatingat353K for2h.ThecuredPDMSwaspeeled-offfromthesiliconwaferanddicedtoobtain individualchips.Thechannelswerebondedirreversiblytoglassslidesbybringingthe channelandtheglassslideintointimatecontactafterabrieftreatmentofthePDMS 38

PAGE 39

chipswithoxygenplasmaAnatechSCE600Asher.Thechannelswere2cmlongwith acrosssectionalareaof20 60 m 2 2.2.3ShearingofSWNTsuspensions TheSWNTsuspensionswereshearedusingaCouettegeometryinanARESLS-1 strain-controlledrheometerTAInstruments.Themaximumattainableshearratein therheometerwas4000s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 .Foreachshearrate,thesuspensionswereshearedfor5 min.Ontheotherhand,meanratesofshearashighas1.72 10 5 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 wereachievedby owingthesuspensionthroughmicrochannelsatdifferentowrates.Asyringepump Pump11PicoPlus;HarvardApparatuswasusedtocontroltheowoftheSWNT suspensionthroughthemicrochannelwithowratesbetween5and4000 L/hr.Itis importanttoremarkthattherateoffailureofthechannelincreasesrapidlywhentheow rateishigherthan2500 L/hr.Forward,reverse,andrandomsweepsoftheshearrate wereusedtoverifythattheobservedspectralchangesatagivenshearratewerenot duetopre-shearingwithinthetubingleadingtothemicrouidicchannel. 2.2.4Dielectrophoresis Theelectrodeswerefabricatedusingthelift-offtechnique.Briey,a100nmlayer ofAuwassputteredonaglassslidethatwaspreviouslypatternedwiththeAZ9260 photoresist.A10nmlayerofCrwasusedastheadhesionlayer.Aftermetaldeposition, theelectrodeswereobtainedbydippingtheglassslideinacetonetoremovethe photoresist.AdropletoftheSWNTsuspensionwascastontotheelectrodesandthen anelectriceldV p-p ,10MHzwasappliedfor1mintominimizesolventevaporation. Thedropletwasblownoffwithastreamofnitrogengas,andthesubstratewasrinsed withethanolandde-ionizedwatertoremovethesurfactantfromthedepositedSWNTs. ThedepositedSWNTswerethenanalyzedwithRamanspectroscopy. 2.2.5Characterization AllSWNTsuspensionswerecharacterizedbyVis-NIRabsorbanceandNIR-uorescence spectroscopyusinganAppliedNanoFluorescenceNanospectralyzer Houston,TX 39

PAGE 40

withexcitationfrom662and784nmdiodelasers.Aminimumof200 LofSWNT suspensionwasneededtoanalyzetheefuentfromthemicrochannel.Twoorthree channelswereusedinparalleltoreducetheexperimentaltimewhenusingsmallow ratesshearrates.RamanspectrawererecordedusingaRenishawInviaBioRaman withexcitationfroma785nmdiodelaser. 2.3Results SuspensionsofSDS-coatedSWNTswerepumpedthroughmicrouidicchannels, suchasthoseshowninFigure2-1a,atdifferentratesofow.Theefuentwascollected andcharacterizedwithuorescenceandabsorptionspectroscopy.ThePLemission spectrafortheinitialSDS-SWNTsuspensionsshowninFigures2-1band2-1cpresent high-intensityandwell-denedpeaksdueto E 11 transitionsfromspecic n m SWNT types,whicharecharacteristicofwell-dispersedsemiconductingSWNTs.After processingtheSWNTsuspensionsthroughthemicrouidicchannel,theintensityof theemissionpeaksincreasessignicantlyasshowninFigures2-1band2-1c.These increasesinintensitydependontheowratewhilethepeakwidthdoesnotchange. Thepressure-drivenowthroughthemicrouidicchannelgenerateshighrates ofshearasshowninFigure2-1d.Themeanrateofshearisapproximately3.0 10 4 and10.7 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 fortheowratesof700and2500 L/h,respectively.TheSWNT suspensionswerealsoshearedinarheometerusingaCouettegeometry.Shearingat arateof1s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 changedthePLemissionintensitymarginally,asshowninFigures2-1e and2-1f.However,alargerintensitychangeisobservedwhenshearingat0.2 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 AdirectcomparisonofPLintensitiesofSWNTsprocessedinthemicrouidic channelandtherheometercannotbemade.WithintheCouettecell,onlyasmall fractionoftheuidvolumeresidesintheshearinggap;hence,mostoftheSWNTsdo notexperienceanyshear.Nonetheless,theincreaseinintensityqualitativelysupports theconceptthatshearaffectsthePLemission. 40

PAGE 41

Figure2-1.aPDMSmicrochannelsusedtoshearSDS-SWNTsuspensions.PL emissionspectrafromSWNTsthatowedthroughthemicrouidicchannels areshownfordifferentowratesormeanshearrateswithb662andc 784nmexcitation.dCartoonrepresentingtheparabolicvelocityproleand shearingofthesuspensionowingthroughthemicrochannel.PLemission spectraareshownforSDS-SWNTsuspensionsshearedinarheometer usingaCouettegeometrywithe662andf784nmexcitation.gThe intensityofthe,5and,7SWNTspeciesateachshearrateinthe microchannelisnormalizedbyitsintensityintheinitialnonshearedsample PL /PL 0 andplottedasafunctionofshearrate. Inbothshearingexperiments,theSWNTswiththesmallestdiameterse.g.,,5 typesshowsmallintensityincreaseswhilethepeaksofthenanotubeshavingthe largestdiameterse.g.,,7types,whicharepredominantlyexcitedat784nmFigure 2-1c,showmoresignicantincreases.ThechangesinPLemissionintensityofthe ,5and,7SWNTtypesareplottedinFigure2-1goverawiderangeofshear rates.ThePLintensityofeach n m typenormalizedbyitsinitialemissionintensity isusuallygreaterfollowingshear.AsseeninFigure2-1g,theemissionintensityof the,7nanotubescanmorethandoubleaftershearing.Most n m typesalsoshow awell-denedmaximuminthePLintensity.Althoughthepositionsofthemaxima varywitheachsuspensionandSWNTtype,itistypicallylocatedat 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 .The 41

PAGE 42

PLintensitythenslowlydecaysashighershearratesareapplieduntilitplateausat approximately3 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 fortheSWNTsoflargerdiameter.ForsmalldiameterSWNTs atlargeshearrates,theintensityoftheshearedsuspensioncanbeequalto,orlower than,theintensityoftheinitialsuspension.Notethatincreasingthemeanrateof shearsimultaneouslydecreasestheresidencetimeofSWNTsinthemicrochannel; consequently,theobservationsofthechangesinPLintensitymaydependonnot onlythemeanshearrate,butalsothelengthofthechannel.However,experiments investigatingtheeffectofresidencetimehaveyettobeperformed. 2.4Discussion 2.4.1Shear-InducedAlignment SWNTsarehighlyanisotropicparticleswithopticaltransitionsthatarehighly polarizedalongthenanotubemainaxis. 93 TheopticalanisotropyofSWNTshasbeen studiedbyaligningnanotubeswithmechanicalstretchingofpolymerlms, 94 electric elds, 95 anduidow. 96 Weismanandco-workersusedtheopticalanisotropyof SWNTstomeasurelengthdistributionsasthenanotubeswerealignedbyshearow. 91 Theauthorsreportedemissionintensitiesthatdoubledasthenanotubesaligned inthesuspension.Similarly,SWNTsprocessedintherheometerormicrouidic channelscouldalsoalignundertheappliedshear.However,thePLemissionfrom thesuspensionsinFigure2-1weremeasuredaftercollectingatleast200 Lofthe suspensionfromthemicrouidicchannels,whichtookseveralhoursforsomeshear ratesowrates.ThesesuspensionsarediluteaccordingtothecriteriaofDoiand Edwardssince nL 3 0.6 ,where n isthenumberdensityand L istheaveragelength oftheSWNTs.Consequently,thetimerequiredforBrownianforcestorandomizethe orientationsoftheensembleofalignednanotubesscaleswiththeinverserotational diffusivity, 1 = D r 97 ofasinglerod.Thoughsuspendedinadilutesolutionofsurfactants, therotationaldiffusivityoftheSWNTsismodeledasrigidrodsofhighaspectratio 42

PAGE 43

suspendedinaNewtonianuid 97 togive D r = 3 k B Tln L = d L 3 where k B T isthethermalenergy, istheviscosityoftheuidmedia.001Pa sfor water,with L and d beingthelengthandeffectivediameteroftherod,respectively. 98,99 Assumingavaluebetween2and5nmfortheeffectivediameterincludingsurfactant shelland500nmforthelengthofananotube,arandomizationtimeof 10 = D r is estimatedtofallbetween51and60msatroomtemperature. 99,100,101 ThePLspectraweretypicallymeasuredmorethantwohoursaftershearingand, asdiscussedfurtherbelow,theincreasesinPLintensitypersistforweeks.Hence, alignmentisnotresponsiblefortheincreasesinPLintensityofSWNTsaftershearing. 2.4.2DisaggregationofSWNTBundles Thecreationofadditional,isolatedSWNTsfromtheshear-inducedbreak-upof bundlesisonepossibleexplanationfortheincreaseinPLintensityaftershearing. However,nosignicantchangesareobservedtoeithertheabsorbanceorRaman spectra,whichprobetheaggregationstateofSWNTsuspensions.Figure2-2ashows thevis-NIRabsorbancespectraofSWNTsuspensionsshearedatratescorresponding tothemaximumandplateauregionsinFigure2-1g.Thespectrashowthatthe absorbancechangesaftershearingareminorlessthan 5%atanyshearrate. SlightpeakshiftsareobservedinsomeshearedSWNTsuspensions;however,these changeswerenotsystematic.Further,thepeakshiftsareoftenred-shiftedasshown forthespectracorrespondingtoSWNTsuspensionsshearedat0.08 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 inFigure 2-2a.Redshiftstypicallyindicatethatthesuspensionhasaggregated. 55 However,this suspensionexhibitssomeofthehighestPLintensityincreasesasshowninFigure 2-1g.Furthermore,theSANSdatashowninChapter4demonstratethatwelldispersed SWNTsareobtainedusingthemethodexplainedinSection2.2.1. 43

PAGE 44

Figure2-2.aAbsorbanceandbRamanRBMspectraex=785nmofaSWNT suspensionaftershearingataverageratesof0.08 10 4 and10.7 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 TheRamanspectraarenormalizedtotheGpeak.cDigitalimageofthe electrodesusedfordielectrophoreticdepositionofSWNTs.Digitalimagesof theelectrodes afterdepositionfromdshearedande noncentrifugedsuspensions.Theshearrate.4 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 waschosen wherethesuspensionshowedthemaximumintensityincreases.Noticethe aggregatesofSWNTsafterdepositionofthenoncentrifugedsuspensionin e.fRamanspectraex=785nmoftheinitial,sheared,and noncentrifugedSWNTsuspensionsdepositedontheelectrodesby dielectrophoresis.Theexcitationwavelengthonlyprobessemiconducting SWNTs. ThenormalizedRamanspectraoftheSWNTradialbreathingmodesRBMsalso shownoevidenceofchangestotheaggregationstateaftershearingasshowninFigure 2-2b.TheRBMpeakforthe,2SWNTtypeisverysensitivetoaggregationand shiftsto 270cm )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 whenbundledwithothernanotubes.Therefore,thispeakprovidesa measureofthedegreeofaggregationinSWNTsuspensions. 102 AsseeninFigure2-2b, 44

PAGE 45

theaggregationpeakhasnochangesatanyshearrate.Thereisasmallincreaseinthe intensityoftheRBMmodesfromthe,5and,7nanotubes,whereastheintensity ofthe,3nanotubesremainsconstant.Theintensityincreasesofthe,5and ,7nanotubesfollowasimilartrendasseeninFigure2-1gwiththehighestintensity occurringaftermoderateshearrates.Theseincreasescouldindicatesomechanges totheaggregationstate;however,researchershaveshowntheintensityofthe,3 SWNTshouldincreasemoresignicantlythaneitheroftheseSWNTtypesifbundles areseparatedintoindividualSWNTs. 103 ThefactthattheRBMintensityfromthe,3 SWNTremainsconstantandtheotherSWNTsshowonlyslightincreasesinintensity suggeststhatifdisaggregationoccurs,itmustbeaminoreffect.Theintensityofthe RBMsisalsosensitivetoquenchingeffects 103 andwillbediscussedfurtherbelow. Toprobetheaggregationstatefurther,thedielectrophoresisproceduredeveloped byKumataniandWarburtonwasusedtocharacterizethesuspensions. 104 The frequencyMHzoftheappliedelectriceldischosensometallicSWNTsare attractedtotheelectrodeswhilesemiconductorsarerepelledorexperienceamuch smallerattractiveforce. 105 Hence,semiconductingSWNTsaredepositedonthe electrodesonlyifaggregatedwithmetallicSWNTs,asshowninFigure2-2c-e.After depositionofanoncentrifugedsample,largeaggregatesareevidentontheelectrodes Figure2-2e.Theextentofaggregationcanthenbemeasuredbyselectivelyprobing thesemiconductingSWNTswithRamanspectroscopy.TheRamanspectraofSWNTs depositedfromtheinitialnoshear,sheared =0.4 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 ,andnoncentrifuged SWNTsuspensionsareshowninFigure2-2f.Asexpected,theRamanspectrafor noncentrifugedsuspensionscontainingbundledSWNTsshowverystrongRBMand GpeakstypicalofdepositedsemiconductingSWNTs.Ontheotherhand,theRBM andGpeaksareabsentfromboththeinitialandshearedSWNTsuspensions.Once again,theseresultssuggesttheinitialSDS-SWNTsuspensionsarewell-dispersed 45

PAGE 46

andthestateofaggregationdoesnotchangesignicantlyafterowingthroughthe microchannel. 2.4.3FluidFlowSegregationorRemovalofImpurities Researchershaveobservedthatshearowwithinaconninggeometrycan segregatesemidilutesuspensionsofcarbonnanotubesbylength. 96 Ifsegregationwere occurringinthemicrouidicchannels,thePLintensityoftheSWNTsuspensionscould beaffectedbynonlinearchangestothequantumyieldsofnanotubesasafunction oflength. 84,48 However,developmentofthelength-dependentdistributionforadilute suspensionofBrownianrodsrequiresalargeresidencetime. 106 Estimatesbased upontheparametersfortheseexperimentsindicatethattherodswouldhavetoreside withinthechannelforatleast90s,whereastheactualresidencetimeis1satmost. Therefore,itisunlikelythatsegregationoccursundertheseowconditions.Inaddition, similarPLintensityincreasesareobservedinchannelswithdifferentdimensions, acriticalparameterthatwouldaffectthesegregationifoccurring.Furthermore,no evidenceofSWNTremovalineitherthechannelorintheabsorbancespectrasee Figure2-2aisobserved. Similarargumentswouldsuggestthattheremovalofotherimpurities,suchasmetal catalysts,fullerenes,andothercarbonaceousmaterialwouldbeminimal.Thisisfurther supportedbyinductivelycoupledplasmaatomicemissionspectrometryICP-AES, whichshowsthattheamountofironinthesuspensiondoesnotchangeafterowing throughthemicrochannelnotshown.Whileselectiveinteractionwiththewallsofthe channelcouldoccur,therheologydatainFigures2-1eand2-1fstillshowPLincreases. Channelsconstructedofglass/parafnalsoshowedsimilarPLintensityincreases,ruling outselectiveinteractionwiththechannelwalls. 2.4.4SurfactantReorganization AnotherplausibleexplanationfortheincreaseinPLintensityofspecic n m SWNT typesistherearrangementofthesurfactantshellsurroundingthenanotubes.Our 46

PAGE 47

grouprecentlydemonstratedtheabilitytomanipulatetheSDSshellsurrounding SWNTsusingimmiscibleorganicsolvents. 51 Aftermixingwithorganicsolventsand allowingittoevaporate,thesurfactantassumedaconformationthatprovidedbetter protectiontoextrinsicquenchingevents,suchasprotons. 69,51,71,52,77,82 Thischemical annealingprocesswasfoundtoincreasethePLintensityofthelargestdiameter SWNTsbyasmuchas175%withoutchangestoeitherabsorbanceorRamanspectra. Inaddition,Doornandcollaboratorsrecentlyreportedthattheadditionofsaltsto SDS-SWNTsuspensionscanalterthesurfactantstructuresurroundingthenanotubes, resultinginincreasesofPLintensityaswellasnarrowingofspectrallinewidthof SWNTspecies. 28 Inasimilarfashion,theshearappliedtoSWNTsuspensionsfrom eithermicrouidicchannelsorarheometermayalsorearrangethesurfactantstructure, resultinginreducedquenchingfromextrinsicfactors.Infact,shearhasbeenshown tocausephasetransitionsandinducechangesintheself-assemblingpropertiesof micellarsystems. 107,108 Forexample,micellesofCTATcetyltrimethylammonium p-toluenesulphonate,whicharecylindrical,cangrowwhenshearedandreturntothe initialstateonceshearinghasstopped. 109,55 IfshearcanimprovetheabilityofthesurfactanttoprotectSWNTsfromquenching protons,thesuspensionshouldmaintainhighPLintensityasthepHoftheaqueous phaseisaltered.Figure2-3showsthatPLintensityfromthe,7and,3SWNT typesintheinitialSDS-SWNTsuspensiondecreasebyatleastanorderofmagnitude atacidicpHvalues,similartopreviousobservationsdataforthe,5and,5SWNT typesareavailableinAppendixA. 51,52,103 However,thePLintensityofsheared aqueousSWNTsuspensionsarelesssensitivetothequenchingeffectsoftheacid medium.AllSWNT n m typesshowimprovedPLintensityatallpHvalues.The effectisevenmoredramaticforthesmallerdiameternanotubese.g.,,5and,3 SWNTsatthelowestpHvalues.Atthesehighlyacidicconditions,theintensityfrom 47

PAGE 48

Figure2-3.pH-dependentPLintensityofthea,7andb,3SWNTspeciesinan initialSDS-SWNTsuspensionandaftershearing =0.4 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 theshearedSWNTsuspensionremainsclosetoanorderofmagnitudehigherthanthe initialsuspension. Theresistancetoquenchingsuggeststhatshearingprovidesabetterprotective layeraroundthenanotube.AsimilarphenomenonwasexploitedbyMcDonaldetal. 110 toselectivelyquenchtheemissionofSWNTspeciesbasedondifferencesinsurfactant adsorption.BetterprotectionoftheSWNTfromprotonsalsoexplainstheintensity increasestotheRBMsofthe,7and,5nanotubesintheRamanspectraseen inFigure2-2b.Thesephononmodesarebroadenedandlessintenseathighproton concentrations. 77 Asthesurfactantreorganizesaroundthenanotubeaftershearing,the SWNTminimizesitsinteractionswithprotonsresultinginmoreintenseRBMpeaks. TheresultsshowninFigures2-12-3conrmourpreviousobservationthatSDSis capableofsuspendinglargediameterSWNTs;however,SDSisoftenassembledina 48

PAGE 49

structurethatfailstoprotectthenanotubefromextrinsicquenching. 51 Throughall-atom MDsimulations,TummalaandStriolofoundthatSDSmoleculesdonotspreadevenly ontheSWNTsurfaceand,athighsurfacedensity,prefertomaximizeinteractions amongthemselves. 53 Hence,partsofthenanotubesidewallremainexposedtowater. Further,thereareseveralreportsthatsuggestSDShaslowersurfacecoverageor weakerinteractionswithlargediameterSWNTs. 110,111 Thesedifferencesininteraction strengthmayhelpexplainthehighersensitivityofas-preparedSDS-coatedSWNTs, especiallythelargediameterSWNTs,toquenchingevents.Aftershearingtheinitial SWNTsuspensionsbyeitherthemicrouidicchannelorrheometer,thesurfactant shellacquiresadifferentcongurationthatprovidesbetterprotectiontotheextrinsic factors.Thisprocessmaybefacilitatedbyshear-inducedchangestothemicellar structure. 107,108 2.4.5ImprovedDispersion TheformationofamorerobustsurfactantstructuresurroundingSWNTsshould alsoprovidelargerrepulsiveforcesasnanotubesinteractwitheachotherinsolution. 112 Therefore,thereorganizedsurfactantstructuresshouldimprovedispersionstability.The PLintensityoftheinitialandsheared =0.4 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 SDS-SWNTsuspensions wererecordedoverthecourseof14weekstomeasurethedispersionstability,as showninFigure2-4datafor,5and,7SWNTtypesareavailableinAppendix A.Interestingly,theenhancedPLemissionintensitypersistsforseveralmonths.The PLintensityofallSWNT n m typesfromtheshearedsuspensionisalwayshigher thanthatoftheinitialsuspension.Typically,theintensityoftheshearedsuspension decaysslowerthantheinitialsuspension.Forexample,the,3SWNTsintheinitial suspensiondecreaseinintensityby50%Figure2-4a.However,thesesameSWNT typesonlydecreaseby15%intheshearedsuspensions.Ontheotherhand,the behaviorofthelargediameternanotubesintheinitialandshearedsuspensionswas different.Theintensityfromthelargestnanotubese.g.,,5SWNTsintheinitial 49

PAGE 50

Figure2-4.Time-dependentPLintensityofa,3andb,5SWNTspeciesinan initialSDS-SWNTsuspensionandaftershearing =0.4 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 suspensionstayedapproximatelyconstantfortherst2weeks,andthentheintensity increaseduntilreachingaplateauFigure2-4b;thisobservationisattributedtoenergy transferfromthesmallernanotubeslargebandgaps. 74,111 Afterthesixthweek,the intensitystarteddecreasingagainuntilanotherperiodofslowincreasesisobserved aftertheeleventhweek.Insharpcontrast,theintensityfromthelargediameterSWNTs intheshearedsuspensionstabilizedafterapproximately2weeks.Therelatively constantintensityfromthelargediameterSWNTsisattributedtothebettersurface coverageofthenanotubeswithSDS.Theresultsshowthatshearinduceslong-term changestotheSWNTPLandagingdynamicsofthesuspension. 50

PAGE 51

2.4.6DiscrepanciesamongSuspensions Althoughthesamedispersionprocedureisalwaysusedtopreparethesuspensions, occasionallySDS-SWNTsuspensionsshowlargevariationsintheinitialPLintensity. Forexample,suspensionsAandBcurves1and2inFigure2-5ashowalarge deviationintheinitialintensityofthesuspension.Themostsignicantdifferencesare observedforthelargediameternanotubeslowestemissionenergy.Insuspension B,thelargediameterSWNTsarecompletelyabsentfromthespectra,andthesmall diameterSWNTsshowsignicantdecreasesinintensitywhencomparedtosuspension A.TherstimpressionfromthePLspectraisthatsuspensionBisapoorsuspension thatwasunabletoadequatelydispersethenanotubes.AsdiscussedinAppendixA, thissuspensionwasbath-sonicatedagaininanattempttodisperseSWNTsbetter.This showedmarginalimprovementstothePL,butthelargestdiameterSWNTsremained absentfromthespectra.Indeed,researchershaveoftenconcludedthatthelargest diameternanotubesdonotsuspendwellinSDSsuspensions. 70 Itcouldalsobeargued thatSWNTsinsampleBhaveahigherdefectdensitycausingSWNTstobeoptically inactive 113 orthatlargerSWNTsareinanoxidizednonuorescingstate. 79 However, afterbothsuspensionswereshearedatidenticalshearrates =0.4 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 ,the PLemissionforboththepoorandgoodsuspensionsincreasetothesameintensity foralltheSWNTspecies.Themostfascinatingincreasesareseenforthepeaks correspondingtothe,5and,7speciesinsuspensionB,whichwereabsent fromtheinitialSWNTsuspensionspectra.Thesespeciesshowa20 increasein intensityafterthesuspensionwassheared,asshowninFigure2-5b.Thesimilarityin thenalPLemissionFigure2-5aandabsorbanceFigureA-3,AppendixAspectra forbothsuspensionsisalsosignicant.Again,theseresultssuggestSDSiscapable ofsuspendinglargediameterSWNTs;however,extrinsicfactorsoftenquenchtheirPL. Theeffectsofextrinsicfactorscanbeminimizedbyreorganizingthesurfactantstructure 51

PAGE 52

Figure2-5.aPLemissionspectraex. = 784nmfrominitialandshearedSWNT suspensions.NoticethattheinitialPLintensityfromthelargerdiameter SWNTspecieslowerenergyinsuspensionBarealmostentirelyabsent fromthespectra.However,thePLemissionintensityofAandBisnearly identicalaftershearingeachsuspension.bIntensityratioPL /PL 0 ofthe ,5speciesinsuspensionBasafunctionofshearrate aroundthenanotubeeitherbyshearingthesuspension,asshowninFigure2-1,orby mixingwithorganicsolvents. 51 ThePLemissionintensityincreasesobservedinFigures2-1and2-5occurwithout signicantchangestotheabsorbance.Therefore,theensembleQYsofSWNTspecies hasincreasedconsiderablyaftershearing. Table1showsQYscalculatedforvariousSWNT n m typesinshearedsuspensions bycomparisontothedyeIR26seeAppendixA.Forsuspensionsthatareinitially bright,theQYsfortheselectedspeciesareapproximately0.4%,whicharesimilar 52

PAGE 53

Table2-1.Quantumyields%forselected n m speciesinshearedsuspensionswith PLintensitiesthatwereinitiallygoodandinitiallypoor.Seesuspensions describedinFigure2-5.Moredetailsofquantumyieldcalculationsaregiven inAppendixA n m initiallygoodQYinitiallypoorQY ,30.710.88 ,50.710.93 ,50.881.05 ,70.580.66 totheensemblevaluesreportedbyJonesetal. 85 AlthoughBlackburnetal.didnot reportQYsforSDS-suspendedSWNTs,theydidcommentthatQYswere5times lowerthaninsodiumcholatesuspensions 1%. 82 AsshowninTables1andA-2 AppendixA,shearinginitiallygoodsuspensionsresultsinsignicantincreasesto theaverageQY.AlthoughallSWNTspeciesshowedimprovedQYs,shearinghada moresignicanteffectonthelargerdiameterSWNTs.Forexample,the,3species initiallyhadaQYof0.54%thatincreasedto0.71%aftershearing.Ontheotherhand, the,7SWNTsinitiallyhadaQYof0.13%andshoweda4-foldincreaseto0.58%. TheresultsareevenmoreimpressiveforthelargediameterSWNTsininitiallypoor suspensions.TheaverageQYof,5speciesincreasesfrom0.02%to1.05%in theseSWNTsuspensions.Hence,shearedsuspensionsofSDS-SWNTsshowQYs thatarecomparabletoensemblemeasurementsperformedonseparatedSWNTs 72,114 orSWNTssuspendedinothersurfactants. 82 AninterestingfeatureoftheaverageQYs calculatedinTable1isthattheinitiallypoorsuspensionsconsistentlyyieldedbetternal valuesfortheQYthanthegoodsuspensions.Theseresultshighlighttheimportance inunderstandingthesurfactantstructuresurroundingSWNTssincethePLspectraof as-preparedSWNTsuspensionscouldprovidemisleadinginformation. 70,79,86 2.4.7ShearingEffectsonotherSWNT-SurfactantSystems Preliminaryexperimentswereperformedwithothersurfactants,suchasSDBS, SC,andcetyltrimethylammoniumbromideCTAB.Enhancements 10%are observedwithSDBS-SWNTsuspensions,whereasSCandCTABsuspensions 53

PAGE 54

haveslightdecreasesinPLintensity,especiallyatsmallshearratesnotshown. However,additionalexperimentsareneededtobetterunderstandthebehaviorofthese SWNT-surfactantsystemsandwhySDSsystemsshowsuchalargeincreaseinPL intensity. 2.5Conclusions Thehigh-shearforcesthatSWNTsexperiencewhenowingthroughmicrochannels caninducesignicantincreasestothePLofSDS-suspendedSWNTs.ThesePL increasesaremostsignicantforthelargediameterSWNTs,resultinginPLintensities thatcanbeasmuchas2-20timeshigherthantheinitialsuspension.TheQYsofthe shearedsuspensionsarenear1%,makingthemcomparabletotheQYsobtained withsodiumcholate.ThepersistenceofthePLandtheincreasedstabilityofthe suspensionssuggestthatalignmentofSWNTsisnotresponsible.Inaddition, Ramanandabsorbancespectroscopyanddielectrophoresisexperimentsshowthat disaggregationofbundlednanotubescannotexplaintheseincreases.Therefore,these PLintensityincreasesareattributedtosurfactantreorganizationaroundSWNTs.The abilitytoalterthesurfactantstructuredemonstratespotentialopportunitiestofurther improvePLbytailoringthesurfactantstructurearoundSWNTs. 54

PAGE 55

CHAPTER3 SOLVATOCHROMICSHIFTSOFSINGLEWALLCARBONNANOTUBESIN NONPOLARMICROENVIRONMENTS 3.1Introduction SWNTshavecharacteristicopticaltransitionenergies E ii associatedwiththeir n m chirality.The E ii associatedwitheachSWNT n m typewereassignedby dispersingnanotubeswiththeaidofsurfactants, 55,62,63 enablingcharacterization throughabsorbanceandphotoluminescencePLspectroscopy. 64,65 Typically,sodium dodecylsulfateSDS,sodiumdodecylbenzenesulfonateSDBS,andsodiumcholate SCareusedtosuspendSWNTs,althoughothersurfactantsarealsoused. 55,62,63,61 WhilethesestudieshavefacilitatedtheapplicationandunderstandingofSWNTs, theeffectoftheenvironmentsurroundingthenanotubesonSWNTpropertiesisstill notwellunderstood. 115,116,76,117,118,119,120,121,122,123,124 Theinabilitytosuspend SWNTsinknowndielectricenvironments,suchasorganicsolvents,complicatesthese studies. 125 Severalstudieshaveusednanotubessuspendedacrosstrenchestostudy environmentaleffectsonSWNTPLemission; 120,121,122 however,mechanicalstrain andchargetransfercanaffectthemeasurements. 122,126 Thelimitedunderstandingof environmentaleffectsonSWNTpropertiesisalsopartiallyduetotheunknownorpoorly characterizedsurfactantstructure, 53,87 whichmakesitdifculttoassessthedielectric constantofthemediasurroundingSWNTs. ThepolarizabilityofindividualSWNTsisalsoimportanttomanyapplicationsas wellasthephotophysicalandphotochemicalresponseofSWNTs.Thepolarizabilityis highlyanisotropicwiththelongitudinalcomponent k atleastanorderofmagnitude largerthanthetransversecomponent ? .Therefore,thelongitudinalpolarizabilityis thedominantcontributionandthedipolemomentisorientedalongthenanotubeaxis. Severalmodelshavebeendevelopedtocalculatethelongitudinalpolarizability.These modelssuggestthelongitudinalpolarizabilityisafunctionofthenanotuberadius R andbandgapenergyofSWNTsbutdisagreeontherelationship.Researchershave 55

PAGE 56

foundtheoreticalrelationshipswherethelongitudinalpolarizabilityisproportionalto either R = E ii 2 1 = E ii 1/ E ii 2 ,or R 2 127,128,129,130 WhilethereareseveraltheoreticalstudiesonthepolarizabilityofSWNTs,thereare fewexperimentalstudies. 76,131 Measuringthephotophysicalpropertiesofmolecules, suchassolvatochromicshifts,canprovideinformationabouttheexcitedstatesofthe moleculesandeventheirpolarizability.ChoiandStranousedsolvatochromicshiftsof SWNTstodeterminethatlongitudinalpolarizabilityvarieswith 1 = R E ii 76 However, thesemeasurementswerebasedonrelativelyfewsystems,includingsurfactant suspendedSWNTs,whosedielectricenvironmentsarestillnotwellcharacterized. Thetheoreticalandexperimentalstudiesdescribedaboveshowthattheexact formofthelongitudinalpolarizabilityremainsunknown.Experimentalmeasurement ofsolvatochromicshiftsinavarietyofsolventsremainsagoodapproachtostudy polarizability,providedthatSWNTscanbesuspendedinseveralsolventsofknown dielectricconstant.However,dispersingSWNTsinorganicsolventsremainsa challengingtask. 125 Recently,ourgrouphasobservedthatmixingasuspension containingSDBS-coatedSWNTswithimmiscibleorganicsolventsinducessolvatochromic shiftsordielectricscreeningeffects,whicharedependentonthesolvent. 51 ThePL spectraofSDBS-SWNTsuspensionsmixedwitho-dichlorobenzeneODCBshowed identicalshiftstothePLfromSWNTssuspendedinonlyODCBi.e.nosurfactant orwater. 125 Thesimilarityinthepeakpositionsindicatesthatthehydrophobiccore ofthesurfactantformsanemulsion-likemicroenvironmentofODCBaroundthe nanotube.Here,weusethesenewmicroenvironmentsaroundSWNTstoinvestigate thephotophysicalpropertiesofSWNTsinavarietyofsolvents.Whilethepresenceof theorganicsolventcanmodifythesurfactantassemblyaroundtheSWNTsurface 51 andsubsequentlyaffectthePLemission, 82,71,77 thepresentworkusesSDBSasthe surfactant,whichhasminimalrearrangementofthesurfactantstructureinthepresence oforganicsolvents. 51 Thesolvatochromicshiftsfollowtheexpectedbehaviorfroma 56

PAGE 57

polarizablesoluteinapolarizablesolvent.ThePLintensityisshowntobeverysensitive topolarsolvents.Relationshipsaredevelopedtodescribethesolvatochromicshiftsand PLintensityofdifferent n m typesinlowdielectricmedia. 3.2Methods 3.2.1Reagents Deionizedwaterwasusedinallexperiments.Thesurfactant,sodiumdodecylbenzene sulfonateSDBS,wasobtainedfromSigma-AldrichSt.Louis,MO,USAandusedas receivedTechnicalGrade.HiPcoSWNTswereobtainedfromRiceUniversityRice HPR145.1andusedasreceived.Thenonpolarsolventsexaminedinthisstudywere obtainedfromSigma-Aldrichhexane%,heptane%,cyclohexane%, carbontetrachloride.9%,1-chlorohexane%,1,6-dichlorohexane%, 2-heptanol%,3-heptanol%,1-chlorobutane.8%,2,6-dichlorotoluene %,3,4-dichlorotoluene%,o-dichlorobenzene%,FisherScientic Pittsburgh,PA,USAchloroform.8%,p-xylene.7%,toluene%andFluka benzene.5%,1,3-dichlorobenzene%.Allsolventswereusedasreceived. 3.2.2AqueousSWNTSuspensions AllexperimentsusedSDBSsincethissurfactantdidnotshowanyquenching effectsuponmixingwithorganicsolvents,asobservedforSDS-SWNTsuspensions. 51 Aqueoussuspensionsofnanotubeswerepreparedbymixing20mgofrawSWNTswith 200mLofaSDBSsolutionwt%.High-shearhomogenizationIKAT-25Ultra-Turrax for1.52handultrasonicationMisonixS3000for10minwereusedtoaiddispersion. Afterultrasonication,themixturewasultracentrifugedat20,000rpmfor5hBeckman CoulterOptimaL-80K. 3.2.3SolventMicroenvironmentsaroundSWNTs ImmisciblesolventswereaddedtoeachSWNTsuspensionsolvent:watervolume ratioof0.5andmixed.Thedielectricconstant,refractiveindex,dipolemoment,and inductionpolarizationforeachsolventisgiveninTable3-1.Themixturewasshaken 57

PAGE 58

vigorouslyfor30swithavortexstirrer.Awhiteemulsionphaseimmediatelystarted tocoalesceaftershaking.Afterwaitingfor1.5htoreachsteadystate,theexcess organicsolventwasthencarefullyremovedfromtheaqueousSWNTsuspension withoutpromotingfurtheremulsication. Table3-1.Propertiesofsolventsusedtoformmicroenvironmentsaround SWNTs. 132,133,134 SolventsRelativeDielectricConstantRefractive Index Dipolemoment/D Inductionpolarization f 2 hexane1.891.370.000.369 heptane1.921.390.000.383 cyclohexane2.021.430.000.411 carbontetrachloride2.231.460.000.430 p-xylene2.271.500.000.455 benzene2.281.500.000.455 toluene2.391.500.380.455 2,6-dichlorotoluene3.361.550.830.483 chloroform4.811.451.040.424 1-chlorohexane6.101.421.940.404 3-heptanol7.071.421.710.404 1-chlorobutane7.281.402.050.390 1,6-dichlorohexane8.601.462.030.430 3,4-dichlorotoluene9.391.553.000.483 2-heptanol9.721.421.710.404 o-dichlorobenzene10.121.552.500.483 3.2.4SWNTCharacterization Theaqueousphasewascharacterizedbyvis-NIRabsorbanceandNIR-uorescence spectroscopyusinganAppliedNanoFluorescenceNanospectralyzer Houston,TX withexcitationfrom662and784nmdiodelasers.Althoughtheorganicsolventshave someabsorbancebandsintheNIRregion,theireffectonthespectralpropertieswere determinedtobeminor.Thisconclusionissupportedbythefactthatnochangesto theabsorbancespectraofSWNTsareobserved,whichisconsistentwithaverysmall volumeofsolventinthesystem.Likewise,absorptionofSWNTphotoluminescence PLbythesolventwasconcludedtohaveminimaleffectsonthePLemissionintensity. Allsolvatochromicshiftsaredescribedrelativetotheiremissionenergyinair,i.e. E solvent 11 )]TJ/F24 11.9552 Tf 9.299 0 Td [( E air 11 58

PAGE 59

3.2.5DeconvolutionofPLSpectra SWNTPLspectracontainsinformationonthe E 11 interbandtransitionsforeach n m SWNTtype.AllPLspectraweredeconvolutedintotheirrespectivebandsfor each n m typeusingtheAppliedNanospectralyzersoftwareHouston,TX.The deconvolutionroutineusesVoigtprolesforeach n m peak.Changestotheposition E 11 andwidthofeach n m peakwerelimitedto0.1%and3%,respectively,for eachiterationtopreventthemisidenticationofpeaks.Afterseveraliterations,the meanstandarddeviationMSDbetweenthesimulatedandexperimentaldatapoints wassmallerthan0.005MSD<0.005.FigureB-1showsatypicalspectra,whichis deconvolutedintothepeakscorrespondingtoeach n m type.Thepeakpositionof every n m typeofSWNTsafterdeconvolutionisusedinthenonlinearoptimization model. 3.3ResultsandDiscussion 3.3.1SolvatochromicShiftsofSWNTSpectrainVariousSolvents Figure3-1aandbshowthePLemissionspectraofSWNTsexcitedwitha662and 784nmlaser,respectively,whenmixedwithavarietyofnonpolarsolventswithdifferent dielectricconstants,refractiveindices,andpolaritiesseeTable3-1.Allfeaturesof thespectrashowpeakshiftsaswellasintensitychangeswitheachsolvent.SWNT suspensionsmixedwithhexaneshowthehighestPLintensityandemissionenergy.In contrast,SWNTsuspensionsmixedwithODCBhavelowerPLintensityandred-shifted peakpositions.Figure3-1cshowssimilarpeakshiftsintheabsorbancespectraof SWNTsinthesamesolvents. Toseethespectralchangesmoreclearly,Figure3-2aandbshowthePLspectra correspondingtoonlythe,6and,5SWNTtypeswhenmixedwiththenonpolar solvents.Theguresshowchangesinboththeintensityandpeakpositionasthe solventisalteredfromhexanetoODCB.Thesedifferentenvironmentsshowthat,in general,theemissionenergytendstored-shiftandtheintensitydecreasesasthe 59

PAGE 60

Figure3-1.PLemissionspectraofSWNTsexcitedata662andb784nmandc absorbancespectraofSWNTsinnonpolarmicroenvironments.Allspectrain acareoffsetforclarityandarrangedbydielectricconstant. dielectricconstantsolventpolarityisincreased.However,therearesomedeviationsin thetrends.Forexample,2,6dichlorotolueneand2heptanolarered-andblue-shifted morethansolventswithsimilardielectricconstants,respectively.Inaddition,the peakwidthsgetbroaderinhighpolaritysolvents,suchas3,4dichlorotolueneand ODCB.Theabsorbancespectraforboththe E 11 and E 22 transitionsofthe,6SWNT areshowninFigure3-2canddandalsoshowsimilarspectralshiftsasthesolvent polarityisincreased.However,the E 11 .85.4eVtransitionsoftheSWNTsaremore sensitivetothesolventenvironmentslargershiftsthanthe E 22 .4.1eVtransitions, inagreementwithpriorobservations. 76 ThepeakshiftsinthespectraofFigure3-1and3-2indicatethatSWNTsare experiencingdifferentenvironmentswhenmixedwithnonpolarorganicsolvents.SWNTs havenonetdipolemomentbutarehighlypolarizable.Priortophotoexcitation,the solventinducesasmallreactioneldontheSWNTsdielectricscreeningandvice versa.Theeldchangesthesolvationenergyassociatedwithstabilizingtheground state,yieldingcharacteristicabsorptionenergiessolvatochromicshiftsindifferent solvents.TheexcitedstateofSWNTsshouldhavealargerdipolemomentafterphoton 60

PAGE 61

Figure3-2.PLspectra E 11 emissionforthea,6andb,5SWNTspeciesand c E 11 andd E 22 absorbanceforthe,6SWNTtypemeasuredin microenvironmentsofvariousnonpolarsolvents.Thelinesarethepeak positionforSWNTsinhexanemicroenvironments. absorption,generatingaeldthatforcesthesolventstructuretorearrangeorrelax aroundtheexcitedstateSWNTdipole.AccordingtotheFranckCondonprinciple, solventreorientationistooslowtobeobservedduringabsorptionbutcanbeobserved inPLemission.Eachsolventrespondsdifferentlytotheexcitedstatedipole,resultingin solvatochromicshiftsofthePLemissionofSWNTsthatmaydifferfromthoseobserved inabsorbance.TheobservedspectralchangesinFigure3-1and3-2suggestthe measuredsolvatochromicshiftsareduetotheformationofasolventmicroenvironment encapsulatingtheSWNTs. 51 Figure3-3ashowsaschematicofthemicroenvironmentsformedaroundSWNTs basedontheseresultsaswellaspriorspectroscopicobservationssummarized inFigure3-3b. 51 TheinitialsurfactantstructurehastheSDBSmoleculesonthe sidewalloftheSWNTs.However,aftermixingwithnonpolarsolvents,thehydrophobic regionbetweentheSWNTsidewallandthesurfactantswells,creatingasmall 61

PAGE 62

Figure3-3.aSchematicofthenonpolarsolventmicroenvironmentsformedaround SWNTs.Thecut-outsectionshowsthesolventlayerencasesthenanotube, providinganapproachtosystematicallyaltertheenvironmentsurrounding SWNTs.bComparisonofPLemissionspectraex.=662nmofSWNTs surroundedbySDBSsurfactant,ODCBmicroenvironments,and pureODCB.Dashedlinesshowtheemissionenergyforselect n m species.Notethatthepositionsforandareconsistent conrmingthatthespectroscopicresultsformicroenvironmentsaround SWNTsaresimilartodispersionsinpuresolvent.Thesimilarityinpeak positionsalsoindicatesthatthesurfactanthasaminoreffect. 51 microenvironmentofsolvent.Itwasshownpreviouslythatthesurfactantstructurecould bedifferentinthepresenceofsomeorganicsolvents;however,thesechangesoccurred whenSDSwasusedtosuspendtheSWNTsandareeasilyobservedbysignicant quenchingofthePL. 47 SolventsthatyieldedsimilarbehaviorintheSDBSSWNTs studiedhere,suchasm-dichlorobenzeneandethylacetate,wereexcludedfromfurther analysis.Further,ourpriorworkshowedthatPLemissionenergiesofSWNTsinan ODCBmicroenvironmentwereidenticaltoSWNTsinODCBaloneseeFigure3-3b. Therefore,thethicknessoftheshellisassumedtobelargeenoughthattheeffectsfrom thesurfactantandwatercanbeneglected.ThefactthatthePLspectrareturntotheir 62

PAGE 63

originalpeakpositionsandintensityafterthesolventisremovednotshownsuggests thataggregationisminimal,asdiscussedpreviously. 47 3.3.2CharacterizingSolvatochromicShiftsofSWNTs Researchershaverelatedtheenvironmentaleffectsorsolvatochromicshifts E 11 ofSWNTstothedielectricconstantoftheenvironment. 118,120,121 Theseshiftswere foundtobeapproximatelylinearwiththedielectricconstantinlowdielectricmedia, suchasthesolventsusedinthisstudy. 120 Figure3-4ashowsthemeasuredPLpeak shiftsfromthedatainFigure3-1relativetoairformultiple n m typesasafunction ofdielectricconstant.Althoughthe,7nanotubehasareasonablelinearttothe dielectricconstantofthesolvent,mostothernanotubes,especiallysmalldiameter nanotubes,showconsiderablescatter,suchasthe,5and,3.Indeed,Miyauchi etal. 120 13foundsignicantdeviationsfromlinearityforSWNTPLmeasuredinhexane =1.89 Figure3-4.Solvatochromicshiftsofvarious n m SWNTtypesinnonpolarsolventsasa functionofadielectricconstantandbsolventinductionpolarization, f 2 .Shiftsobservedforthe,7SWNTtypebyOhnoetal. 121 arealso plottedforcomparison.Notethattherearemultipledatapointsthatare indistinguishableinb.cGeneralizedsolvatochromicshiftsasafunctionof theinversediameteroftheSWNTs/d.Notethatthe,1and,4 SWNTtypeswereexcludedfromcbecauseofdifcultiesassociatedwith deconvolutingthemfromthespectra. 63

PAGE 64

Thesolventsusedfortheseexperimentsareallconsideredtobenonpolarbecause oftherelativelylowdielectricconstants. 135 Therefore,thesolvatochromicshiftsforeach solventareprimarilyinuencedbythepolarizabilitypolarizabilityinteractionsgiven by: 136,137 E 11 = E solvent 11 )]TJ/F24 11.9552 Tf 11.955 0 Td [( E air 11 = )]TJ/F42 11.9552 Tf 9.298 0 Td [(C solvent 11 a 3 f solvent )]TJ/F43 7.9701 Tf 6.587 0 Td [(air where C solvent isauctuationparameterassociatedwithSWNTdispersionforces, 11 isthechangeinpolarizabilityofSWNTsbetweenthegroundandexcitedstates, isa shapefactorfortheSWNT, isaparameterassociatedwiththelocationoftheSWNT dipoleinthevolume a 3 ,and f isthesolventinductionpolarizationdescribedbythe Onsagerpolarityfunctions, f 2 =2 2 )]TJ/F24 11.9552 Tf 12.961 0 Td [(1 = 2 +1 ,where istherefractive indexofthesolvent. 136,137 Figure3-4bshowstheshiftsplottedasafunctionof f 2 Asexpectedfromequation3,thesolvatochromicshiftsinFigure3-4btendtobe linearforall n m types,yieldingsignicantlybetter R 2 valuesthantherelationshipwith dielectricconstant. Thesolvatochromicshiftsrangefromapproximately20meVforSWNTswith thelargestdiameterstoashighas100meVforsmalldiameternanotubes.Adirect comparisonwiththedatafromOhnoetal. 121 forthelargerdiameter,7,,5, ,3,and,1SWNTtypesinhexaneandchloro-formshowexcellentagreement towithinafewmeV.Thesimilarityinspectralshiftsstronglysupporttheformationof asolventmicroenvironmentaroundtheSWNTs,asshowninFigure3-3.Asseenin Figure3-4c,theaveragesolvatochromicshiftforeach n m typeinallsolventsvaries linearlywiththeinversediameter/dofthenanotubes.Nodependencyofthespectral shiftwiththe n-m mod3valueisobserved.AleastsquarestofthedatainFigure 3-4cyieldsanexpressionfortheaverageshiftasafunctionofnanotubediameter, E 11 eV =0.0760.119 = d nm.Thissimpleexpressionexcludesspecicsolvent effectsbutprovidesanestimateoftheanticipatedsolvatochromicshiftofSWNTsinlow dielectricmedia. 64

PAGE 65

Figure3-5.ThePLemissionaintensityandbnormalizedintensityforselected n m SWNTtypesasafunctionofdielectricconstant.Theintensitiesina areoffsetforclarity.Thedashedlinesarepowerlawts.Theopencirclesin aindicatepointsexcludedfromthepowerlawtandthedatainb.The intensitiesinbwerenormalizedtotheintensityofthe n m typemixedwith hexane. Whilethepeakpositionsaredescribedbetterbytheinductionpolarizationof thesolvent,thePLemissionintensityisclearlyafunctionofthedielectricconstantof thesolvent.AsseeninFigure3-5a,thePLintensitydecaysrapidlyasthedielectric constantincreases.Thetrendissmoothandcontinuousuntil 0.5 ,wherethedata startstoscatterasthedielectricconstantcontinuestoincrease.However,itisstill 65

PAGE 66

clearthattheintensitycontinuestodecayasthedielectricisincreased.Someofthese intensitydifferencescouldbeduetoconcentrationdifferencesbetweensuspensions orremovalofasmallamountofaggregatesattheinterface. 46,47 Duetothelarge deviationsassociatedwiththeintensityof1-chlorohexaneand3-heptanol,thesewere excludedfrompowerlawtstotheintensity.TheintensityoftheSWNTPLemission wasapproximatelyproportionalto 0.5 Theintensitiesofeach n m SWNTtypewerenormalizedtotheintensitymeasured inhexane.Thedatashowsconsiderablescatterbutitisagainclearthatintensity decreasesaredependentontheSWNTdiameter.Forexample,thePLemission intensityofthe,5,,5,and,5SWNTtypesdecreasebyapproximately35%, 40%,and65%inhigherdielectricmedia,respectively.TheseresultsshowthatSWNT PLemissionisverysensitivetopolarsolvents.Forexample,alargeportionofthe decreaseinPLintensityforeach n m SWNTtypeisobservedinchloroform,which has 0.5 .ThesensitivityofSWNTPLemissiontopolarenvironmentscouldhave signicantimplicationsinunderstandingthePLemissionintensityfromaqueous suspensions,whereminoramountsofwaterincontactwiththeSWNTsidewallmay signicantlyaffectPLintensity.Indeed,researchershavefoundthatbettersurfactant layersresultinhigherPLintensity. 51,82,52 TheStokesshiftisdenedasthedifferencebetweenpeakpositioninabsorbance andPLemissionspectraandisrelatedtothesolventreorganizationthatoccursduring photonabsorption.AllofthenonpolarmicroenvironmentsyieldedsmallStokesshifts ofapproximately1meVseeAppendixB.ThesesmallStokesshiftsindicatethat boththegroundandexcitedstatesoftheSWNTsareequallystabilizedinthenonpolar solvents.Thesimilarityinsolvationcouldindicatethateitherthedifferenceofthedipole momentbetweenthegroundandexcitedstatesissmallorthatsolventrelaxationhas notoccurredpriortoradiativerecombination.Thislattereffectmightbeexpectedfor adipoleorientedalongthelengthofthenanotube,whichwouldrequirethesolvent 66

PAGE 67

tomovelongdistancesinordertoalignwiththeexcitedstatedipole.Notrendswere observedfortheStokesshiftbutthiscouldbeattributedtothesmallvaluesandthe errorassociatedwithdeconvolutingthespectra. 3.3.3NonlinearOptimizationModelforApproximatingtheLongitudinalPolarizability The E ii transitionsforeach n m SWNTtypefromeitherthePLorabsorbance spectracanbeusedtoestimatetheSWNTpolarizabilityoritsfunctionalform;however, thePLdataisusedbecausethe n m peaksarebetterresolved.Previously,Choiand Strano 76 concludedthatthedifferenceinSWNTpolarizabilityusedinequation3is primarilydeterminedbythelongitudinalpolarizabilityoftheexcitoni.e. 11 11, k SeveralresearchgroupshavedescribedthelongitudinalpolarizabilityofeachSWNT n m typebyafunctionoftheform, 11, k = R a E b 11 ,where isaconstant, a and b are integers, R istheSWNTradius,and E 11 isthebandgapasmeasuredinair. 127,128,130 Asdescribedabove,thereisnoconsensusontherelationshipi.e.constants a and b forthepolarizability.Ifthevolumeassociatedwiththeexcitedstateisassumedtobea spherewitharadiusequivalenttotheSWNTradius,thenequation3becomes: E 11 = )]TJ/F42 11.9552 Tf 9.299 0 Td [(C solvents R a )]TJ/F25 7.9701 Tf 6.586 0 Td [(3 E b 11 f Theconstants C solvent ,and arecombinedintooneconstant D solvent forsimplicity. Sinceequation3isvalidforallsolvents,themeasuredsolvatochromicshiftsinFigure 3-1canthenbecomparedtothisequationtodeterminetheglobalvariables, a and b andthesolventspecicvariables, D solvent Table3-2.Criterionforassigningtheweightingfactorsforbothintensityandamountof overlapwithotherpeaks. WeightingFactorAreaOverlapIntensitynm excitation Intensitynm excitation 1<50%>0.4>0.08 0.5between50and 100% between0.2and0.4between0.04and 0.08 0.25100%<0.2<0.04 67

PAGE 68

AllPLspectrainFigure3-1weredeconvolutedintotheirrespective n m peaks. The E 11 transitionsforeach n m SWNTtypeinairweresubtractedfromthe E 11 recorded fromeachspectrumineachsolventtogivethesolvatochromicshift E 11 .Theoptical transitionsinairwerebasedontheequationdevelopedbyChoiandStrano. 76 Asshown inFigureB-1,therecanbeconsiderableoverlapinthedeconvolutedPLspectraforeach n m SWNTtype.Forexample,the,6SWNTtypeshowninFigureB-1awhasgood intensityandlittleoverlapwithother n m SWNTtypes.Ontheotherhand,the,2 SWNTtypeinFigureB-1bwshowsconsiderableoverlapwithother n m typesand lowintensity.Therefore,thecondencelevelofeachpeakinthedeconvolutedspectra wasevaluatedtoassessitsrelativeimportanceinregressionanalysis.Toaccount forthevaryingcondencelevelsineachdeconvolutedpeak,weightingfactorswere assignedbasedontherelativeintensity w Intensity andamountofoverlap w Overlap in eachsolvent.Thistechniqueisoftenusedinnonlinearoptimizationtoaccountforthe relativeimportanceofdata. 138,139 AsshowninTable3-2,boththeintensityandoverlap werebrokenupintothreecategoriesandassignedvaluesof1,0.5,or0.25.Figure B-1wshowssomeexamplesofthosepeaksassociatedwitheachcategory.Thesetwo weightingfactorsweremultipliedtogethertogettheweightingfactorslistedinTable3-3 forallSWNTtypesinallsolvents.Thegeneraltrendshowninthetableisthatthe,6 and,5spectrahavethehighestcondencewhilethe,4and,1havethelowest ineachsolvent. Aconstrainednonlinearoptimizationmodelwasthenformulatedusingthe generalizedreducedgradientGRGmethod. 138 Anobjectivefunction waschosen tobethesquaredsumofresidualsmultipliedbytheweightingfactors w i where E 11 arethemeasuredshiftsrelativetoairorthecalculatedvaluesfromequation3.These residualsweresummedoverall n m typesandsolventstoobtainthenalobjective function: min = allsolvents X alln,m X w n m Intensity w n m Overlap E n m 11, measured )]TJ/F24 11.9552 Tf 11.955 0 Td [( E n m 11, calculated 2 68

PAGE 69

Table3-3.Combinedweightingfactors w = w Intensity w Overlap assignedforeach n m typeasafunctionofsolvent. solvent,5,3,5,6,2,4,1,6,5,1,3,2,5,7 hexane0.50.5110.50.250.250.250.50.2510.2511 heptane0.51110.50.250.250.250.50.510.2511 cyclohexane0.51110.50.250.250.250.50.510.2511 carbontetrachloride0.50.5110.50.250.250.250.50.2510.2511 p-xylene0.50.5110.50.06250.250.250.50.12510.2510.5 benzene0.50.5110.50.1250.250.50.50.12510.2511 toluene0.50.5110.50.1250.250.50.50.12510.2510.5 2,6-dichlorotoluene0.50.50.510.50.1250.06250.50.06250.062510.250.250.5 chloroform0.50.5110.50.1250.250.50.1250.1250.50.250.251 1-chlorohexane0.50.5110.50.1250.250.50.1250.12510.250.51 3-heptanol0.50.5110.50.250.250.250.50.2510.2511 1-chlorobutane0.50.5110.50.1250.250.50.1250.12510.250.51 1,6-dichlorohexane0.50.5110.50.1250.250.50.1250.06250.50.250.251 3,4-dichlorotoluene0.50.5110.50.06250.250.50.250.06250.250.125-0.0625 2-heptanol0.50.5110.50.250.250.250.50.2510.2511 o-dichlorobenzene0.50.5110.50.1250.250.50.1250.06250.250.1250.25Average0.500.560.971.000.500.160.240.390.340.190.840.230.730.84 69

PAGE 70

Theobjectivefunction,equation3,wassolvedwithalloptimizationvariables constrainedtovaluesbetween-10and10toaidoptimization.Theglobalparameters a and b hadtheadditionalconstraintsofbeingintegers.Centraldifferenceswereused tocalculatethepartialderivativesoftheobjectivefunction.Thesearchdirectionwas determinedateachiterationbyusingtheconjugatemethod.Theobjectivefunctionwas consideredtoconvergeoncetherelativechangebetweeniterationswaslessthan 10 )]TJ/F25 7.9701 Tf 6.587 0 Td [(7 Aminimumoffourdifferentinitialstartingpointsfortheparameterswereusedtoobtain theglobalminimumratherthanlocalminima. Inadditiontothesolutionusingdatafromallsolvents,solutionsusingdatafrom individualsolventswerefoundindependentlytoobtaintheparameters a b ,and D solvent Table3-4and3-5showtheresultsfromalloptimizationcalculations.Someoptimization modelshadminimathatwereweakfunctionsoftheparameters a and b .Inotherwords, thevaluesof a and b couldbechangedby 1 withoutaffectingthevalueoftheobjective function.Forexample,theoptimizationofonlyheptanegavesolutionsof a = )]TJ/F24 11.9552 Tf 9.298 0 Td [(2 b = )]TJ/F24 11.9552 Tf 9.299 0 Td [(3 and a = )]TJ/F24 11.9552 Tf 9.298 0 Td [(3 b = )]TJ/F24 11.9552 Tf 9.299 0 Td [(4 .However,oneofthesesolutionscoincidedwiththe globaloptimizationsolution.Therefore,thisdatasetwaslistedinTable3-4.Nearly alloptimizationsolutionsyieldedsolutionsof a = )]TJ/F24 11.9552 Tf 9.299 0 Td [(2 and b = )]TJ/F24 11.9552 Tf 9.298 0 Td [(3 .AsseeninTable 3-4,aglobalsolutionwithallweightingfactorssetequalto1gaveadifferentsolution. However,asdescribedabove,thissolutionisdiscardedbecauseofthedifferencein dataqualityforeach n m SWNTtype.Thisanalysissuggeststhatthelongitudinal polarizabilityoftheexcitonshouldhavetheform, 11, k = R )]TJ/F25 7.9701 Tf 6.586 0 Td [(2 E )]TJ/F25 7.9701 Tf 6.587 0 Td [(3 11 .Insertingthisrelation intoequation3yieldsanexpressionforthesolvatochromicshift: E 11 = )]TJ/F42 11.9552 Tf 9.299 0 Td [(D solvent R )]TJ/F25 7.9701 Tf 6.586 0 Td [(5 E )]TJ/F25 7.9701 Tf 6.587 0 Td [(3 11 f or E 11 E 3 11 = )]TJ/F42 11.9552 Tf 9.299 0 Td [(D solvent f = R 5 Figure3-6ashowsthettoequation3forthesolvatochromicshiftofSWNTs whensurroundedbychloroformmicroenvironments.Thelineartyields R 2 =0.98 indicatingthattheparametersforthepolarizabilityexpressionshowexcellentagreement 70

PAGE 71

withtheexperimentaldata.AsshowninFigure3-6b,thelargestresidualsareforthe ,1,,4,and,6SWNTtypes;however,twoofthesenanotubes,the,1and ,4,havethesamediameters,soalargeerrorisassociatedwiththedeconvoluted peakpositions,asindicatedinTable3-3. Figure3-6.SolvatochromicshiftsofSWNTsinchloroformmicroenvironmentsrelativeto air.aThepolarizabilityparameters a and b determinedfromthenonlinear optimizationyieldalinearexpressionwithblowresiduals. Theseresultsdeviatefrompriorreports,whichobtainedvaluesof a rangingfrom 2to-1and b from0to-2. 127,128,129,130 However,notethatthosesolutionsthatdeviate fromtheotherresultstendtobethemostpolarsolvents.Thechangesinintensityand peakpositionforthesesystemsinducemoreerrorinpeakassignment.Ontheother hand,someofthesesystemsgiveaandbconstantscomparabletothoseobtained before. 76,128 71

PAGE 72

3.4Conclusions MultiplesolventsformmicroenvironmentsaroundSWNTs,inducingsolvatochromic shiftsthatrangefromapproximately25to100meV.Theshiftsscalewellwiththe solventinductionpolarization, 2 ,asexpectedforinteractionsofapolarizableSWNT withapolarizablesolvent.Thesolventmicroenvironmentsshowthesensitivityof SWNTPLtoslightchangesintheirenvironment.Achangeofthedielectricconstant from2to5couldresultinadropinPLintensityofmorethan50%,whichcould havesignicantimplicationsonthemeasuredintensityofpoorlycoatedSWNTsin aqueousenvironments.Changestotheenvironmenthavethemostsignicanteffect onthepeakpositionofthesmallestdiameterSWNTsandtheintensityofthelargest diameterSWNTs.Aconstrainednonlinearoptimizationmodelwasusedtostudy thepolarizabilitychangeswitheachsolventmicroenvironment.Theresultsyielda longitudinalpolarizabilityoftheform 11, k / 1 = R 2 E 3 11 72

PAGE 73

Table3-4.Optimizationresultsforallttedparameterswhensolvedasindividualsystemsandcollectively ParameterAllsolventsHexaneHeptaneCyclohexaneCarbon Tetrachloride p-xyleneBenzeneToluene2,6-Dichlorotoluene a -2-2-2-2-2-2-2-20 b -3-3-3-3-3-3-3-3-1 D hexane 4.14.1 D heptane 3.93.9 D cyclohexane 3.63.6 D carbontetrachloride 3.63.6 D p-xylene 3.63.6 D benzene 3.63.6 D toluene 3.63.6 D 2,6-dichlorotoluene 3.613.3 D chloroform 3.9 D 1-chlorohexane 4.0 D 3-heptanol 3.8 D 1-chlorobutane 4.1 D 1,6-dichlorohexane 3.9 D 3,4-dichlorotoluene 3.9 D 2-heptanol 3.8 D o-dichlorobenzene 3.9 73

PAGE 74

Table3-5.Optimizationresultsforallttedparameterswhensolvedasindividualsystemsandcollectively ParameterChloro1-Chloro3-Heptanol1-Chloro1,6-Dichloro3,4-Dichloro2-Heptanolo-DichloroAllsolvents -form-hexane-butane-hexane-toluene-benzene w i =1 a -2-2-3-2-20-3-1-3 b -3-3-4-3-3-1-4-2-4 D hexane 2.2 D heptane 2.1 D cyclohexane 1.9 D carbontetrachloride 1.9 D p-xylene 1.9 D benzene 1.9 D toluene 1.9 D 2,6-dichlorotoluene 1.8 D chloroform 3.92.0 D 1-chlorohexane 4.02.1 D 3-heptanol 2.02.0 D 1-chlorobutane 4.12.2 D 1,6-dichlorohexane 3.92.0 D 3,4-dichlorotoluene 14.41.9 D 2-heptanol 2.02.0 D o-dichlorobenzene 7.52.0 74

PAGE 75

CHAPTER4 SWELLINGTHEHYDROPHOBICCOREOFSURFACTANTSUSPENDEDSINGLE WALLCARBONNANOTUBES:ASANSSTUDY 4.1Introduction EnablingtheuseofSWNTsinnewtechnologicalapplicationsrequireslearninghow toprocessthem.ThestrongvanderWaalsinteractionsbetweenSWNTsmakestheir processingdifcult.Whileorganicsolventsarebeingstudiedtopromotetheexfoliation andtruedissolutionofSWNTs,theexfoliationcapacityoforganicsolventsisstillnot optimal. 43,44 StrongacidicmediahavealsoshownacapacitytodissolveSWNTs,but handlingsuchharshmaterialsisproblematic. 140 Forthesereasons,SWNTsarestill commonlydispersedinwaterthroughcovalent 141 ornon-covalentfunctionalization. 63 Suspensionthroughnon-covalentfunctionalizationisachievedusingsyntheticpolymers e.g.,PluronicandTritonX, 63 bio-polymerse.g.,DNA, 142 peptides 143 andgumarabic 61 andsurfactantse.g.,SDS,SDBSandbilesalts. 54,55,144 ThesuspendingagentanditsassemblyontheSWNTsurfaceaffectstheinterfacial propertiesofSWNTs.Thenatureoftheinterfacebetweenthemediumandthe nanotubescanhavesignicantconsequencesonthespectroscopicpropertiesof SWNTsasshownChapter3,theirabilitytopartitionintoorganicinterfaces 47 andthe densityofthesurfactant-SWNTcomplex. 27 Thepermeabilityoftheinterfacedetermines howeasilyquenchingmolecules,suchasprotons,caninteractwithexcitonsonthe SWNTsurfaceand,consequently,reducethePLquantumyield. 71 Forexample,DNA suspendsSWNTsextremelywell,butthehighlyexposedsurfaceofDNA-suspended SWNTstowatercausestheirPLquantumyieldtobeverylowseeFigureC-12.Onthe otherhand,individualandbrightSWNTscanbeproducedbyusingsuspendingagents, suchasanaliphaticdodecylanalogofavinmononucleotide,thatassembletightlyon theSWNTsurface. 145 TheabilityofSWNTssuspendedinwatertopartitiontoorganic interfaceshasbeenexploitedtoremoveSWNTbundlesinamethodcalledinterfacial trapping. 47 However,thepartitionofSWNTstotheinterfaceishighlydependenton 75

PAGE 76

thenatureofthesuspendingagent.Forexample,themethodworkswellwithSWNTs suspendedwiththenon-ionicsurfactantgumarabic,butitismorechallengingfor SWNTssuspendedwithionicsurfactants.Similarly,itisnowrecognizedthatthe effectivenessofdensitygradientultracentrifugationDGUishighlydependentonne tuningofthesurfactant-SWNTinterfacesby,forexample,modifyingthepHandtheionic strengthofthesuspensionorusingmixturesofsurfactants. 27 ThesurfactantmoleculesonthesurfaceofaSWNTaremobileandreorganizein responsetodrivingforces.Specically,thecongurationofSDSmoleculesonSWNTs hasbeenshowntoberesponsivetohighshearstressesinmicrouidicchannels asdescribedinChapter2,changestotheionicstrength, 28,146 andinteractionwith organicsolvents. 51 Ourgrouphasdevelopedandtakenadvantageofamethodthat benetsfromtheabilityofsurfactantmoleculestosolubilizenonpolarorganicsolvents, resultinginalocalizedsolventenvironmentaroundtheSWNTs.SWNTssuspendedin organicsolventsshowedthesamesolvatochromicshiftsasthesurfactant-suspended SWNTsafterswellingwithorganicsolvents.Encasingthenanotubeswiththesesmall solventenvironmentsenabledasystematicstudyoftheeffectofsolventpolarityon PLemissionenergy,whichwasdiscussedinChapter3.Wehavealsousedthese solventenvironmentsasmicro-reactorstoperformin-situpolymerizationontheSWNT surface. 147 Finally,thechangestothesurfactantstructureinducedbythepresenceof thesolventwereusedtostudytheretentionmechanismofSDS-suspendedSWNTs onagarosebeads,asshowninChapter5.Whiletheselocalizedenvironmentsaround SWNTshavebeenusedforfundamentalandappliedstudies,questionsremainabout thesizeofthesolventdomains,theiruniformityanddistributiononthesidewalls,andthe effectofsolventswellingontheaggregationstateofthenanotubes. Small-angleneutronscatteringSANSisanexcellenttechniquetoprobestructure ofsoftmaterialsatthenanoscale.Atsmallscatteringangles,thesmallwavelengthof neutronsprobeslengthscalesfrom1nmto1 m. 148,149 Becauseneutronsinteractwith 76

PAGE 77

otheratomsthroughshortrangenuclearinteractions,theyareabulkprobethatdeeply penetratessamples.Inaddition,theabilitytochangethecontrastfactorbetweenthe scatteringcentersandthemediumthroughselectivedeuterationmakesthestudyof multi-componentorcomplexsamplestractable.IntheeldofSWNTs,SANShasbeen usedtomeasurethedispersionqualityofSWNTs, 150,32 theoptimalconcentrationof dispersingagents,theinuenceofdepletionforces, 151 andalsoenabledstudieson theexfoliatingpowerandassemblyofpolymersonSWNTs. 152,153 Theformationof3D networksathighSWNTconcentrations 154,155 andin-situpolymerization 144,156 processes alsohavebeenstudiedthroughSANS.Ofparticularimportancetothisstudy,Yurekli etal.studiedtheassemblyofSDSonSWNTswithSANSandfoundthattheSDS assemblyischaracterizedbylackoflong-rangeorder. 56 4.2Methods 4.2.1Reagents DeionizedwaterH 2 OanddeuteriumoxideD 2 Owereusedtopreparesurfactant solutionsaswellasSWNTsuspensions.Thesurfactant,hydrogenatedsodiumdodecyl sulfateH-SDS%,anditsdeuteratedcounterpart,sodiumdodecylsulfated 25 D-SDSatom%d,werepurchasedfromSigma-AldrichSt.Louis,MO,USAand usedasreceived.HiPcoSWNTswereobtainedfromRiceUniversityRiceHPR162.3 andusedasreceived.Benzene.9%,o-dichlorobenzene%andD 2 O.9atom %dwerepurchasedfromSigma-Aldrich.Allsolventswereusedasreceived.Table4-1 showsthescatteringlengthdensities ofallthereagentsusedintheexperiments. Thevaluesof forthedifferentmoleculeswerecalculatedfromtheexpression, = N A M X i b i where isthemolecule'sbulkdensity, M isitsmolecularweightand b i isthecoherent neutronscatteringlengthofnucleus i 157 77

PAGE 78

Table4-1.Scatteringlengthdensitiesofallthereagentsusedinthisstudy Reagent 10 cm )]TJ/F25 7.9701 Tf 6.587 0 Td [(2 H 2 O-0.56 D 2 O6.39 H-SDS0.387 D-SDS6.704 H-Benzene1.18 H-ODCB2.35 SWNT 155 4.9 4.2.2AqueousSWNTSuspensions Aqueoussuspensionsofnanotubeswerepreparedbymixing30mgofrawSWNTs with100mLofeither35mMH-orD-SDSsolution.High-shearhomogenization IKAT-25Ultra-Turraxfor30minandultrasonicationMisonixS3000for10min Wwereusedtoaidalldispersions.Duringhomogenizationandsonicationof D 2 O-SWNTsuspensions,thebeakerandsonicatorcupwerecoveredwithparalmto avoidexchangeofD 2 OwithatmosphericH 2 O.Afterultrasonication,themixturewas ultracentrifugedat20,000rpm gfor4husingaswingbucketrotorBeckman CoulterOptimaL-80K. 4.2.3MixingSWNTsuspensionswithinsolubleorganicsolvents SwellingofSWNTsuspensionswithorganicsolventswasperformedbyfollowing apublishedprotocol. 51 Briey,thesuspensionwasmixedwiththeorganicsolvent andvigorouslyagitatedfor30sinavortexstirrer.Thesolvent-swelledsuspensions werethenpreparedandlefttosettleovernighttoallowbulk-phaseseparationpriorto scatteringexperiments.Then,analiquotof0.5mlwascarefullywithdrawnfromthe aqueousphasecontainingswelled-SWNTtoavoidfurtheremulsication.Insomecases, thesolventwasremovedbyevaporationinairfor24h,whichwaspreviouslyshownto removethesolventencasingtheSWNTs. 51 4.2.4SWNTCharacterization SWNTsuspensionswerecharacterizedbyabsorbanceanduorescencespectroscopy usinganAppliedNanoFluorescenceNanospectralyzer Houston,TXwithexcitation 78

PAGE 79

from662and784nmdiodelasers.SANSmeasurementswereperformedonthe NG330minstrumentattheNationalInstituteofStandardsNISTCenterforNeutron ResearchNCNR.Thewavelengthoftheneutronbeamwas6.Threesample-to-detector distanceswereused,4and13.2mtoobtainascatteringvector Q rangefrom 0.003to0.4 )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 .Thesampleswereloadedintodemountabletitaniumcellswithquartz windowsthathadapathlengthof1mm.Multiplesamplesupto10wereplacedona samplerackforSANSmeasurements.Anemptycellandablockedbeamtransmission measurementswereusedtoconvertthedataintoabsolutescale. 4.2.5DataAnalysis TheindependentvariableinaSANSmeasurementisthemodulusofthescattering vector Q .Themagnitude Q isafunctionofthescatteringangle andtheneutron wavelengthasgivenby, Q = 4 sin 2 where istheneutron'swavelength.UsingthepreviousexpressionalongwithBragg's lawofdiffraction,onecanquantifylengthscalesinreciprocalspaceanddeterminethe characteristicsizeofscatteringcentersthroughthefollowingexpression, 148,157 d = 2 Q ThedependentvariableinaSANSmeasurementisthecoherentmacroscopic scatteringcrosssection,whichinabsolutescaleisreferredtoasthescatteringintensity. Thescatteringintensitycanbedecoupledintointraandinterparticlecontributions, and,inthecaseofanincompressiblesystem,itcanbemodeledas, I Q = N V V 2 p 2 P Q S Q + I Inc where )]TJ/F43 7.9701 Tf 6.903 -4.977 Td [(N V isthenumberdensityofparticleswhile V p and aretheparticlevolume andcontrastfactor,respectively. I Inc isaconstantthatdescribestheincoherent scatteringfromthesample.Thesingleparticleformfactor, P Q ,isadimensionless 79

PAGE 80

functionanddependsonthesizeandshapeofthescatteringparticles.Ontheother hand,thestructurefactor, S Q ,whichisadimensionlessfunctionaswell,dependson thedegreeoflocalorderandinteractionpotentialbetweenthescatteringparticles. ThemicellesintheSDSsolutionweremodeledaseithertri-axialoruniform ellipsoids.ThosemodelswerechosenbasedonpreviousSANSmeasurementson SDSsolutions. 158,159 PrvostandGradzielskifoundthatatri-axialellipsoidmodel wasnecessarytosimultaneouslytSANSandSAXSdataofSDSorCTABsurfactant solutions.However,atlowsaltconcentrations,thedatacanbettedtoauniform ellipsoid.Theformfactorforatri-axialellipsoidaxes a b ,and c iscalculatedthrough thefollowingexpression, P Q = Y V ell Z 1 0 Z 1 0 2 f Q [ a 2 cos 2 x 2 + b sin 2 x 2 )]TJ/F42 11.9552 Tf 11.956 0 Td [(y 2 + c 2 y 2 ] 1 2 g dxdy where Y isaconstant,thefunction isdenedas, x =9 sin x )]TJ/F42 11.9552 Tf 11.955 0 Td [(x x 3 2 andtheparticlevolume, V ell ,isgivenby, V ell = 4 3 abc Inthecaseofauniformellipsoid,theformfactorisgivenby, P Q = Y V ell ell )]TJ/F27 11.9552 Tf 11.955 0 Td [( solv Z 1 0 f 2 f Qr a [1+ x 2 v 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [(1] 1 2 g dx withthefunction f denedas, f z =3 V ell sin z )]TJ/F42 11.9552 Tf 11.955 0 Td [(z cos z z 3 2 and v = r a r b Forthetri-axialellipsoid,theaxesmustfullltheconditionthat a < b < c .Dueto theirsizeandsurfacecharge,SDSmicellescanbeconsideredtobemacroions.The structurefactorismodeledbyassumingSDSmicellesinteractonlythrougharepulsive 80

PAGE 81

potential.TheCoulombpotentialthatresultsfromthemutualinteractionofthemicelle's double-layerdeterminestherepulsivepotential. 160,161 Fitstothedatawerecalculatedin IgorProusingtheNCNRSANSanalysispackage. 4.3Results InasuspensionofSDS-coatedSWNTs,scatteringfromthefreesurfactantmicelles andtheH-SDS/SWNTcomplexisexpected.Figure4-1showsthescatteringfroman H-SDS-coatedSWNTsuspensionandanH-SDSsolutioni.e.,withoutnanotubes.At high Q ,thesignalfromtheSDSsolutionischaracterizedbyastructurepeak.05 )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 andashoulder.5 )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 typicalofionicsurfactantsolutions. 159,158 Arisingscattering intensityisobservedatthelow Q -range<0.02 )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 .WhenSWNTsaredispersed withtheaidofSDS,thereisnotasubstantialchangeinthescatteringintensityatthe high Q -range>0.02 )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 .ThescatteringintensityoftheSWNTsuspensionatthe high Q -rangeisslightlylowerthanthatfromtheH-SDSsolutiononlySDSmicelles andmonomers.ThelowerintensitymightbeduetotheremovalofSDSmolecules duringcentrifugationortheadsorptionofSDSmoleculesonSWNTs.However,thereis aremarkabledifferencebetweenthescatteringfromtheSDSsolutionandtheSWNT suspensionatthelow Q -range.Thesedifferencescanonlybeattributedtothepresence ofSWNTs. H-SDSsuspendedSWNTswerethenswelledwithH-benzeneandtheirscattering measured.Figure4-2showsthescatteringfromtheH-SDS/SWNT/H-benzenesample andtherespectiveH-SDS/H-benzenesolution.TheintensityoftheSWNTsuspension atlowandhigh Q valuesislargelyincreasedwhencomparedtotheinitialsuspension. Thestructurepeakalsoshiftstolower Q values.Likewise,thepeakpositionshiftsand thescatteringintensityatthehigh Q regimeincreasesfortheH-SDSsolutionsmixed withbenzene.However,thedifferencebetweenthescatteringsfromthetwosamples islarge.Surprisingly,therisingintensityatlow Q disappearsasthesolutionistreated withbenzene.Figure4-2bshowsthescatteringfromtheH-SDS/SWNT/H-ODCB 81

PAGE 82

Figure4-1.SANSscatteringintensityfromSWNTssuspendedin1wt%H-SDSandthe respectiveSDSsolution. Figure4-2.ScatteringprolefromH-SDSsolutionandSWNTsuspensionswelledwith abenzeneandbODCB.ThescatteringfromtheinitialSWNTsuspension wasplottedforcomparison. sampleandtherespectiveH-SDS/H-ODCBsolution.Similartothecaseofbenzene,the intensityoftheSWNTsuspensionatlowandhigh Q valuesincreaseswhencompared totheinitialSWNTsystem.However,changestothescatteringprolearedifferent. Whiletheintensityatlow Q decreasesforbenzene,itincreasessubstantiallywhenthe H-SDSsolutionismixedwithODCB. Afterswellingthemicellesorsurfactant-coatedSWNTs,thesolventswereremoved byevaporation.Theobjectiveofthisexperimentwastoobserveifthecontactwith 82

PAGE 83

Figure4-3.ScatteringintensityfromSWNTsuspensionsafterbenzeneandODCBhave beenevaporated.ThescatteringfromtheinitialSWNTsuspensionwas plottedaswellforcomparison. organicsolventscausedanyirreversibleeffectsontheSWNTsuspension.Therewas anincreaseintheincoherentbackgroundasindicatedbythesignalatlarge Q -values, afterthesolventwasevaporated.ThisbehavioristobeexpectedsinceD 2 Owould exchangewithatmosphericH 2 Oduringsolventevaporation.Interestingly,athigh Q values,itisobservedthatSWNTsuspensionsmixedwithsolventsreturntotheirinitial stateaftertheevaporationprocess.Itisobservedthatthesuspensionswelledwith benzenematchesthepeakpositionandintensityoftheinitialSWNTsuspensionatthe high Q -range.Thepeakpositionalsoreturnstohigher Q valuesforthesuspension swelledwithODCB.However,theintensitydoesnotcompletelymatchthatofthe initialsuspensionH-SDS/SWNT.Atthelow Q -range,theslopeofthescattering proledecreasesforbothofthesamplesafterevaporationascomparedtotheinitial suspensionandtheirsolvent-swelledcounterparts. 4.4Discussion 4.4.1H-SDSSolution Itiswell-knownthatSDSmoleculesself-assembleintoaggregatesmicellesonce thecriticalmicelleconcentrationCMCisreached. 162 Inwater,thechargesonthe SDSmoleculesdissociate.Therefore,micellescanbeconsideredtobemacroionsthat 83

PAGE 84

Table4-2.ParametersusedtotthescatteringfromtheH-SDSsolution.Theresults fromtheH-SDSsolutionsmodiedwithNaClarealsoshown.Thedatafor Q valuesabove0.05 )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 wasttedtoauniformellipsoidwiththestructure factorforascreenedCoulombpotential. ParametersH-SDS0.1MNaCl0.2MNaCl MinorAxis14.5514.4314.62 MajorAxis21.5525.6728.97 AggregationNumber, N agg 69.0097.00125.00 DegreeofIonization, z = N agg 0.200.170.064 2 = n 0.5 2.182.261.86 interactwitheachotherthroughCoulombicinteractions.Allthesefeaturesareevident intheSANSscatteringproleoftheSDSsolution.Atthehighest Q values,themajor contributiontoSANSintensitycomesfromtheformfactor,whichisafunctionofthe sizeandshapeoftheSDSmicelles.Thepeakat0.05 )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 ,whichcorrespondto130 inrealspace,isduetotheCoulombicinteractionamongmicelles.Thesefeaturescan bemodeledusinganoblateoruniformellipsoidformfactorandthescreenedCoulomb potentialformacroions. Figure4-4.FittothescatteringfromtheH-SDSsolutionandthescatteringfromH-SDS solutionsmodiedbyaddingNaCltonalconcentrationsof0.1and0.2M. Table4-2showsthegeometricalandphysicalparametersobtainedfromthets tothe1wt%H-SDSsolution.TheresultingcurveisshowninFigure4-4.Themicelles showminorandmajoraxesof14.5and21.5,respectively,whichareinagreement 84

PAGE 85

withpreviouslypublisheddata. 159,158 Fromthevolumeofthemicelles,anaggregation numberof69SDSmolecules/micellecanbecalculatedusingthemolecularvolume ofSDS,0.41nm 3 ,reportedbyPrvostandGradzielski. 158 Likewise,thisresultisin agreementwithvaluesofthemicellaraggregationnumberspreviouslyreportedforSDS inaqueoussolutions. 163,159 Itispossiblethattherisingsignalatlow Q isduetoclusteringornon-uniform distributionofSDSmicelles,aspreviouslysuggestedbyWangetal. 151 andextensively describedinthecaseofpolymersolutions. 164,165 Anexperimentperformedwitha solutionofD-SDSinsteadofH-SDSshowsthattheintensityathigh Q becomesat,as expectedbecausethemicellesarecontrast-matcheddatanotshown.However,the risingintensityatlow Q stillsappears.ThisbehaviorsuggeststhelowQ intensity hasitsoriginsindensityuctuationsofthesystemwhichareamanifestationof clustering. 149,165 Tofurtherprobethenatureofthesignalinthisregion,theionic strengthoftheH-SDSsolutionwaschangedbyaddingNaCl.Astheconcentrationof NaClincreases,Figure4-4showsthatthepolyelectrolytepeakdisappearsbecause thechargesonH-SDSmicellesarescreened.Likewise,theeliminationoftherising signalatlow Q indicatesthatCoulombicinteractionsplayasignicantroleincreating non-uniformdistributionofscatteringcentersinthesample.TheresultsinTable4-2also showthatthemicellesgrowinsizeandasymmetrywhenscreenedbyNaCl. 4.4.2SWNTSuspension Interpretationoftheresultsontheswellingwithorganicsolventsdependsonthe stateoftheinitialsuspension.TherehavebeenreportsthatquestiontheabilityofSDS toexfoliateandindividuallysuspendSWNTs. 70,62,166 Ithasalsobeenshownthatthe effectivenessofsurfactantstosuspendSWNTsdependsontheSWNTdiameter. 146 ThestudybyQuintonetal. 50 isthemostrelevanttoourparticularsystem.Theauthors showedthatSDSisabletodispersenanotubeswithanaveragediameterof1nm producedthroughtheHiPcoprocess,butitishighlyineffectiveatsuspendingnanotubes 85

PAGE 86

Figure4-5.SWNTcontributiontoscatteringaftersubtractingthecontributionfromthe H-SDSsolution.ThescatteringintensityofH-SDSwasmultipliedbyafactor of0.86tomatchthetwopeaksat Q =0.05 .Thedataisdescribedbya functionoftheform I Q =.5 10 )]TJ/F25 7.9701 Tf 6.587 0 Td [(3 Q )]TJ/F25 7.9701 Tf 6.587 0 Td [(0.97 +2.2 10 )]TJ/F25 7.9701 Tf 6.587 0 Td [(2 2 = n 0.5 .A Q )]TJ/F25 7.9701 Tf 6.587 0 Td [(1.2 dependenceisobtainedifthescatteringintensityofH-SDSis subtractedasitis. oflargerdiameter,suchasthoseproducedthroughthearcdischargemethod.SANS ishighlysensitivetotheaggregationstateofSWNTs. 151,150,167 Thescatteringintensity fromrigidrodsischaracterizedby I Q )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 behaviorwhileaggregatedrodsappear asmassfractalsthatexhibita Q )]TJ/F25 7.9701 Tf 6.586 0 Td [(2.5 dependence.Figure4-5showsthecontribution ofSWNTstothescatteringintensityaftersubtractingtheintensityfromtheSDS solution.Thedataisonlyshownupto0.02sincethesignaliscompletelydominated bytheSDSmicellesabovethis Q -range.Thedatacanbedescribedbyapowerlaw functionwithanexponentof.97,whichistherelationshipexpectedofwell-dispersed rigidrods.ThisdatastronglysuggestthatSWNTsinthisparticularsystemHiPco SWNTsbehaveasrigidrods,whichisinagreementwiththeresultsofQuintonand co-workers. 50 TheabsenceofastrongsignalfromtheSWNTsorthesurfactant-SWNTcomplexes athigh Q valuesindicatesthatsurfactantsdonotformwell-organizedaggregatesonthe SWNTsurfaceasshowninFigure4-5.Yureklietal.performedadetailedSANSstudy ontheassemblyofSDSonSWNTsandreachedsimilarconclusions. 56 Theauthors 86

PAGE 87

Figure4-6.CartoondepictinglociofsolubilizationfortheorganiccompoundsintheSDS micelles. foundthatthescatteringintensityathigh Q couldnotbettedbyanycombinationofthe scatteringfrominteractingspheresmicellesandSWNTswithcylindricalmicelleson theirsurface. 4.4.3EffectsofSolventSolubilizationonSDSMicelles Thesolubilityofthetwoorganicsolventsinwaterissmall,being1.77g/Lfor benzeneand0.15g/LforODCB. 132 However,inthepresenceofSDSmicellestheir solubilityincreaseslinearlywiththeconcentrationofmicelles.Thehydrophobicmicelle core,thesurfactantpalisadeareabetweenthepolarheadandthehydrophobiccore andeventhepolarheadserveaslociofsolubilizationforaromaticcompounds,as showninFigure4-6. 168,169 Thepartitionofthetwocompoundsbetweenwaterandmicellesischaracterizedby themicelle-waterphasedistributioncoefcientthatisdenedas, 170,171 K XM = Molefractionofsoluteinmicellarphase Molefractionofsoluteinaqueousphase The log K XM valuesforbenzeneandODCBinSDSsolutionsareapproximately3 and3.89. 171 WiththeconcentrationofSDSusedinourexperimentsandthegiven valuesof log K XM ,thesolubilityofbenzeneandODCBincreasesbyafactorof1.4 and4.7,respectively.Noticethelargeincreaseintheabilityofwatertosolubilizethese compoundsinthepresenceofSDSmicelles.Asthesecompoundsaresolubilized,the micellevolume,symmetryandaggregationnumberareexpectedtochange. 172 These 87

PAGE 88

changesareexpectedbecausethehydrophobicvolumechangesduringsolubilization, asexplainedbelow.Notethatsolubilizationoftheorganiccompoundisadifferent phenomenontothespontaneousformationofmicro-emulsions.Inthelattercase,the solventformsasegregatedregioninthehydrophobiccoreofthemicelle.Intheformer, themicellesareswelledasthesolutecanbelocatedatdifferentpointswithinthemicelle dependingontheirstructure.Itwaspreviouslymentionedthattheintensityfromthe H-SDSsolutionincreasesatthehighQ regimeasthesolutionismixedwithODCB andbenzene.Itislikelythatthelargechangesinintensitycorrespondtochangesin thesizeoftheSDSmicellesandconsequentlythenumberofcharges.Table4-3shows thegeometricalparametersoftheH-SDSmicellesinthepresenceofsolvents.Itis observedthatthemajoraswellastheminoraxesofthemicelleincrease,whichresults inavolumechangeof62%and72%forbenzeneandODCB,respectively. Table4-3.ParametersusedtotthescatteringfromtheH-SDSsolutionduringswelling. Thedatawasttedtoauniformellipsoidwiththestructurefactorfora screenedCoulombpotential.Theconcentrationoffreeionswas0.008M whichisaroundthecriticalmicelleconcentration ParametersBenzeneODCB MinorAxis17.6118.06 MajorAxis24.9926.21 Charge,z21.1619.98 2 = n 0.5 4.082.44 4.4.4EffectsofSolventSolubilizationonSWNTSuspension ExtractingmeaningfulinformationonthescatteringbehaviorofSWNTsafter swellingwiththesolventrequiresthesubtractionofthecontributiontoscattering intensityfromswelledmicellesasdescribebelow, I 1 = I sol )]TJ/F43 7.9701 Tf 6.587 0 Td [(SDS + I sol )]TJ/F43 7.9701 Tf 6.586 0 Td [(SWNT + I Inc I 2 = I sol )]TJ/F43 7.9701 Tf 6.586 0 Td [(SDS + I Inc ,and I 1 )]TJ/F42 11.9552 Tf 11.955 0 Td [(I 2 = I sol )]TJ/F43 7.9701 Tf 6.587 0 Td [(SWNT I sol )]TJ/F43 7.9701 Tf 6.587 0 Td [(SDS and I sol )]TJ/F43 7.9701 Tf 6.587 0 Td [(SWNT arethecontributiontoscatteringbytheswelledSDSmicelles andSDS-coatedSWNTs,respectively.I Inc isabackgroundduetoincoherentscattering. Figure4-7showsthescatteringproleaftersubtractingthecontributionfromfree 88

PAGE 89

Figure4-7.Scatteringfromabenzene-andbODCB-swelledSWNTsuspensions afterthescatteringfromtheswelledH-SDSsolutionwassubtracted. swelledmicellesforSWNTsuspensionsmixedwithbothbenzeneandODCB.An importantquestioniswhethertheaggregationstateofSWNTschangesasthe suspensionisswelledwiththeorganicsolvents.Thelow Q signalhasa Q )]TJ/F25 7.9701 Tf 6.586 0 Td [(2.5 and Q )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 formforbenzeneandODCB,respectively.Thischangeinslopesuggeststhat SWNTsinthesuspensionaggregateafterswellingwithbenzene.However,thereisno visiblesignofaggregationinthesuspensions,andPLandabsorbancespectradonot showsignsofaggregationeither. 51 Inaddition,andmostimportantly,thescatteringat low Q returnstoasimilarintensityoncethesolventisevaporatedforbothbenzeneand ODCB,asshowninFigure4-3.OncethecontributionfromSDSmicellesissubtracted, theslopeatlow Q hasavalueclosetoforbothcompounds,suggestingthatthe swellingprocessandpossiblechangestotheaggregationstatearereversible.This reversibilitywasalsoobservedinpriorspectroscopyexperiments. 51 Figure4-7alsoshowsanewfeatureforSWNTsuspensionsmixedwithboth benzeneandODCBatthehigh Q range.Thisresultisinstarkcontrasttotheresults fromSDS-coatedSWNTs,whichshownostructureontheSWNTsidewall.Since thesignalfromtheswelledmicelleswassubtracted,itisprobablethatthestructure producingtheexcessscatteringontheSWNTsidewall.Accordingtothe Q -rangethat 89

PAGE 90

itisbeingprobed,thestructurehasacharacteristiclengthscaleofatleast15.Itis highlyprobablethatthemicrostructuresareamixtureoftheorganicsolventsandthe SDSmolecules.Acoatingformedbythesolventalone,excludingtheSDSmolecules, wouldallowtheaggregationofSWNTsduetotheexclusionoftherepulsiveelectrostatic interactions.Thefeaturesofthemicrostructuresarepotentiallydependentonthenature ofthesolventasdiscussedbelow. Whendescribingthestructureandself-assemblybehaviorofsurfactantsinto micelles,thenon-dimensionalpackingparameterisfrequentlyused, v = a o l a ,where v isthehydrocarbonhydrophobicvolume, a o istheinterfacialpolarareaand l a is thecriticaloralkylchainlength. 162 Thevalueofthepackingparameterdetermines themicellecharacteristics,includingwhetherasurfactantassemblesintosphericalor cylindricalmicelles.Thevalueofthepackingparametercanbemodiedbychanging thevalueof a o byusingsalttoscreenthechargesonthesurfactantmolecules.The changesobservedinFigure4-4andTable4-3areadirectconsequenceofthechanges inthepackingparameteruponsaltaddition.Morerelevanttoourexperimentsisthe factthat v canalsobechangedbythesolubilizationofaromaticcompoundsintoSDS micelles.Thepresenceofthearomaticcompoundscanchangethecontributionofthe hydrophobicvolumetothepackingparameter.However,theseeffectsaredependenton thearomaticspeciesandthedominantlocusofsolubilization.Forexample,itwasfound thattheresponseofcetylpyridiniumbromideCPB,whichformcylindricalmicelles,to thesolubilizationofhydrocarbonswasdependentonthelocusofsolubilization.The compoundsthatwerepreferablylocalizedatthepolarheadregioncausedanincrease intheaggregationnumberwhilethosethatweresolubilizeatthehydrophobiccore simplyswelledtheCPBmicelles. 173 Inasimilarapproach,thepackingparametercan alsobeusedasaguidetounderstandtheassemblyofsurfactantsonsolidsurfaces, suchasSWNTsidewalls. 174 SincebenzeneandODCBhavedifferentmolarvolumes 90

PAGE 91

Figure4-8.Cartoondepictingtheinitialandswelledsurfactantmicellesforab H-SDSandcdD-SDS. andlociofsolubilizationinSDSmicelles,thesetwosolventscouldpotentiallyinducethe formationofmicrostructuresofdifferentcharacteristicsonthesurfaceofSWNTs. 4.4.5CharacterizingSolventMicro-StructuresontheSWNTSidewall TheresultsinFigure4-7suggestthatsmalldomainsofsolventareassociated withtheSWNTs.Therefore,D-SDSwasusedinsimilarexperimentstoeliminatethe scatteringfromtheSDSmicelles.Asmentionedabove,oncealltheHatomsare replacebyDatomsintheSDSmolecule,thescatteringabove0.02 )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 becomes atindicatingthatthemaincontributiontotheintensityinthisregioncomesfromthe incoherentbackgrounddatanotshown.However,thesignalatlow Q matchesforboth samples.TheseresultsindicatemicellesarecontrastmatchedwhenSDSisdeuterated, asshowninFigure4-8. Figure4-9a-bshowsthescatteringprolefromaD-SDSsolutionmixedwith benzeneandODCB.Thesignalathigh Q ismostlybackgroundalthoughaveryweak peakisobservedaround0.04 )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 forbenzene.ThisindicatesthatbothODCBand benzenearedistributedthroughthemicelleanddonotformanaggregatestructurewith 91

PAGE 92

Figure4-9.aIntensityfromD-SDS-suspendedSWNTs,D-SDSswelledwithbenzene andSWNTsuspensionswelledwithbenzene.bIntensityfrom D-SDS-suspendedSWNTs,D-SDSswelledwithODCBandSWNT suspensionswelledwithODCB. characteristiclengthofatleast15insidethemicelles.Likewise,theatproleatthe high Q regioninSWNTsuspensionsindicatestherearenotvisibleaggregatestructures ontheSWNTsidewall.However,aggregatestructuresbecomevisibleoncethesolvents areadded.WhenD-SDS-coatedSWNTsareswelledwithbenzene,adistinctpeakis observedwhich,conrmsthatthesolventisformingamicrostructurearoundSWNTs. InthecaseofODCB,thepeakisnotasdistinctasinthebenzene-swelledsuspensions butitisstillvisible.Noticethatthepeakisonlyobservedwhennanotubesarepresent. AlthoughthesolventsdonotinduceconsiderablestructuralchangestoSDSmicelles, theydocauseavisibleeffectontheSWNTsidewall.ThedatainFigure4-9strongly suggeststhatthesolventsformaggregatesontheSWNTsurface.Asdiscussedabove, thesolventsactasareorganizingforceforthesurfactantsontheSWNTs. Acorrelationlengthmodelhasbeenusedtotthedataatlowandhigh Q .The modelhasthefollowingform, I = A Q m + C 1+ Q n + I Inc 92

PAGE 93

where A and C aretworealcoefcients, isthecorrelationlengthwhile m and n arethe PorodexponentsforthelowandhighQ regions,respectively. 165 Table4-4showsthe ttingparametersforthedatainFigure4-9forD-SDS-suspendedSWNTsafterswelling withbothbenzeneandODCB. Table4-4.ParametersusedtotthescatteringfromtheD-SDS-coatedSWNTstreated withbenzeneandODCB. PametersBenzeneODCB A 4.84 10 )]TJ/F25 7.9701 Tf 6.587 0 Td [(7 1.02 10 )]TJ/F25 7.9701 Tf 6.586 0 Td [(7 m 2.632.82 C 0.040.02 12.0219.92 n 3.623.03 The m and n exponentsgivesthemassandsurfacefractaldimensionofthesystem atlowandhigh Q ,respectively.Thecorrelationlength,inthecaseofpolymers,isthe distancebetweenthetwopointsatwhichapolymerchainisintersectedbytwoother chains.Thepreviousinterpretationcannotbetranslateddirectlyintooursystem,nor anunambiguousinterpretationcanbefoundindependentlyofthevaluesof n and m TheregimeofinterestinoursystemsisthehighQ regions.Theexponent n canbe interpretedasthedimensionalityofafractalsurface.Theexponent n showsavalueof4 forasmoothsurfacewhilearoughsurfacedisplaysavalueof3.Thecorrelationlength inourcasecanbeinterpretedastheaveragedistancebetweensolventaggregateson theSWNTsurface. Figure4-10dshowsthepossiblestructuresthat mightcorrespondto.Although wecannotconclusivelydisregardL 1 or0.5 L 1 ,itispossibletoarguethatL 2 isthecase basedonthevaluesofthe n exponent.ThesuspensionswelledwithODCBshowsa veryroughsurfacewhilethesurfaceforSWNTswelledwithbenzeneisnotasrough. Thecorrelationlengthforbenzeneis12.02whichis40%lowerthanthevaluefor theODCBsystem.Thesystemswelledwithbenzeneshows n =3.6 whileforODCB n =3 .Hence,thesurfaceofthenanotubescoveredwithbenzenearelessroughthan 93

PAGE 94

Figure4-10.CartoondepictingtheSWNTsandthesurfactantassemblyonthesidewall withaH-SDSandbD-SDS.TheH-SDS/andD-SDS/solventassembly onSWNTsidewallarealsoshownincandd. theonecoveredwithODCB.Inotherwords,thedomainsofbenzeneonSWNTsare closertooneanotherthanthoseforODCB.Thisinterpretationisinagreementwiththe interpretationthatwasgivenpreviouslybasedonPLdata.ItwaspostulatedthatODCB increasethepermeabilityoftheSWNTinterface,whichasconsequenceallowedthe quenchingoftheSWNTPL.Ontheotherhand,benzenepromotedasurfactant-solvent structurethatisolatedthetubesbetterfromtheaqueousenvironment. 4.5Conclusions Thestructureofnonpolarsolventmicroenvironmentsanditseffectonthe aggregationstateofSDS-coatedSWNTswasstudiedusingSANS.Itwasshownthat SWNTsproducedthroughtheHiPcoprocesscanbewell-dispersedbySDS.Thisresult isofcriticalimportancefortworeasons:rst,theabilityofSDStoexfoliateSWNTs appropriatelyhasbeenquestioned;second,theinterpretationoftheresultspreviously presenteddependsontheaggregationstateoftheinitialSDS-SWNTsuspension.The formationofnonpolarsolventmicroenvironmentsdonotinduceaggregationofSWNTs. ThisstronglysupportsourpreviousclaimsthatthedramaticchangesobservedinPL spectraofSWNTsduringswellingareduetothesurfactantreorganizationontheSWNT 94

PAGE 95

surface.TheSANSmeasurementsalsocorroboratedthenotionthatthestructure formedbythenonpolarsolventsandtheSDSmoleculesmainlydependsonthenature ofthesolvent.Somesolventscaninducetheformationofamoreuniformcoatingon theSWNTsidewallthatresultsinbettercoverage.Onthecontrary,othersolvents inducetheformationofastructurethatexposesthesidewallsofSWNTsmakingthem susceptibletoextrinsicquenchingeffects. 95

PAGE 96

CHAPTER5 AMECHANISTICSTUDYOFTHESELECTIVERETENTIONOFSDS-SUSPENDED SINGLEWALLCARBONNANOTUBESONAGAROSEGELS 5.1Introduction SWNTshavephysicalpropertiesthatareenticingforalargenumberofapplications. However,thepresenceofbothmetallicandsemiconductingSWNTtypespreventstheir useinmanyapplications.Althoughprogresshasbeenachievedinthesynthesisof SWNTswithanarrowsizedistribution,post-synthesismethodsarestilltheonlyrouteto achievemonodispersitybydiameter,lengthandtype. 175,176 SWNTsmustbedispersedpriortoseparation.SWNTsaretypicallydispersedin aqueousmediausingeithercovalentfunctionalization,surfactants, 63,55 DNA, 24,142 orpolymers. 61,114 Themostpromisingmethodsofseparationtendtoexploitthe differencesinthesurfactantassemblysurroundingSWNTs.Forexample,density gradientultracentrifugationDGUtakesadvantageofthedifferenceinbuoyantdensity providedbythespecicinteractionsofthesurfactantshellwiththeSWNTs. 27 Another interestingmethodrecentlydevelopedbyKatauraandco-workersistheseparation ofSWNTsintometallicm-andsemiconductings-speciesbychromatographywith agarose-gelbeadsasthestationaryphase. 177,178 Intheseseparations,thes-SWNTs areretainedbytheagarosebeadswhilethem-SWNTsareeluted.Thes-SWNTsare onlyreleasedfromthecolumnoncethesodiumdodecylsulfateSDSeluentisreplaced withasodiumcholateSCsolution.Althoughthechromatographicseparationwith agarosebeadsisnotyetaseffectiveasDGU,itismoreconducivetoscaleup.The retentionofSWNTsonagarosegelbeadsisstillnotwellunderstood,withtwodifferent retentionmechanismsbeingproposed.TherstmechanismproposedbyKataura andco-workersisbasedontheselectiveadsorptionofs-SWNTsontotheagarose gel. 177,178,33 Thealternativemechanismisthatseparationiscausedbysize-exclusion ofs-SWNTs,whichwasproposedbyMoshameretal. 166 Theydemonstratedthat m-SWNTssynthesizedthroughthearc-dischargemethodarebetterdispersedthantheir 96

PAGE 97

semiconductingcounterparts.Theauthorsthenproposedthatthepoorlydispersedand, consequently,largerbundlesofs-SWNTsareretainedbytheagarosebeads.However, theselectiveretentionofs-SWNTbyvirtueoftheirsizeiscontrarytotheprinciples associatedwithsize-exclusionchromatography,whichdictatesthatlargerobjects, suchasbundledSWNTs,passthroughthecolumnfasterbecauseoftheirinability todiffusethroughsmallerpores.Onlyextremelylargeobjectsrelativetothepore dimensionswouldbetrappedbythebeads.Therefore,bundleds-SWNTswouldhave tobesignicantlylargerinsizethantheindividualizedm-SWNTs,whichiscontraryto mostanalyticalreportsofthesesuspensionsbasedoneitherAFM,photoluminescence, absorbanceandRamanspectroscopy. 63,55,166,62,179,180,50 Inthispaper,weexplorethemechanismofchromatographicseparationofSWNTs inagarosecolumns.Experimentsaredesignedtoprobeeachpotentialmechanism. Thesizeexclusionofaggregateds-SWNTsistestedbyanalyzingtheretention characteristicsofSDS-SWNTsuspensionsthathavebeenultracentrifugedforvarying lengthsoftime.Alternatively,gumarabicGA,whichisasugar-basedsurfactantwith structuralsimilaritiestoagarose,isaddedtotheinitialSDS-SWNTsuspensiontoalter thesurfaceofthenanotubes.ThissurfactantadsorbsontothesurfaceofSDS-coated SWNTswithoutchangingtheiraggregationstateandpreventstheiradsorptiononto thebeads,conrmingthatselectiveadsorptionisresponsiblefortheseparationofsfromm-SWNTs.Finally,thefeaturesofthesurfactantshellthatenableadsorptionare investigatedaswellasthenatureofanactiveadsorptionsiteontheSWNTsurface. 5.2Methods 5.2.1Reagents Thesurfactants,sodiumdodecylsulfateSDS 99%,sodiumcholateSC 99%,gumarabic,anddouble-strandeddeoxyribonucleicacidDNAfromsalmontestes werepurchasedfromSigma-AldrichSt.Louis,MO,USAandusedasreceived.HiPco SWNTswereobtainedfromRiceUniversityRiceHPR162.3andusedasreceived. 97

PAGE 98

Carbontetrachloride.9%ando-dichlorobenzene%werepurchasedfrom Sigma-Aldrich.Benzene.5%andp-xylene.7%werepurchasedfromFlukaand FisherScienticPittsburgh,PA,USA,respectively.Allsolventswereusedasreceived. 5.2.2AqueousSWNTSuspensions Aqueoussuspensionsofnanotubeswerepreparedbymixing60mgofraw SWNTswith200mLofa1wt%SDSsolution.High-shearhomogenizationIKAT-25 Ultra-Turraxfor30minandultrasonicationMisonixS3000for10minWwere usedtoaidalldispersions.Afterultrasonication,themixturewasultracentrifugedat 20,000rpm gfor4husingaswingbucketrotorBeckmanCoulterOptima L-80K.TheDNA-coatedSWNTswerepreparedbymixing50mgofSWNTsand50 mgofDNAin60mLofwater.Thesuspensionwassonicatedfor5min.5Wand ultracentrifugedfor2hat20,000rpm.Beforesonication,thesizeoftheDNAstrandsis between500and10,000basepairs.GelelectrophoresisshowedthattheDNAsizeis reducedtoabout500basepairsaftersonication. 5.2.3ColumnExperiments Thecolumnexperimentswereperformedusinglow-pressurechromatography columnsfromBio-RadseeFigureC-1.Thecolumnsweremadeofglassandhadan innerdiameterof1.5cm.ThecolumnswerepackedwithagarosebeadsSepharose 6B,45-165 mindiameterupto9cminheight.AowadapterconnectedanEcono gradientpumpBio-Radtothecolumn.Atypicalexperimentalsequenceconsisted ofrststabilizingthecolumnwithatleasttwocolumn-volumesCV,approx.32 mL,of1wt%SDSsolution.One-halfofacolumn-volumeofeithertheinitialorthe surfactant-modiedsuspensionwastheninjectedintothecolumn.Theearlyfractions ofSWNTswereelutedwithoneCVofSDSsolutionfollowedbytwoCVsof2wt%SC solutiontoremovetheretainedSWNTsfromthecolumn. 98

PAGE 99

5.2.4DensityGradientUltracentrifugation Lineardensitygradientswereformedbylayeringaqueoussolutionsofiodixanol boughtasOptiPrep fromSigma-AldrichincentrifugetubesmLthick-wall polycarbonatefromBeckman-Coulter.Volumesof1.5,3.75,and10mLof60,40, and30%w/viodixanol,respectively,weresequentiallyaddedtothetubefollowedby 10mLofDIwater.TheSWNTsuspensions.75mLwereinsertedontopofthe30% iodixanollayerusingasyringepump.Priortoinsertion,theSDS-SWNTsuspension wasadjustedto20%w/viniodixanol.TheSDSconcentrationwas1wt%throughout thetube.Thecentrifugetubeswerelefttorestforatleast5htoestablishthelinear gradient.Thesampleswerethenultracentrifugedfor17hat27,000rpm g. 5.2.5SWNTCharacterization Theefuentfromthecolumnwascharacterizedbyabsorbanceanduorescence spectroscopyusinganAppliedNanoFluorescenceNanospectralyzer Houston, TXwithexcitationfrom662and784nmdiodelasers.Theefuentwascontinuously characterizedin-situusingauorometerowcellfromStarnaCells.Typically,absorption spectraweretakenevery20swhiletheefuentowedthroughthecell.Insomecases, aBio-RadfractioncollectorModel2110wasusedtoobtainsamplesevery1/3CV duringelution. 5.3ResultsandDiscussion 5.3.1RetentionBehaviorofas-PreparedSDS-SWNTSuspensions TheretentionbehaviorofanSDS-coatedwt%SWNTsuspensionwasmeasured toestablishtheelutioncharacteristicsoftheinitialsuspension.Asobservedbyothers,a signicantfractionofnanotubesisinitiallyretainedontheagarosebeadswhiletherest ofthenanotubesareelutedoutofthecolumnwith1wt%SDS. Figure5-1ashowstheelutioncurvechromatogramoftheseSWNTsuspensions astheypassthroughthecolumn.Therstpeakisobservedat 20min.Thevisible absorbancespectraattherstpeakinFigure5-1bshowsthattherelativeconcentration 99

PAGE 100

Figure5-1.GeneralretentionbehaviorofSWNTssuspendedin1wt%SDS.aElution andcumulativemasscurves.Theelutioncurvesarepresentedintermsof theabsorbanceoftheefuentnormalizedbytheabsorbanceoftheinitial suspension.Allabsorbancedatapointsareat = 763nm.Detailsofthe calculationsforthecumulativemasscurvesareprovidedinthesupporting information.bAbsorbancespectrafromtheinitialsampleandtheefuent attherstP1andsecondP2peaksoftheelutioncurve,whichareshown asdashedlinesinparta.TheSWNTsuspensionisinjectedattimezero. Theplotalsoshowsthetimesatwhichtheeluents,SDSandSCsolutions, areinjected. ofmetallicspecies-600nmhasincreasedbutthesuspensionstillconsistsof bothm-ands-SWNTs.Thisresultislikelybecauseofthelargevolumeofsuspension loadedintothecolumn.Thishighloadingwasdoneintentionallyfortheseexperiments sothatthedifferencesinretentionbehaviordescribedbelowcouldbeeasilyobserved. TheremainingnanotubesinthecolumncannotbeelutedwithSDS.Theretentionof theseSWNTsisreversibleoncetheeluentisswitchedfrom1wt%SDSto2wt%SC 100

PAGE 101

after24min.Asecondpeakintheelutioncurveofthesuspensionisthenobservedat approximately48min,asshowninFigure5-1a.Thispeakisbroaderandlesssmooth thantherstpeak.Themuchlowerrelativeabsorbancebetween400and600nm inFigure5-1bshowsthatthespecieselutinginthispeakaremainlys-SWNTs.In thisparticularsample,thesecondpeakaccountsfor65%ofallthenanotubesthat areeluted,asobservedfromthecumulativemasscurve.Ingeneral,62 3%ofthe nanotubesarerecoveredinthesecondpeak.Duetotheirreversibleretentionofsome nanotubes,thecumulativemasscurvehasbeennormalizedtothetotalamountof SWNTsthatarerecovered,asexplainedinthesupplementaryinformationsection.For as-preparedSWNTsuspensionsinSDS,theamountofnanotubesthatareirreversibly retainedinthecolumnisverysmall,somostoftheinjectednanotubesarerecovered i.e.,recoveryof99 10%.Thesamecannotbesaidforothersamplesthathavebeen modiedasshownbelow. 5.3.2ProbingSize-ExclusionasaMechanismforSelectiveRetentionofsSWNTs Inanidealsize-exclusionseparation,samplesareelutedisocratically;inother words,thereisnoneedtochangetheeluentsorbuffer.Furthermore,allthesample componentsshouldeluteinapproximatelyoneCV. 181 However,theelutionofSWNTs fromthecolumnrequiresmorethanoneCVFigure5-1aofeluentandthechangefrom SDStoSCsolutions.Thesegeneralobservationspointtothefactthats-SWNTsare selectivelyretainedduetoadsorption. Tofurtherexplorethepossibleselectiveretentionduetoaggregationofs-SWNTs andsubsequentsize-exclusion,experimentswereconductedonsuspensionswith differentdegreesofaggregation,whichwerepreparedusingdifferentultracentrifugation times.FigureC-2showsthatthesuspensionsultracentrifugedthelongesthavethe mostdistinctfeatures.Thesespectralchangesarerelatedtotheamountofbundled SWNTsandindicatethatbundledSWNTsareremovedwithlongerultracentrifugation 101

PAGE 102

Figure5-2.Retentionbehaviorof1wt%SDS-SWNTsuspensionspreparedwith differentaggregationstates.Elutionandcumulativemasscurvesforthe SWNTsuspensionssubjectedtoa240andb30minof ultracentrifugation.cMassfractionofSWNTsthatareelutedinPeak1 P1andPeak2P2aswellasthosethatareirreversiblyretainedwithinthe column. times. 182,183 Figures5-2aand5-2bshowthattheelutioncurvesforsuspensionsafter 240and30minofultracentrifugationareverysimilar.Theretentionbehaviorofthe suspensionsultracentrifugedbetween240and30minisalsosimilarseeFigures 5-2candC-3.However,Figure5-2cshowsthatthesuspensionsubjectedto30minof 102

PAGE 103

ultracentrifugationhassomeirreversibleretention,whichcanbevisuallyobservedinthe columnaftertheexperiment.Theamountofirreversibleretentionincreasesdramatically fornon-centrifugedsuspensions.Inthiscase,approx.80%ofthenanotubesinjected areirreversiblyretainedinthecolumn.TheabsorbancespectrainFigureC-3alsoshow thataggregationaffectsthequalityoftheseparation,whichislikelyduetothefactthat m-ands-SWNTsarebundledtogether.Theseresultsshowthatwhileaggregationand size-exclusiondohavesomeimpactontheelutioncharacteristics,itisnotresponsible fortheselectiveretentionofs-SWNTs.Ifsizeexclusionwereresponsibleforselectivity, thenthesecondpeakintheelutioncurveshoulddecreaseasSWNTbundlesare removedatlongerultracentrifugationtimes.However,Figure5-2cshowsnosignicant changestotheelutioncharacteristicsbetween30and240min.Instead,irreversible retentionisobservedoncetheaggregatesreachathresholdsizeandarenotableto owthroughthespacebetweenthebeads,whichisconsistentwiththeprinciplesof size-exclusionchromatography. 5.3.3ProbingSelectiveAdsorptionasaMechanismforSelectiveRetentionof s-SWNTs Althoughnon-functionalizedagarosebeadsSepharose aremainlydesignedfor size-exclusionchromatography,itisknownthattheycanpresentstrongnon-selective adsorptionofanalytes.Theseeffectsaredetrimentaltoseparationbutinsomecases canbeusedadvantageously. 184,185 Ifselectiveadsorptionofs-SWNTstoagaroseis theprimarymechanismforseparationinthecolumn,thenretentioncharacteristics canbemodiedbydisruptingthisadsorptionprocess.Thiscanbeachievedby addingacompoundthatactssimilartoagaroseinbindingwiththeSWNTs.Once thiscompoundisadded,theadsorptionsitesonthenanotube-surfactantcomplex wouldnotbeavailabletointeractwiththeagarosecolumn.Therefore,theretention characteristicsshouldbedramaticallyalteredandthenanotubesshouldpassthrough thecolumnwithsignicantlylowerretention. 103

PAGE 104

Figure5-3.RetentionbehaviorofSDS-SWNTsuspensionsmodiedwith0.1wt%of GA.Elutionandcumulativemasscurvesfortheainitialandb GA-modiedSWNTsuspensions ThestructureofgumarabicGAiscomplex;however,bothGAandagarosehave carbohydrates,inparticulargalactose,astheirmainstructuralunits. 186 Althoughthe dataaboveshowthataggregationmainlyaffectsirreversibleretention,thesuspension wastitratedwithGAtoobservethechangestotheopticalspectraseeFigureC-4and ensurethattheSWNTsdonotaggregate.TheresultsshowthatGAconcentrations below0.3wt%signicantlyincreasesthePLintensityoflargediameterSWNTs,which isattributedtobettercoverageofthenanotubes,withoutinducingaggregation. 187,51 Furthermore,thepronouncedred-shiftintheemissionpeaksindicatesthatthepolarity ofthemediumsurroundingSWNTshaschanged,seeChapter3.AsshowninFigure 5-3aand5-3b,theretentioncharacteristicsaresignicantlyalteredwhen0.1wt%GA isaddedtothe1wt%SDS-SWNTsuspension.TheamountofSWNTsretainedinthe columnhasfallenfromapprox.58to30%withasmallamountofGA.Theabsorbance spectrashowstheSWNTsretainedarestillpredominantlys-SWNTsseeFigureC-5. Thesedramaticchangessuggestthattheinteractionswiththecolumnmediahave beensubstantiallyreduced.Infact,theinteractionoftheSWNTsinpeak1hasalso beengreatlyreduced.NearlyalloftheSWNTsinthisfractionelutewithinoneCV16 min,indicatingthattheeluentissimplypushingtheSWNTsthroughthecolumn.The 104

PAGE 105

dramaticchangesobservedafteraddingGAstronglysuggestthatselectiveadsorption ofSWNTsisresponsiblefortheselectiveretentionofSDS-SWNTsonagarosegel beads. 5.3.4EffectofSurfactantStructureontheSelectiveInteractionofs-SWNTswith Agarose ThetotalinteractionbetweenSWNTsandtheagarosebeadswillbedependenton thevanderWaalsforcesbetweentheSWNTsandagaroseaswellasthestericand electrostaticinteractionsfromthesurroundingsurfactantlayer.Althoughthereisstilla debateonhowthesurfactantsassembleontheSWNTsurface,theoreticalcalculations haveshownthatSDSmoleculestendtolieatalongtheaxiallengthoftheSWNT surfaceatlowsurfacedensityapprox.1molecule/nm 2 53,57 Thisstructureisespecially dominantforsmalldiameterSWNTs,suchasthe,6nanotube.Theformationofthis orientationwasduetothehighenergeticpenaltyrequiredtobendtheSDSmolecules aroundthenanotube.Therefore,theSDSheadgroupsaresurprisinglyveryclosetothe surfaceofSWNTs.Anotherimportantresultfromthesecalculationswasthatthesurface ofSWNTsisonlypartiallycoveredbytheSDSmolecules. Anydifferencesinsurfactantstructurewillaffecttheinteractionsbetweenthe SWNTsandagarose.Figure5-4showshowthesurfactantmorphologycoulddictate theinteractionsoftheSWNTwiththeagarosematrix.Regardlessofthenatureofthe repulsiveforcee.g.,electrostaticorsteric,athickbrush-likesurfactantstructurehigh aggregationnumberstatewouldprovidemorerepulsion,likelypreventingananotube fromadsorbingontothesurfacebecauseoftherepulsivebarrierthatovercomesthe vanderWaalsinteractions.Ontheotherhand,alowaggregationnumberofsurfactant cansignicantlyaltertheinteractionenergyallowingthenanotubetobeadsorbedonto thesurface.Experimental 187,51,68,28 andtheoreticalinvestigations 53,57 haveshownthat surfactantmoleculescanhavedifferentorientationsaroundSWNTsandthatpartsofthe nanotubemaybehighlyexposedtothesurroundingmedia.Therefore,theorientationof 105

PAGE 106

Figure5-4.EffectofsurfactantstructureontheinteractionpotentialbetweentheSWNT andtheagarosesubstrate.aThoseSWNTswithalowaggregation numberofsurfactantmoleculesontheSWNTwilladsorbontothesurface whilethoseSWNTswithahighaggregationnumberwillremaininthemobile phase.bDiagramillustratinghowtheaggregationnumberofsurfactantwill changetheinteractionenergyastheSWNTsapproachtheagarose substrate. surfactantsorlackofcoverageonSWNTscouldinuencetheretentioncharacteristics. Understandingthefactorsthatcontroltheadsorptionprocessmayleadtoimprovements intheseparationefciencyofthisprocess.Toinvestigatetheroleofthesurfactant, thesurfactantshellofaninitialSDS-SWNTsuspensionwillbemodiedbyelectrolyte tuning 68,28 andswellingwithorganicsolvents. 51 Table5-1summarizestheeffectof themodiedsurfactantstructuresonSWNTPL,theirsurfacecoverageandelution characteristics. 5.3.5ChangestoRetentionCharacteristicsbyAlteringtheSurfactantStructure throughElectrolyteTuning TherstmethodusedtoalterthesurfactantstructurewastheadditionofNaCl totheSDS-SWNTsuspension.Previously,Doornandco-workersshowedthat SDS-suspendedSWNTsexperienceintensitychangesandablueshiftintheuorescence spectrawhensaltisadded. 28 FigureC-6conrmsthatlargerSWNTsaremoresensitive totheincreaseinNaClconcentration. 68 Theseintensityandemissionenergydifferences indicatethattheimmediateenvironmentsurroundingSWNTshaschanged.These 106

PAGE 107

Table5-1.SummaryofSWCNTsuspensionsstudiedandtheeffectthatthemodiers hadonPLintensityandretentionbehavior.Thechangesaredescribedin referencetotheinitialSDS-SWCNTsuspension. SampleeffectonPLintensitySWNTcoveragepeak1%peak2% InitialSDS-SWNTs38 362 3 0.1wt%GAIncreasedIncreased7030 60mMNaClIncreasedIncreased7030 ODCB-SwelledGreatlyreducedDecreased7624 p-xylene-SwelledGreatlyreducedDecreased7624 CCl 4 -SwelledIncreasedIncreased5347 Benzene-SwelledIncreasedIncreased5248 ODCB-EvaporatedSimilarSimilar5050 CCl 4 -EvaporatedIncreasedIncreased50.549.5 DNA-SWNTsGreatlyreducedDecreased7030 changesaffectthebuoyantdensityofSWNTs.Thechangesaremoredramaticfor m-SWNTs,whichformadistinctredbandinthecentrifugationtubeseeFigureC-6, indicatingthatm-SWNTshavelargerdensitychangesdecreasesthans-SWNTs.The differenceinPLemissionenergyandintensityalongwiththedecreaseinbuoyant densityindicateschangestotheconformationandaggregationnumberofSDS moleculesontheSWNTsurface. 28 Thepartialscreeningofthechargeonthehead groupofthesurfactantnotonlyallowsahigherpackingdensityontheSWNTsbutitalso reorientsthemolecules,causingtheheadgroupstomovefurtherawayfromthesidewall ofthenanotube.ArecentstudybyDoornetal.showsthatthesechangesarenotabrupt andtakeplaceincrementallyasNaClisadded. 29 Thechangesintheretentioncharacteristicsareseenbycomparingtheresults forNaCl-modiedSWNTsuspensionsinFigure5-5atotheinitialSWNTsuspensions inFigure5-1.Althoughtherearestilltwopeaksintheelutioncurve,itisclearthat theretentionhasbeensignicantlyaffectedbymodifyingthesurfactantstructurewith 60mMNaCl.Whiletheinitialas-preparedsuspensionshadapproximately62%of allSWNTsreversiblyretainedinthecolumn,thesalt-modiedSWNTsuspensions inFigure5ahadapproximately30%retention.Inaddition,ashouldernowappears ontherstpeak.Thisshoulderbecomeslargerasthesaltconcentrationincreases 107

PAGE 108

Figure5-5.Retentionbehaviorofsalt-modiedSWNTsuspensions.aElutionand cumulativemasscurvesfor1wt%SDS-SWNTsmodiedwith60mMNaCl. TheelutioncurvesaswellasabsorbancespectraatotherNaCl concentrationsareshowninFigureC-7.bMassfractionofSWNTsthat areelutedinPeak1P1andPeak2P2aswellasthosethatare irreversiblyretainedwithinthecolumn. upto80mMseeFigureC-7.Thechangesintheretentionbehaviorinducedby increasingtheNaClconcentrationarebetterobservedinFigure5-5b.Theamountof SWNTsretainedinthecolumnsecondpeaksteadilyfallsfrom62to20%asthesalt concentrationincreasesfromto0to80mMNaCl.AtthehighNaClconcentrationof110 mM,aggregationisalreadysignicantforlargerSWNTs.Asaconsequence,approx. 20%ofSWNTsareirreversiblyretainedatthetopofthecolumn.Inordertoruleout thereducedretentiontopossiblechargescreeningontheagarosematrix,elutionof SWNTswasattemptedwithaSDS-NaClsolutiongradient.TheSDSconcentrationwas 1wt%andtheNaClconcentrationwasslowlyincreasedfrom0to80mMover6CV 108

PAGE 109

seeFigureC-8.ItwasobservedthatonlyasmallfractionofSWNTsareelutedwith thismethod.Hence,theeventsthatinducechangestotheretentionbehavioroccurred priortotheSWNTsinteractingwiththeagarosematrix.Finally,thelackofpeaksin theabsorptionspectrabetween400and600nmshowninFigureC-7conrmthat thes-ratherthanm-SWNTsarethoseretainedinthecolumn.Becausethesurfactant structureisalteredwithoutchangestotheaggregationstate,theseresultsagain conrmthatselectiveadsorptionisresponsiblefortheselectiveretentionofs-SWNTs. ThesmoothchangesintheretentionbehaviorofSWNTsalsoreecttheincremental changesinthesurfactantstructurethattakeplaceasNaClisadded.Thesurfactant makesagradualtransitionfromastructurewithlowaggregationnumbertoonewith higherpackingdensityassaltisadded.Theelutionprolesshowthatthetransitionin aggregationnumbergreatlyaffectstheinteractionofthenanotubeswiththeagarose matrix.AsshowninFigure5-4,thesedifferenceswillaffecttheinteractionpotential betweentheSWNTsandtheagarose,alteringtheretentioncharacteristics. 5.3.6ChangestoRetentionCharacteristicsbyAlteringSurfactantStructure throughSolventSwelling Thenextmethodusedtomodifytheas-preparedSWNTsuspensionwastoswell thesurfactantstructuresurroundingSWNTswithimmiscibleorganicsolvents,suchas carbontetrachloride,benzene,o-dichlorobenzeneODCBandp-xyleneseeFigure C-9andChapter3. 51 Theintentionwastoprobetheretentionbehaviorwithsolvents thatinducesignicantchangestothePLspectraofSWNTsduringswelling. ItwaspreviouslyobservedthatthePLintensityofmostSWNTtypesfalls dramaticallywhenSDS-SWNTsareswelledwithODCBseeFigure5-6aandp-xylene datanotshown.Inaddition,largered-shiftsaredetectedforthosenanotubeswhose PLintensityisnotquenched.Theseobservationsindicatethatdramaticchanges intheimmediateenvironmentsurroundingSWNTshavetakenplace.Aftersolvent evaporation,thePLintensityoftheinitialsuspensioninFigure5-6aisrecoveredfor 109

PAGE 110

Figure5-6.AlteringthesurfactantstructurearoundSWNTswithODCBandCCl 4 .aPL spectraforSWNTstreatedwithODCB.Imagesofthecentrifugationtubes forthebinitial,cODCB-swelled,anddODCB-evaporatedSWNT suspensionsinadensitygradientmedium.ePLspectraforSWNTs treatedwithCCl 4 .ImagesofthecentrifugationtubesforthefCCl 4 -swelled andgCCl 4 -evaporatedSWNTsuspensionsinadensitygradientmedium. thesmallestdiameterSWNTswhilethelargernanotubeslongerwavelengthshow smallPLintensityincreases.Hence,swellingdoesnotchangetheaggregation stateofSWNTs.Inaddition,thepresenceofthesolventaltersthenatureofthe surfactant-SWNTinterfacebothbeforeandafterevaporation,resultinginchanges tothebuoyantdensityofSWNTsseeFigure5-6bd.ThequenchingofPLintensity indicatesthatduringswellingwithODCBtheSWNTsurfaceisgreatlyexposedtothe medium. 51 ItisprobablethatsmallrandomsurfactantdomainsformontheSWNT surfacewheretheorientationofSDSmoleculeschangeswhiletheirpackingdensity increasesresultinginalowerbuoyantdensityfortheSDS-SWNTs. Figure5-7bshowsthechangestoretentionbehaviorcausedbyswellingSDS-SWNTs withODCB.NotethatthelowamountofsolventintheSWNTsuspensionuponswelling 110

PAGE 111

Figure5-7.RetentionbehaviorofODCB-andCCl 4 -modiedSWNTsuspensions. Elutionandcumulativemasscurvesfortheainitial,bODCB-swelled,and cCCl 4 -swelledSWNTsuspensions. willnotpoisontheagarosematrix.ItisobservedthatthepresenceofODCBgreatly reducestheretentionofSWNTsbytheagarosebeads.ThesecondpeakofFigure5-7b isdecreasedwhiletherstpeakislargerandsharperthanthepeakinFigure5-7a. Inthiscase,theretainedSWNTsinthesecondpeakaccountforapprox.25%ofall thenanotubesthatareinitiallyinjectedintothecolumn.Thesevaluesaresubstantially lowerthanthe62%retainedinthecontrolsample.Accordingtotheabsorbance measurements,thenanotubesinjectedintothecolumnarecompletelyrecovered.Once again,theabsorbancespectrashowthatthenatureofthenanotubeselutedinthe secondpeakhasnotchangedseeFigureC-10.AsshowninTable5-1,suspensions modiedwithp-xylenenotonlyhavesimilarspectralchangestoODCBbutidentical retentioncharacteristicsaswell.Similartotheresultsforsalt-modiedsuspensions, achangeinthesurfactantstructureofSWNTbyadditionofeitherODCBorp-xylene resultsinreducedretentionontheagarosematrix.Thisresultisstrikingbecause, despitep-xylene-andODCB-swelledSWNTshavingtheirsurfacelargelyexposed,their retentionismuchlowerthantheinitialsuspension.Thissuggeststhatexposedsidewalls fromalackofsurfactantarenottheactiveadsorptionsites.Iftheseexposedregions constitutedtheactiveadsorptionsites,increasingtheiramountwouldhaveincreased retention.Interestingly,SDS-SWNTsuspensionsswelledwithCCl 4 orbenzenedata 111

PAGE 112

onlyshowninTable5-1showdifferentPLbehaviorthanp-xylene-andODCB-swelled SWNTs.Figure5-6eshowsthePLspectrafromSWNTstreatedwithCCl 4 .Increasesin intensityaswellasblue-shiftsinemissionenergyareobservedformostSWNTspecies. Onceagain,thesechangesindicatethatthelocalenvironmentaroundSWNTshas beenmodied.WhilethespectraforODCB-swelledSWNTslargelyreturnedtothe initialspectraafterevaporationofthesolvent,thePLspectraforCCl 4 -swelledSWNTs remainmoreintense.ThebuoyantdensityofallCCl 4 -modiedSWNTsdecreasessee Figure5-6fg,conrmingthatthesolventaltersthenatureofthesurfactant-SWNT interface.ItwaspreviouslyspeculatedthattheSDSmoleculesandsolventsbenzene andCCl 4 uniformlycovertheSWNTsurfaceduringswelling. 51 Contrarytotheresults withODCBandp-xylene,thepresenceofCCl 4 orbenzeneonlyslightlyreducesthe extentofSWNTretentionontheagarosebeads.Figure5-7cshowsthedifferencesin theretentionbehavior.TheelutionandcumulativemasscurvesofCCl 4 -swelledSWNTs aresimilartothe1wt%SDS-suspendedSWNTs.However,astherstpeakdecays, thereisadistinctpeakaround30minthatdoesnotappearinthecontrolsample. Theappearanceofthispeakisprobablyduetotheelutionofsomenanotubeswhose bindingstrengthhasbeenreducedandthatotherwisewouldcomeoutinthesecond peak.Therefore,thepercentageofnanotubescomingoutinthesecondpeakisreduced from60%toapprox.50%byswellingthehydrophobiccorewitheitherCCl 4 orbenzene. TheabsorbancespectrafromtheCCl 4 -swelledSWNTsatdifferentelutiontimesshows thattheselectivityhasnotbeenaffectedandthespectralookverysimilartothatofthe controlsampleseeFigureC-11. 5.3.7EffectofExposedNanotubeSurfaceonRetentionCharacteristics Theselectiveinteractionand,consequently,theselectiveadsorptionofSWNTs ontotheagarosematrixmustbeduetodifferentsurfactantmorphologiesonthe surfaceofeachSWNTtype.Althoughalteringthesurfactantstructureshowsthat theaggregationnumberaffectsretention,itdoesnoteliminatethepotentialrolethat 112

PAGE 113

exposedsidewallsfromalackofsurfactantmayhaveinadsorption.Asstatedabove, despitethesignicantamountofexposednanotubesurface,Figure5-7bshowsthat theretentionforODCB-swelledsuspensionsissubstantiallylowerthanthepristine SDS-SWNTsapprox.24vs.62%.Therefore,itcanbearguedthatexposedregions onthenanotubesarenotresponsibleforthestronginteractionofSDS-SWNTwiththe agarosematrix.Toinvestigatethisissuefurther,columnexperimentswereconducted withDNA-suspendedSWNTs.DNA-suspendedSWNTsareagoodprobesinceitis knownthattheDNAassemblyonSWNTsleavessignicantportionsofthenanotube surfaceexposedtothemedium.Itisimportanttonotethatalthoughthedispersion qualityofDNA-suspendedSWNTsishigh,theirPLintensityisverylowduetotheir considerableexposuretothesuspendingmediumand,consequently,PLquenchers seeFigureC-12. 24,188,189,73 TheconcentrationofSWNTsinthissuspensionwas adjustedtomatchtheabsorbanceoftheSDS-SWNTsuspensionsat763nm. Figure5-8ashowstheelutioncurveoftheDNA-suspendedSWNTs.The absorbancespectrainFigure5-8bshowsthattheretentionofDNA-suspendedSWNTs isnotselectivesincethepeaksaresimilaranddonotshowselectiveenrichmentof eitherm-ors-SWNTs.ThisresultisnotsurprisingsincetheDNAusedinthiswork doesnotassembledifferentlyonm-ands-SWNTs.Mostimportantly,theretentionin thecolumnislowerforDNA-SWNTsapprox.30vs.62%forSDSsuspensions.The reducedretentionindicatesthattheorientationofthesurfactantratherthananexposed nanotubesidewallmustbethedominantfeatureofadsorption.Thelackofsignicant retentionforeithertheODCB-swelledorDNAsuspensionofSWNTsaswellasthe effectofionicstrengthonretentionbehaviorstronglysuggeststhattheorientationofthe surfactantanditssubsequenteffectontheinteractionpotentialbetweentheSWNTand theagaroseisthemostcriticalparameterthatdeterminesselectiveadsorption. 113

PAGE 114

Figure5-8.RetentionbehaviorofSWNTssuspendedwithDNA.aElutionand cumulativemasscurves.bAbsorbancespectrafromtheinitialsampleand theefuentattherstP1,secondP2,andthirdP3peaksoftheelution curve,whichareshownasdashedlinesinparta.TheSWNTsuspension isinjectedattimezero.Theplotalsoshowsthetimesatwhichtheeluents, waterand1wt%SC,areinjected. 5.3.8SummarizingtheEffectofSurfactantStructureontheSelectiveInteraction ofs-SWNTswithAgarose BydisruptingtheretentionprocesswithGA,ithasbeendemonstratedthatselective adsorptionisresponsiblefortheselectiveretentionofSWNTsontheagarosematrix. Thegradualtransitionfromastructurewithlowaggregationnumbertoonewithhigher packingdensityandtheconcomitantsurfactantreorientationasNaClisaddedtothe initialSWNTsuspensionsuggeststhatsurfactantorientationisthedominantfactor intheadsorptionprocess.Inaddition,modifyingthesurfactantstructurearoundthe SWNTsbyswellingwithODCBorp-xyleneshowsthattheexposedregionsonthe 114

PAGE 115

surfaceofnanotubesarenottheactiveadsorptionsites.Performingexperimentswith DNA-coatedSWNTs,whichhaveahighlyexposedsurfacetothemedium,further supportthisconclusion.ThebehaviorofODCB-andp-xylene-modiedSWNTswas contrastedwiththatofCCl 4 -andbenzene-swelledSWNTs.Theselatterswelled systems,whichformauniformsolvent-surfactantshell,showsimilarbehaviortothe initialSWNTsuspension. Katauraetal.havecorrectlyattributedthecauseofseparationtotheselective interactionofs-SWNTswithagarose. 176,190,191 However,theyarguedthatthesurfaceof s-SWNTispartiallycoveredbytheSDSmolecules,whichinturncausestheirstronger interactionwithagaroseduetoelectrostaticinteractionsorvanderWaalsforces. 176 Ina moredetailedstudy,Lietal.usedthioninedyetotrackchangestotheSDSmolecules ontheSWNTsurface. 192 Theyconcludedthatduringelectrophoresiss-SWNTsare strippedofSDSmoleculesandconsequentlyinteractstronglywithagarosethrough hydrophobicinteractions.However,theseresultsshouldnotbeextendedtothecaseof elutionchromatographyduetotheabsenceofelectricelds. Basedontheexperimentalresultsshownabove,weproposethattheorientationof thesurfactantisresponsiblefortheselectiveadsorptionofs-SWNTs.Thismechanism requiresthatdifferencesinsurfactantorientationarepresentintheinitialas-prepared suspension.Althoughthetheoreticalsurfactantconcentrationsusedinpriorsimulations 53,57,193 cannotbedirectlytranslatedtoexperimentalconcentrations,Xuetal.predictedthat SDSmoleculestransitionfromlyingatontheSWNTsurfacetolyingperpendicular whentheirsurfacedensitychangesfromlowtohigh. 57 Basedonthedifferencesin surfactantcoveragebetweenm-ands-SWNTsdescribedpreviouslybyDoornet al., 29 them-SWNTsarelikelysurroundedbyalargernumberofsurfactantmolecules. Therefore,weproposethats-SWNTsinthesuspensionhaveasurfactantstructure withthemoleculeslyingatonthesidewalloftheSWNTswhilem-SWNTshavethe surfactantorientedawayfromthenanotubesurface.Consequently,m-ands-SWNTs 115

PAGE 116

willhaveweakerandstrongerinteractionswiththeagarosematrix,respectively,as depictedbythehighandlowaggregationnumberstatesinFigure5-4.Theweaker interactionsform-SWNTsallowthemtopassthroughthecolumnwhilethes-SWNTs areadsorbed.OncetheeluentischangedtoanSCsolution,theSDScongurationon theSWNTsisdisrupted,alteringtheinteractionbetweentheSWNTsandtheagarose matrix.TheSCmoleculesmightalsoadsorbstronglytoagaroseanddisplacethe adsorbednanotubes. 5.4Conclusions Ithasbeenshownthatselectiveadsorptionratherthansize-exclusionisresponsible fortheselectiveretentionofSWNTsonagarosebeads.Moreover,thesystematic modicationofthesurfactantshellallowedthestudyofthefactorsthatdetermine adsorptionSWNTsontheagarosematrix.Forfuturework,itwillbeinterestingto takeacloserlookattheforcesthatdrivetheadsorptionofSWNTsontheagarose matrixandtoinvestigatethenatureoftheselectivityfurther.Itshouldbenotedthatthe observationspresentedabovearenotlimitedtoagarose-basedmedia.Recentwork hasshownthatSephacrylcross-linkedcopolymerofallyldextranandN,N'-methylene bisacrylamidealsoprovideselectiveadsorptionofs-SWNTs. 166,194 116

PAGE 117

CHAPTER6 BROWNIANDYNAMICSSIMULATIONSOFSWNTSEPARATIONINMICROFLUIDIC CHANNELS 6.1Introduction DielectrophoresisDEPisthephenomenonthatdescribestheresponseof polarizableparticlestothepresenceofanon-uniformelectriceldwhereaselectrophoresis occursinauniformeld. 195 Theparticles'responsedependsontheelectrical propertiesoftheparticlesandthemediuminwhichtheyareimmersed.Thefrequency oftheappliedelectricelddeterminesthemagnitudeanddirectionofthedielectrophoretic force.Dielectrophoresisandmicrouidicsystemscanbeusedsynergisticallytodevelop anefcientseparationsystemwithcontinuousow.Inaddition,thecharacteristic sizeandthelevelofowcontrolinmicrouidicsystemscanbeexploitedtoperform highlysensitivespectrometricmeasurementsonavarietyofuorescentmolecules. Forexample,Dittrichetal.havebeenabletofabricateasystemtocharacterizeand sortuorescentbeadsaswellascellsbasedontheirspectra. 196 Similarly,Morgan andcollaboratorscharacterizedlatexmicroparticlesanduseddielectrophoresisto manipulatetheirtrajectory. 197,198,199 Devicescapableoftrapping,focusing,sortingcells havebeenreported. 200 DEPisapromisingapproachformanipulatingSWNTs.TherststudyofSWNT behaviorinanon-uniformelectriceldwasdonebyKrupkeetal. 22 Inthatwork,it wasobservedthatmetallicm-SWNTswereattractedtotheelectriceldmaxima positiveDEP,whilesemiconductors-SWNTswererepelledfromtheelectric eldmaximanegativeDEP.Theyalsofoundthatthebehaviorofbothmetallicand semiconductorSWNTsdependedonthefrequencyoftheappliedelectriceldas predictedbytheory. 201 Asaresult,theywereabletodepositlmswithapproximately 80%metallicspecies. Asoftoday,itisnotclearhowtheelectrokineticbehaviorofSWNTsisaffected bythesurfactantmoleculesthathelpsuspendSWNTs.Theorysuggeststhatthe 117

PAGE 118

DEPforceoverm-SWNTsisseveralordersofmagnitudelargerthantheDEPforce experiencedbys-SWNTsinagivenelectriceldwhichshouldleadtocomplete separationofthetwospecies.Recently,Kimetal.achievedlmsdepositedthrough dielectrophoresiswhichwere100%metallicinnature. 202 Theauthorsusedacombination ofcationicandanionicsurfactantstoneutralizethenitesurfaceconductancecaused bythechargeofthesurfactants.However,themagnitudeoftheDEPforceexperienced byeachspeciesremainsanunansweredquestion.Also,largeelectriceldsareneeded tomoveSWNTs,demandingmicron-sizedelectrodes.Asaconsequence,onlysmall amountsofSWNTshavebeenprocessed 100pg. AsinitiallyproposedbyKrupkeetal.,thethroughputofDEP-basedseparationscan beincreasedbyusingmicrouidicsystems.ModelsdevelopedbyDimakiandBggil showSWNTscanbemanipulatedbydielectrophoresiswhileowingpastelectrodes. 203 Mattsonetal.builtasystemtoseparatem-ands-SWNTscontinuouslywhileowing throughanH-shapedchannel. 204 However,lowseparationefciencieswerereached anditwasrecognizedthattherewasroomforimprovement.Buildingonthatwork,Shin etal.usedasimilarmicrouidicsystembutwithanoptimizedelectrodegeometry. 205 TheelectrodesweredesignedtoincreasethelateralDEPforceandextractm-SWNTs fromastreamoftheinitialsuspensiontoaonecontainingabuffersolution.Therefore, itwaspossibletoobtainfractionswithmorethan90%enrichmentinm-ors-SWNTs.In anattempttoincreasethethroughputofseparationsthroughDEP-microuidicsystems, Pasqualietal.proposedtheuseofacoaxialcylindricalchannel,wheretheinnerand outercylinderswouldworkastheelectrodes. 206 Inthissystem,theseparationwould beaccomplishedbycollectingm-SWNTsontheinnerwirewhiles-SWNTwouldbe collectedatthechanneloutlet. Here,wepresentadetailedmodeltosimulatethetrajectoryofSWNTsunder electricandoweldsusingBrowniandynamicssimulations.SWNTs,whichhave adiameterofaround1nmandlengthsaslargeasafewmicrons,aremodeledas 118

PAGE 119

Figure6-1.IllustrationoftheDEPseparationdevice.Thechannellengthandgap betweentheelectrodesare L ch and L el slenderbodies.Theobjectiveistoprovideatoolforthedesignofeffectiveseparation systemsbasedondielectrophoresis.Wehavecombinedthepoweroftheniteelement methodforthepredictionofelectricandoweldswiththeabilityofBrowniandynamics simulationstondthetrajectoriesofastatisticallysignicantnumberofparticles. 6.2DescriptionofDEPSeparationDevice ThegeometryofthemicrouidicdeviceisshowninFigure6-1.Thedeviceconsists ofamicrouidicchannelwithside-wallverticalelectrodes. 207,208 Theelectrodescreate anelectriceldsothattheDEPforceontheparticlesisperpendiculartotheuidow withinthechannel.Thedeviceconsistsofthreeinletsandthreeoutlets.TheSWNT suspensionisinputthroughthechannelatthecenter.Theparticlesenterthechannel throughthecenterinletandareinitiallyforcedtostayinthecenterofthechannelby thetwoankingstreams.Theankingstreamsaresolutionsofthesurfactantatthe sameconcentrationastheoneusedtosuspendtheincomingSWNTs.Onceavoltage isapplied,theelectrodescreateanelectriceldthatisnon-uniforminthey-direction. AstheSWNTsowdownthechannel,theyexperiencea F DEP proportionaltotheir polarizabilityandnon-uniformityoftheelectriceld.Thepolarizabilityofm-SWNTs ishigherthanthatofs-SWNTsandconsequentlyexperiencealarger F DEP .Asa result,m-SWNTsareattractedtowardstheelectrodes,whiles-SWNTspassthrough thechannelwithnonetdeviation.Inthisway,thecentralstreamisstrippedofmetallic speciesandenrichedins-SWNTs.Incontrast,thelowerpartofthechannelisenriched inm-SWNTs.Thefateofthenanotubesdependsonthegeometryofthesystemand 119

PAGE 120

theoperatingparameters.If F DEP issmall,itispossiblethatm-SWNTssimplymigrate towardthelowerpartofthechannelandexitthechannelthroughOutlet1.However, if F DEP islarge,m-SWNTscanbetrappedbytheelectrodes.Ontheotherhand, s-SWNTswillremaininthecentralareaofthechannelandexitthroughOutlet2unless theresidencetimeislongenoughforthemtodiffusetootherpartsofthechannel. 6.3FabricationofDevices TofabricateprototypesofthedesigninFigure6-1,theproceduredevelopedby Wangetal.wasadopted. 207 Thesubstrateinthiscasewasa1mmthickborosilicate glassslidewithanareaof60 40mm 2 .Theglassslidewascleanedbysonicating inconsecutivebathsofacetone,methanolandwater.Afterdryingwithastreamof nitrogengas,theglassslidewasexposedtooxygenplasmaAnatechSCE600Asher for5min.Afterdehydrationat433Kfor20mininaconvectionoven,thegoldseed layerwasdepositedontheglassslide.Theseedlayerwasfabricatedusingthelift-off technique.A100nmlayerofAuwassputteredonaglassslidethatwaspreviously patternedwitha7 mthicklayerofAZ9260photoresist.A10nmlayerofCrwasused astheadhesionlayer.Theelectrodeseedlayerwasobtainedbydippingtheglassslide inacetonetoremovethephotoresist.Afterdehydration,theglassslidewaspatterned witha20 mthicklayerofAZ9260photoresist.TheAZ9260layerwasspunintwo stepstoachieveanalthicknessof20 m.Aftereachstep,thelayerwasbakedfor 10minonahotplateat363K.Afterexposureanddeveloping,themoldswereready fordepositionoftheelectrodes.Theelectrodepositionwasperformedinabathofgold electroplatingsolutionTechnicsgold25ES,TechnicsInc,RDwithwellcontrolledstir ratesandcurrentdensities.Then,sonicationinacetonewasusedtostriptheAZmold. ThesubstratewiththeelectrodeswasrinsewithDIwateranddehydratedat433K for20minbeforecoatingthenextSU-8layer.SU-82015Microchemwasspunon thesubstratetomatchtheheightoftheelectrodes.Thechannelsweremade10 m widerthanthespaceinbetweentheelectrodestomakealignmentbeforeexposure 120

PAGE 121

Figure6-2.IllustrationoftheDEPseparationdeviceintroducedbyShinetal.SampleP istheinitialsurfactant-suspendedSWNTs,whilethebufferisasurfactant solution.Thedeviceissupposedtoproduceastreamenrichedinm-SWNTs SampleMandanotherenrichedins-SWNTsSampleS.Thegurewas takenfromShinetal. 205 easier.ThechannelsweresealedwithaslabofPDMSthatwaspreviouslycoatedwith a5 mlayerofSU-8.ThePDMSslabhasholesinitthatserveastheinletandoutlet reservoirs.Afteralignment,thetopSU-8layeriscross-linkedwiththebottomlayerby oodexposureandbakingat393Kfor20min. Experimentswithtwootherdesignswereperformedaswell.Onedesignconsisted ofamicrochannelwithinterdigitatedelectrodesatthebottomwallofthechannel.The channelhadacrosssectionof300 20 m 2 andwas2cmlong.Thechannelhadone inletandoneoutlet.Individualelectrodeswere40 mwideand300 mlong.Thegap betweenadjacentelectrodeswas20 m.Electrodeswereplacedalongthewholelength ofthechannel.TheotherdesignwasintroducedbyShinetal. 205 Thesystemconsisted ofanH-channelwithembeddedelectrodesatthebottomwall.Thechannelhadacross sectionof500 150 m 2 .TheelectrodedesignisshowninFigure6-2.Theup-and down-electrodeswere200 mand250 mwideatthebase,respectively.Theminimum distancebetweentheelectrodeswas50 m.Thepartofthechannelwithelectrodes was2.5cmlong. 121

PAGE 122

Thefabricationofthosedesignsconsistedofthreesteps:fabricationofplanar electrodesonglassslides,fabricationofPDMSchannels,andbondingoftheglass slidesandthePDMSchannels.Microuidicchannelswerefabricatedusingsoft lithography 92 withPDMSasthestructuralmaterial.Briey,a4inchsiliconwafer waspatternedwiththenegativephotoresistSU-82015orSU-82150Microchem usingaprintedtransparencyastheshadowmask.A1/10ratiomixtureofPDMScuring agentandbaseSylgard184;DowCorningwaspouredoverthepatternedsilicon waferandcuredbyheatingat353Kfor2h.ThecuredPDMSwaspeeled-offfrom thesiliconwaferanddicedtoobtainindividualchips.Theelectrodeswerefabricated usingthelift-offtechnique.Briey,a100nmlayerofAuwassputteredonaglassslide thatwaspreviouslypatternedwiththeAZ9260photoresist.A10nmlayerofCrwas usedastheadhesionlayer.Aftermetaldeposition,theelectrodeswereobtainedby dippingtheglassslideinacetonetoremovethephotoresist.Thechannelswerebonded irreversiblytoglassslideswithelectrodesbybringingthechannelandtheglassslide intointimatecontactafterabrieftreatmentofthePDMSchipswithoxygenplasma AnatechSCE600Asher.Thechannelsandtheelectrodeswerealignedbyhandusing anstereomicroscope. 6.4ModelDescription SWNTsareparticleswithahighaspectratio,generallyhigherthan100,sotheywill bemodelledasrigidslenderbodies.SuspensionsofSWNTs,asproducedbytheHiPco process,aredispersedusingshearmixing,highenergysonicationandcentrifugation. Forthesystemofinterest,theconcentrationofSWNTsisafewmilligramsperliterand canbeconsidereddilute,with nL 3 1 where n and L arethenumberdensityandthe particlelength.Asaconsequence,thehydrodynamicinteractionsbetweennanotubes areignored. 97,209 Also,itisassumedthattherearenootherparticle-particleinteractions. 122

PAGE 123

6.4.1EquationsofMotion ThemotionoftheparticlesisdescribedbytheLangevinequation, m @ x @ t = F DEP )]TJ/F97 11.9552 Tf 11.955 0 Td [(F Drag + F Br where x isthecenterofmassvelocityoftheparticle, m ismass, t istime, F Drag isthe dragforce, F DEP isthedeterministicdielectrophoreticforceand F Br istheuctuating Brownianforce.DuetothesmalldimensionsofSWNTs,theinertialtermisnegligible andtheforcesontheparticlebalance, 209 F Drag = F DEP + F Br Thevelocityofanypointontherigidrod,atverysmallReynoldsnumbers Re ,is u x + s p )]TJ/F93 11.9552 Tf 11.955 0 Td [(u 1 x + s p = ln L = d 4 I )]TJ/F93 11.9552 Tf 11.955 0 Td [(pp f x + s p where u 1 istheunperturbedvelocityeldwhichisthevelocityoftheuidasifthe particlewasnotpresent.Theparticlevelocityis u s isthedistancefromthecenterof masstoanypointonthecenterlineofthetube, istheviscosityoftheuid,and d is theSWNTdiameter. 210 Theidentitymatrixis I .Theline-forcedensity, f ,isrelatedtothe totalforceactingontheparticle, F = Z L = 2 )]TJ/F43 7.9701 Tf 6.586 0 Td [(L = 2 f s ds wheretheforcesincludeboththeBrowniananddielectrophoreticforce.Theorientation vector, p ,istheunitvectoralignedwiththemajoraxisoftheSWNT.Thevelocityofany point s alongthelengthoftherigidrodcanbewrittenasthesumoftherateofchange ofthecenterofmassandorientationvector, u x + s p =_ x + s p 123

PAGE 124

Ifthepreviousexpressionfortheparticlevelocityisinsertedin6,andintegratedalong theparticlecenterline,from )]TJ/F42 11.9552 Tf 9.298 0 Td [(L = 2 to L = 2 ,anexplicitequationforthecenterofmass velocityisobtained, x = 1 L Z L = 2 )]TJ/F43 7.9701 Tf 6.586 0 Td [(L = 2 u 1 ds + ln L = d 4 L I + pp F DEP + F Br Thersttermontherighthandsideaccountsfortheunperturbedoweldwhilethelast termrepresentstheBrowniancontributiontotheparticlevelocity. Likewise,anexplicitexpressionfor p isgeneratedbycrossingequation6with s p andthenintegrating, p = 12 L 3 I )]TJ/F93 11.9552 Tf 11.956 0 Td [(pp Z L = 2 )]TJ/F43 7.9701 Tf 6.586 0 Td [(L = 2 s u 1 ds + 3 ln L = d L 3 I + pp h e F DEP + e F Br i where e F Br istheBrowniancontributionand e F DEP istheweightedlineforcedensity e F DEP = Z L = 2 )]TJ/F43 7.9701 Tf 6.586 0 Td [(L = 2 s f s ds Theweightedlineforcedensityiscloselyrelatedtothetorqueactingontherod, T DEP = p Z L = 2 )]TJ/F43 7.9701 Tf 6.586 0 Td [(L = 2 s f s ds and p T DEP = I )]TJ/F93 11.9552 Tf 11.955 0 Td [(pp e F Inderivingequations6,thevectoridentity p p e F DEP = )]TJ/F24 11.9552 Tf 9.299 0 Td [( I )]TJ/F93 11.9552 Tf 12.229 0 Td [(pp e F DEP was used. Thediffusivitiesfortheslenderbodymodelare: D T = B T )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 T I + pp and D R = B T )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 R I )]TJ/F93 11.9552 Tf 11.955 0 Td [(pp where B T isthethermalenergy.Themobilities )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 T and )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 R aregivenby: )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 T = ln L = d 4 L and )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 R = 3 ln L = d L 3 124

PAGE 125

Theuctuation-dissipationtheoremdenesthestatisticalcharacteristicsoftheBrownian forceandtorque,demandingthatthemeaniszeroandthevarianceequals 2 B T t )]TJ/F42 11.9552 Tf -455.188 -23.908 Td [(t 0 F Br t =0, F Br t F Br t 0 =2 D )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 T t )]TJ/F42 11.9552 Tf 11.955 0 Td [(t 0 T Br t =0, T Br t T Br t 0 =2 D )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 R t )]TJ/F42 11.9552 Tf 11.955 0 Td [(t 0 where t )]TJ/F42 11.9552 Tf 12.192 0 Td [(t 0 istheDiracdeltafunctionandtheanglebrackets, hi ,indicateensemble averages. 211,212 Thecenterofmassandrotationaldisplacementsresultingfromthe Brownianforcesandtorques,togetherwiththeequationsofmotion,aregivenby, x Br t =0, x Br t x Br t 0 =2 D T t )]TJ/F42 11.9552 Tf 11.956 0 Td [(t 0 p Br t =0, p Br t p Br t 0 =2 D R t )]TJ/F42 11.9552 Tf 11.955 0 Td [(t 0 Fromtheseexpressions,thenitestochasticdisplacementsaregivenas, x Br t = D T t 1 = 2 W T p Br t = D R t 1 = 2 W R where W T and W R arerandomvectorswiththefollowingcharacteristics, h W T t i =0, h W T t W T t 0 i = I t )]TJ/F42 11.9552 Tf 11.955 0 Td [(t 0 h W R t i =0, h W R t W R t 0 i = I t )]TJ/F42 11.9552 Tf 11.955 0 Td [(t 0 Inthevicinityofthechannelwalls,arepulsiveforcethatisexponentiallydecayingwith distanceisimposedontheparticles.Theforceusedforrepulsionisexpressedas, F w = 0.01 u 1 x T L )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 exp )]TJ/F24 11.9552 Tf 9.299 0 Td [(10 h = L 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [(exp )]TJ/F24 11.9552 Tf 9.299 0 Td [(10 h = L n where h istheclosestdistancebetweenthetipofthenanotubeandthewall, u 1 x isthe x -componentof u 1 and n isaunitvectorperpendiculartothewall.Thesignoftheforce densityisdeterminedsothatthetubeispushedawayfromthewall.Thisshort-range 125

PAGE 126

repulsiveforcepreventsthetubefromcrossingthechannelboundaries.Thereisalso atorqueassociatedwiththerepulsiveforce.However,whenastrongDEPattraction towardstheelectrodesoverwhelmsBrownianmotionanduidow,thereisahigh probabilitythattheparticlewillbetrappedormoveextremelyslowonceinthevicinityof theelectrodes;hence,thecalculationisstopped. 6.4.2EvaluationoftheDielectrophoreticForce Theforcethataparticleinadielectricmediumexperienceswheninteracting withanon-uniformelectricelddependsonthepropertiesandsizeoftheparticle, thepropertiesofthemediumandthestrengthandnon-uniformityoftheelectriceld. Generally,thedielectrophoreticforceiscalculatedusingexpressionsthatinclude momentsoftheelectriceldandtheinduceddipolemoment.Inthepresentwork,only expressionstruncatedattherstordertermareused, F DEP = P r E x where P istheinduceddipolemomentontheparticleand r E representsthegradient oftheelectriceldevaluatedatthecenteroftheparticle.Thisassumptionisgenerally validwhenparticlesarewidelyspacedornon-interacting,andwhentheelectriceld changeslinearlyoverthelengthoftherod. 195 Thetorqueduetotheelectriceldequals thecrossproductoftheeffectivedipolemomentandtheelectriceld, T DEP = P E x Noticethatthetorquedependsonthemagnitudeandorientationoftheelectriceld insteadofthegradientoftheelectriceld.Thedielectrophoreticforceandtorque dependontimeastheelectriceldisrapidlyvarying.However,inasinusoidaleld,only theaverageofthedielectrophoreticforceandtorqueisconsideredsincetheviscosityof 126

PAGE 127

theuiddampensthetimedependentcomponent, h F DEP i = 1 2 Re r P E and h T DEP i = 1 2 Re P E Theeffectivedipolemomentapproximationisusedtondanexpressionfortheinduced dipolemoment.Inthisapproximation,theeffectivedipolemoment,asdenedbyJones, isthemomentofanequivalent,freecharge,pointdipolethat,whenimmersedinthe samedielectricliquidandpositionedatthesamelocationasthecenteroftheparticle producesthesamedipolarelectrostaticpotential. 195 Foradielectricandconductive material,theeffectivedipolemomentiswrittenasfollows, P = E Thepolarizabilityis anditisgivenas = 4 Ld 2 3 m Re K Inexpression6, Re K extractstherealpartof K ,theClasius-Mossottifactor,and m isthepermittivityofthemedium.Thevalueof K dependsonthefrequencyofthe appliedelectriceld.Theequationforthedielectrophoreticforceandtorquemustbe writtenappropriatelytoaccountfortheanisotropicdielectricpropertiesofSWNTs, 213 F DEP = 1 2 r k E k E k + ? E ? E ? T DEP = k E k E ? + E ? E k Inequation6and6theelectriceldandthepolarizabilityhavebeenwrittenin termsoftheparallelandperpendicularcomponentstotheparticlemajoraxis. E k = pp E and E ? = I )]TJ/F93 11.9552 Tf 11.955 0 Td [(pp E 127

PAGE 128

Thevalueof K andconsequentlythepolarizabilityareestimatedbyapproximatingthe carbonnanotubesasprolateellipsoidswithhighaspectratios, K = e p k ? )]TJ/F33 11.9552 Tf 11.648 0 Td [(e m e m + e p k ? )]TJ/F33 11.9552 Tf 11.648 0 Td [(e m e = )]TJ/F42 11.9552 Tf 11.955 0 Td [(i where e isthecomplexpermittivity, theconductivityofthemediumorparticleand is theangularfrequency.Thedepolarizationfactorfactor, L p ,parallelandperpendicularto themainaxisisgivenas, L p k = d L 2 ln 2 L d )]TJ/F24 11.9552 Tf 11.955 0 Td [(1 L p ? = 1 2 Inaowingsystem,itisveryimportanttoproperlyaccountfortheanisotropic propertiesofSWNTsbecausetheorientationandthereforethepositionaredetermined bythreecompetingfactors:Brownianmotion, T DEP and T Drag .TheBrownianmotion hasthetendencytorandomizetheSWNTs.The T DEP tendstoalignthemajoraxisof theSWNTswiththeelectriceld,whereas T Drag tendstoalignthemajoraxiswiththe ow. 195 6.4.3NumericalIntegrationoftheGoverningEquations ThepositionandorientationofeachSWNTasafunctionoftimeareobtainedby integratingequations6and6usinganexplicitrstorderEulermethod, 214 x = x 0 +_ x t p = p 0 +_ p t However,duetothestochasticcontribution,acorrectiontermmustbeincludedinthe rotationalvelocityexpression,6.Thecorrectiontermisnecessarytoavoiderrors thatcanaccumulateafteranumberoftimesteps,anditmaintainstheaveragelengthof theunitvector p duringeachtimestep.Thecorrectiontermisequaltothedivergenceof 128

PAGE 129

therotationaldiffusivity, @ @ p D R = )]TJ/F27 11.9552 Tf 9.298 0 Td [( B T )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 R p Adetailedexplanationofthenatureandderivationofthecorrectiontermisgivenby CobbandButler. 211 Afteraddingthecorrectionterm,equation6takesthefollowing form, p = p 0 +_ p t + )]TJ/F27 11.9552 Tf 5.479 -9.684 Td [( B T )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 R p t Inaddition,thevector p mustberenormalizedaftereachtimesteptoensurethat p isof unitlength.Therenormalizationisnecessarytocorrectforthedeterministicforcingdue tothedielectrophoreticforce. 6.4.4PropertiesofSWNTsandtheSuspendingMedium ItisnaturaltoexpectthatthedielectricpropertiesofsuspendedSWNTsnotonly dependontheintrinsicpropertiesofSWNTsbutalsoonthesurfactantshellaround thenanotubes.Hence,inmanyworksconductivityvaluesusedasinputtothemodel equation6spanseveralordersofmagnitude.Somehaveusedtheintrinsic valuesofSWNTsconductivityi.e., 10 6 10 8 S/m 203 andothersavalueoftheorderof 0.3S/mforsemiconductorandmetallicSWNTsassumingthatsurfaceconductance duetoionicsurfactantchargedominatesthepolarizationofthesurfactant-SWNT complex. 213,201 Othershaveassumedthattheconductivityform-SWNTsisatleast 1000timesthatofs-SWNTs. 206 Similarly,theexactvaluesofthedielectricconstant forsurfactant-suspendedSWNTsarenotknown,andusuallyordersofmagnitude approximationsareusedasinputstothemodel.Theoreticalpredictionsestimate thattheintrinsicvaluesof k fors-SWNTsarearound10. 215,216 Experimentalvalues areingoodagreementwiththeoreticalpredictions.Forexample,usingabsorption spectroscopyandpolarizationdependentRamanscattering,valuescloseto5were obtainedforthelongitudinalcomponentofthedielectricconstant k ofs-SWNTs dispersedinapolymermatrix. 217 Luetal.measuredtheintrinsiclongitudinaland transversecomponentofthedielectricconstant k and ? ofindividualSWNTsthrough 129

PAGE 130

atomicforcemicroscopy. 218 Theyfoundthat ? fors-SWNTshadvaluesbetween5and 10,whileavalueof30wasmeasuredfor k .Ontheotherhand,theauthorsmeasured valuesof10and1000for ? and k inm-SWNTs. Ifarangeofvaluesforthepropertiesareused,asoutlinedabove,thevalue ofReKwouldspanseveralordersofmagnitudei.e1to10000whichcreates considerableuncertaintyforthevalueoftheDEPforce.Thevaluesusedinthe calculationsareoutlinedbelow.Inthecaseofs-SWNTs,the p k and p ? weretakento be0.35S/m. 213,201 Thevalueof p ? form-SWNTwasassumedtobe0.5S/m,while p k wasallowedtotakevaluesof 1 10 3 1 10 4 and 1 10 5 .Fors-SWNTspecies, p k and p ? wereassumedequalto5and8,accordingtoexperimentalmeasurements. 217 Ontheotherhand,thevaluesof p k and p ? form-SWNTswere1000and8. ThecalculationswereperformedforSWNTssynthesizedthroughtheHiPco process,whichproducesSWNTswithanaveragediameterof1nm.Thediameterof thetubesplusthesurfactantshellwasassumedtobe2nm,whichimpliesthatthe surfactantmoleculesarelyingatontheSWNTsidewall. 100,57 ThelengthofSWNTs affectsthemagnitudeofthedifferentforcesandtorquesactingonthem.Suspending SWNTsinaqueousmediarequiressonicationtobreakupthebundlesofSWNTs.The processofsonicationcutsSWNTsandcreatesanensemblepolydisperseinlength. Thelengthdistribution,asdeterminedbyAFMmeasurements,canbemodeledusinga Weibulldistributionfunction, W L = A B L )]TJ/F42 11.9552 Tf 11.955 0 Td [(L min B )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 e [ )]TJ/F43 7.9701 Tf 6.587 0 Td [(A L )]TJ/F43 7.9701 Tf 6.587 0 Td [(L min B ] where A and B arethescaleandshapeparameters. 206 Thedistributionisgivenfor valuesof L above L min .Figure6-3showsthelengthdistributionusedforthecalculations. SWNTsaredispersedinwaterwiththeaidofsurfactants.Hence,themedium consistsofwaterwithfreesurfactantmicellesandmonomerinsolution.Itisassumed thatthepresenceofsurfactantdoesnotchangethepermittivityofwater,sothevalue 130

PAGE 131

Figure6-3.WeibulldistributionofSWNTlength.Thevaluesof A B and L min were 6 10 )]TJ/F25 7.9701 Tf 6.586 0 Td [(6 ,2.29and300nm. forpurewaterisused, m =80 .Theconductivityofthemediumisaffectedbythefree surfactantmonomers,so0.1S/m,whichisthevalueofwatercontaining1wt%SDBS,is used. 213 6.4.5EvaluatingtheDevicePerformance Theperformanceofthedeviceishighlydependentontheresidencetimeofthe particlesinsidethechannel t r aswellasonthetimeittakesaparticletomigratefrom onepartofthechanneltoanother t mig undertheinuenceof F DEP .Theresidence timeisgivenbytheratioofthechannellengthandthemeanvelocityinsidethechannel, t r = L ch = U .Itisassumedthatasthechannellengthincreases,thenumberofelectrodes increasesaswell.Ontheotherhand, t mig isgivenbytheratiooftheaveragedistance necessarytomovefromthecentralstreamtothebottomstreamandtheaverage dielectrophoreticvelocityinthey-direction, t mig = L d = U DEP .Anapproximatedvalueof thedielectrophoreticvelocityisgivenby, U DEP Re K k +2 Re K ? ln L = d d 2 m 144 V 2 p L 3 el @ E 2 @ y avg Asshownbythepreviousexpression,thevalueof U DEP dependsontheSWNT properties,specicallytheirsizeandvalueof K ,andtheelectriceld. U DEP is dependentonthenon-uniformityoftheelectriceld @ E 2 @ y avg ,theappliedvoltage 131

PAGE 132

V p andthedistancebetweentheelectrodes L el .Theperformanceofthedevice isultimatelydependentontheoperatingparameters, V p and U ,aswellasonthe geometryofthesystem,asspeciedby L ch and L el .Foragivenvalueof F DEP ,if t r is tooshortforthem-SWNTtobedeected,bothm-ands-SWNTsremaininthecentral areaofthechannel.Ifenoughtimeisallowed,thetrajectoryofm-SWNTscouldbe changed,andtheycouldbecollectedinthebottomchannel.Ontheotherhand,fora longertimetheparticlewouldbedirectedtowardstheelectrodesandgettrapped.Also, when t r islong,thes-SWNTspeciescandiffusefromthecentralareaofthechannel tothebottomstreamandconsequentlyreducethepurityofm-SWNTs.Theoperating parameterscanbechosentomakeanyofthosesituationsdominant.Theeffectofthe operatingparameters V p and U andtheeffectofthechannellengthorthenumberof electrodeswillbeexploredlater. 6.4.6NumericalCalculationofFlowandElectricFields Electricandoweldsarecalculatednumericallyusingtheniteelementmethod withCOMSOLMultiphysics.Themediumisconsideredahomogeneousandlinear dielectricwithzerofreechargedensity.Hence,theelectriceldiscalculatedbysolving Poisson'sequation, r r =0, where istheelectricpotential.Theelectrodesaredescribedbyequipotentialboundary conditions.Theinterfacebetweenthewallsofthechannelandtheliquidmediais describedbythejumpcondition, i m + m @ m @ y = i w + w @ w @ y GiventhelowpermittivityandconductivityofglassorPDMSincomparisontothe aqueousmedia,thejumpconditionsimpliestotheNewmancondition, 219 @ m @ y =0. 132

PAGE 133

Thevalidityofthisapproximationwastestedbysolvingmodelsthatincludedthe channelwallsinthecalculationdomain.Itwasobservedthatforthesystemofinterest theNewmanconditionadequatelydescribethepotentialinthevicinityofthechannel walls.Poisson'sequationisscaledbytheamplitudeoftheappliedsignalandthe characteristiclength L c .Thecharacteristiclengthisdenedastheheightofthe channel.Thescaledboundaryconditionsare, 0 =1, 0 = )]TJ/F24 11.9552 Tf 9.299 0 Td [(1, and @ 0 @ y =0, where wasscaledbytheappliedpotentialorvoltage V p .Theoweldiscalculated bysolvingtheStokesequationforanincompressibleuid, 0= r P + r 2 u 1 where P istheuidpressure.TheStokesequationisalsosolvedinthescaledformas, 0= r P 0 + 1 Re r 2 u 1 where u 1 hasbeenscaledbythecharacteristicmeanvelocityinthechannel U .The pressurewasscaledby U 2 ,with beingthedensityoftheuid.Thenon-slipboundary conditionisappliedatthewallsofthechannel,whereasaninputvelocityproleisgiven attheentranceofthechannel.Theboundaryconditionatthechannelexitisavalue forthepressure,usually P =0 .Inthenon-dimensionalform,theeldequationsare onlysolvedonceforagivenelectrodeandchannelgeometry.Therefore,avarietyof conditions V 0 and U canbeexploredwithminimalcomputationaleffort. 6.4.7PerformingtheCalculations ThecalculationswereperformedusingaFORTRANcode.Particlesarerandomly placedattheentranceofthechannel.Theelectricandoweldsolutionsobtainedfrom COMSOLaremappedontoa3Dregulargrid.Asthecalculationproceeded,theelectric andoweldvalueswereevaluatedthroughalinearinterpolationroutineasneeded. 133

PAGE 134

Figure6-4.Cartoonshowingtheproceduretointerpolatethevaluesofthevelocityand electriceldsatdifferentpositionsduringtheimplementationofthecode. Figure6-4depictstheprocedureusedtoperformtheinterpolation.First,itisnecessary tondtheeightpointsP i closesttothepointinquestionPwheretheelectriceld orvelocityareknown.ThevaluesofthevelocityorelectriceldsatParecalculated byrstobtainingthevaluesatK3usingtheapproximatedvaluesatK1andK2.The sameprocedureisusedtocalculatethevaluesatK6,onthesecondz-plane.Then,the componentsof u 1 and E atPareapproximatedbyinterpolatingbetweentheprojected valuesontheupperandlowerz-planes,K6andK3respectively,asshownfor E y inthe followingexpression, E y = E y K6 )]TJ/F42 11.9552 Tf 11.955 0 Td [(E y K3 z z P )]TJ/F42 11.9552 Tf 11.955 0 Td [(z P1 + E y K3. SolutionofCOMSOLmodelscanbememoryintensive.Hence,onlymodelsfor shortchannelswithafewelectrodesarefullysolvedinCOMSOL.Theowforlonger channelsareapproximatedbyrepeatingthesegmentawayfromtheinletandoutlet oftheshortchannelwheretheowisconstant.Similarly,theelectriceldinlonger channelsisobtainedthroughtherepetitionofaperiodicunitcellasshowninFigure6-5. TomapthedatafromthesolutionoftheCOMSOLmodelintoaregular3D-grid,the domainwasdividedintotwozones.Therstzoneistheareacoveringonlythecentral 134

PAGE 135

Figure6-5.Unitcellsforowandelectriceldcalculationsandtheapproximationof geometriesthatdemandalargernumberofelements.Thedomainzone1 andzone2thataremappedintotheregular3D-gridarealsoshown. inletchannel,whilethesecondzoneistheareabetweentheinletandoutletchannels asshowninFigure6-5.ThelecontainingthegridisoutputfromCOMSOLwiththe followingformat: Row.Numberofnodesforzone1#nodes1 Row.Initialcoordinatesforzone1 Row.Numberofnodesalongeachcoordinateforzone1 Row.Stepsizeforeachcoordinateforzone1 RowRow+numberofnodes. x u 1 and E forzone1 Row+#nodes1+1.Numberofnodesforzone2 Row+#nodes1+2.Initialcoordinatesforzone2 Row+#nodes1+3.Numberofnodesalongeachcoordinateforzone2 Row+#nodes1+4.Stepsizeforeachcoordinateforzone2 Row.... x u 1 and E forzone2 Asmentionedpreviously,particlesarenotallowedtocrosstheboundariesofthe channelbecausearepulsiveforceisimposedontheboundaries.IftheDEPforceis largeenough,particlesmightgettrapped,moveextremelyslowinthevicinityofthe boundaries,andconsequentlyinduceaninniteloop.Therefore,thecalculationfor aparticletrajectoryisstoppediftheparticlegetsclosetotheelectrodesbyagiven distance.5 m,andtheparticleiscountedasbeingtrappedontheelectrodes. 135

PAGE 136

6.5ResultsandDiscussion 6.5.1DevicePerformance Thedeviceperformancewasevaluatedusingdifferentvaluesof V p U ,and L ch Thevaluesof V p U and L ch werevariedfrom5to17.5V,1to16mm/s,and1.8to4.8 mm,respectively.Figure6-6showshowthedeviceperformancechangesastheapplied voltageincreases.Figure6-6ashowsthefractionofm-SWNTsthatarerecovered throughOutlet1Recoveredm-SWNTs.Theplotalsoshowsthetotalamount ofm-SWNTsthatareremovedfromthecentralstream,whichincludesm-SWNTs exitingthroughtheOutlet1andthosetrappedontheelectrodesortheirvicinityTotal m-SWNTsremoved.Figure6-6bshowsthepurityofrecoveredm-SWNTsaswellas thepurityofSWNTsrecoveredthroughtheOutlet2Recovereds-SWNTs.Asthe valueof V p increases,thetotalnumberofm-SWNTsremovedfromthecentralchannel increaseslinearly.Asaconsequence,thepurityofrecovereds-SWNT,increases linearlyaswell,asshowninFigure6-6b.Theamountofm-SWNTsthatcanbe recoveredincreasesupto V P =12.5 Vbutdecreasessubstantiallyforlargervalues of V p .Thepurityoftherecoveredm-SWNTstreamincreaseslinearlyupto V p =10 Vbutdecreasesslowlywithlargervaluesof V p .Thereasonisthatasthevoltage increases,alargernumberofm-SWNTsgettrappedontheelectrodesinsteadofexiting fromOutlet1. Figure6-7showshowthedeviceperformancechangeswiththemeanvelocity insidethechannel.Thetotalamountofm-SWNTsthatareremovedfromthecentral channeldecreasesas U increases.Thepurityofthecentralstreamdecreasesina similarway.Thefractionofrecoveredm-SWNTsinitiallyincreasesbutlevelsoffwhen U =8 mm/s.Thepurityofrecoveredm-SWNTsincreasesrapidlyas U changesfrom1 to8mm/s,butlevelsoffatavalueof0.95at11mm/s. Figure6-8showshowthedeviceperformancechangeswiththechannellength ornumberofelectrodes.Thetotalamountofremovedm-SWNTsaswellasthepurity 136

PAGE 137

Figure6-6.Effectofthemagnitudeoftheappliedvoltagepeakvoltage, V p onthe deviceperformance.Theresultswerecalculatedwiththefollowing parameters: U =8 mm/s, L ch =30 mm,andforthemetallicspecies, k =1000 .aFractionofm-SWNTsthatarerecoveredfromthebottom channel,andthetotalfractionofm-SWNTsthatisextractedfromthecenter channel.bPurityoftherecoveredfractionsthatareenrichedins-SWNT centralchannelandm-SWNTsbottomchannel. oftherecoveredm-SWNTsincreaseslinearlywith L ch .Theamountofrecovered m-SWNTspeakandthendecreaseslinearly.Thepurityofrecoveredm-SWNTs decreasesas L ch increases.Thisisduetoanincreaseinthenumberoftubesthatare trappedandthediffusionofs-SWNTsintothebottomstream. Asasummary,toremovethemaximumamountofm-SWNTs,itisbesttoapplythe highestvoltagepossible.However,thiscausestrappingofahighnumberofm-SWNTs. Toavoiddiffusionofs-SWNTsintothebottomstream,itisbestnottousealong channel.Ontheotherhand,tomaximizetheamountofm-SWNTsthatarerecovered andthepurityofthestream,itisbesttouseahigh U .However,thetotalamountof removedm-SWNTandconsequentlythepurityofthecentralstreamwouldbelow. Thedifferentregimesunderwhichthedevicecanoperatecanbecharacterized bytheratioof t mig and t r = t mig = t r .Table6-1showstherangeof thatproducesa givenseparationperformance.Forexample,torecoveralargeamountofm-SWNTs, avaluesof between2.08and8.5arenecessary.Asstatedearlier,thisconditionis attainedwhentheresidencetimeisshorterthatthetimeofmigration.Thatcanbe 137

PAGE 138

Figure6-7.Effectofthemeanvelocityonthedevice'sperformance.Theresultswere calculatedwiththefollowingparameters: V p =12.5 V, L ch =30 mm,andfor themetallicspecies, k =1000 .aFractionofm-SWNTsthatarerecovered fromthebottomchannel,andthetotalfractionofm-SWNTsthatisextracted fromthecenterchannel.bPurityoftherecoveredfractionsthatare enrichedins-SWNTcentralchannelandm-SWNTsbottomchannel. accomplishedusingshortchannelswithhighvaluesof U .Ontheotherhand,large removalofm-SWNTsandconsequentlyhighpurityofs-SWNTsareattainedwhen theresidencetimeislargerorslightlylowerthanthemigrationtime.Thatcanbe accomplishedusinghighvoltages,longchannelsandsmallvaluesof U Table6-1.Operationalregimeofthedevice. TargetCriteriaRangeof GoodRecoveryofm-SWNTs0.20.3172.08 8.5 LargeRemovalofm-SWNTs0.71.00.13 2.29 HighPurityofm-SWNTs0.71.01.38 15 HighPuritys-SWNTs0.91.00.13 2.08 6.5.2ExperimentalValidation Attemptsweremadetoexperimentallyconrmthemodelpredictionsforthe designinFigure6-1.However,oureffortswerefrustratedbyfabricationissues. Themostprominentproblemwasthebondingandsealingofthechannelafterthe electrodepositionoftheelectrodesandfabricationoftheSU-8walls.Althoughitwas possibletoassemblethedevice,thechannelsleakedeasily.Anothercommonproblem wasthedelaminationofSU-8fromtheglassslide.Evenwhena2 mthickSU-8layeris 138

PAGE 139

Figure6-8.Effectofthechannellengthnumberofelectrodesonthedevice performance.Theresultswerecalculatedwiththefollowingparameters: V p =12.5 V, U =8 mm/s,andforthemetallicspecies, k =1000 .a Fractionofm-SWNTsthatarerecoveredfromthebottomchannel,andthe totalfractionofm-SWNTsthatisextractedfromthecenterchannel.b Purityoftherecoveredfractionsthatareenrichedins-SWNTcentral channelandm-SWNTsbottomchannel. depositedpriortothepatterningoftheseedlayerseeSection6.3,delaminationtook place. Experimentswereperformedwithchannelsthathadinterdigitatedelectrodesonthe bottomwall,asdescribedinSection6.3.Thefabricationofthesedeviceswaseasier. Theobjectivewastoenrichs-SWNTsbytrappingm-SWNTs. 203 However,therewere problemswithelectrodedegradationwhenvoltagesabove1Vpeak-to-peakwere applied.Bridgingoftheelectrodesbythedepositednanotubeswasalsoacommon problem.Tominimizebridging,adevicewithvechannelsinparallelwasused.The channelshadonecommoninletand5differentoutlets.Themaindifcultytoachieve separationinallthechannelswastheunevendistributionoftheowthroughthe channels.Forexample,theowwouldsometimespassthroughthechannelwiththe leastresistanceratherthanevenlydistributeamongthevechannels. ThesystemintroducedbyShinetal. 205 wasbuilttoconrmandfurtherexpand theirresults.ThissystemrequireshydrodynamicfocusingofSWNTstoonesideof thechannel.Hydrodynamicfocusingisachievedbyankingthesamplestreamwitha 139

PAGE 140

Figure6-9.NIRimageofSWNTsthathavebeenfocusedwhileowingthrougha microchannel.Themicrochannelhadacrosssectionof20 60 m 2 andwas fabricatedthroughsoftlithography.Theowwasprovidedbyasyringepump HarvardPicoPlus.TheSWNTsweresuspendedina1wt%solutionof SDBS.A40 water-immersedobjectivewasused,andtheacquisitiontime was20ms. sheathstream.Thesamplestreamiscompressed,anditswidthconsequentlydepends ontherelativeowrateofthesampleandsheathstreams.Figure6-9isanNIRimage ofasuspensionofSWNTsowingthroughamicrochannel,anddemonstratesthat SWNTshavebeenconnedtoonesideofthechannelbyasheathstream.The focusedstreamappearsasabrightareaontherightsideofthemicrochanneldueto theuorescenceofs-SWNTs.However,inthefocusedSWNTsuspension,individual nanotubesarenoteasilyidentiedsincetheaveragelinearspeedisabout12mm/s. ThesheathstreamdidnotcontainSWNTs,butithadthesameconcentrationof surfactantastheSWNTsuspension.Althoughthefabricationandassemblyofthe separationsystemwasrelativelyeasy,operationaldifcultieswereencountered.Itwas difculttoachievehydrodynamicfocusingofSWNTsforextendedperiodsoftime.Since thetargetowratewas60 L/hrandtheminimumvolumeneededforcharacterization purposeswas200 L,atleast6.6hrwerenecessary.Usingsyringepumps,itwasnot 140

PAGE 141

possibletocontroltheowconditionstocarryouttheseparationsuccessfullyforthis lengthoftime. 6.6Conclusions Browniandynamicssimulationswereperformedtoassesstheperformanceofa microuidicdevicetoseparateamixtureofm-ands-SWNTs.Theeffectofoperational parameterssuchasthemeanowvelocityinsidethechannelandtheappliedvoltage wasdetermined.Theeffectofthechannellengthwasalsostudied.Itwasshownthatby choosingtherightchannelgeometryandoperationalparameters,itispossibletoobtain fractionsthatareeitherhighlyenrichedinm-SWNTsors-SWNTs.Theresultsshow thatobtaininghighlyenrichedfractionsinm-ors-SWNTssimultaneouslyisdifcult withtheproposeddevicegeometry.Experimentalvalidationwasattempted.However, therewerefabricationandoperationalissuesthatthwartedthesuccessfulexperimental assessmentofthedevices. 141

PAGE 142

CHAPTER7 CONCLUSIONSANDFUTUREDIRECTIONS 7.1Conclusions TheseparationofSWNTsbytype,diameterorchiralityrequiresanabilityto characterizeandcontroltheinterfacesaroundnanotubes.Theworkpresentedinthis dissertationhasprovidednewinsightsintothecharacteroftheinterfacesurrounding SWNTsanditsmanipulation.ThedifferentstructuresthatformaroundSWNTsarethen showntoaffectthechromatographicseparationofSWNTs. BothPLspectroscopyandSANSarecharacterizationtoolsthathelpdene thestructureandnatureoftheenvironmentsurroundingthenanotubes.Although PLspectroscopyiscommonlyusedtocharacterizethe n m typeofSWNTs,this workshowsthatPLemissionalsoprovidesdetailsregardingthelocalenvironment surroundingthesidewallofthenanotube.Thepermeabilityofthesurfactantassembly isobtainedbymeasuringthechangestoPLintensityasafunctionofthequencher concentrationinthebulkphase.Ontheotherhand,thePLpeakpositionindicates thepolarityoftheenvironmentaroundthenanotube.SANSprovidesadditionaldata relatedtothelocalenvironmentaroundtheSWNTs.SANSscatteringprolesareused todeterminethestructureoftheassembliesofsurfactantandsolventsurroundingthe nanotubes. ThesurfactantassemblysurroundingSWNTsishighlymobileanddynamic.PL spectroscopyandSANSstudiesshowthatthesurfactantassemblyaroundSWNTs canbemanipulatedbyshearingthesuspension,changingtheionicstrength,oradding immisciblesolvents.Thesechangescanbeunderstoodbyusingthepackingparameter, whichdescribesthetypesofsurfactantstructuresobservedwithmicellarsystems. Changingthepolarheadareachargescreeningorthehydrophobicvolumeofthe surfactantsolubilizationofnon-polarorganiccompoundsadjustthepackingparameter. Hence,theresponseofthesurfactantmolecules,andconsequentlythebehaviorof 142

PAGE 143

theinterfacedependsonthesoluteandthestructureofthesurfactantmolecules. ModifyingthesurfactantassemblycanenhancethePLquantumyieldofSWNTsdue tothechangesinthesurfactantpermeabilityandprovidebetterstabilizationofthe suspensions.Theseresultssuggestthatsurfactantmoleculesoftendonotassemblein afashionthatreturnsthehighestPLintensity. ControllingtheSWNTinterfaceisimportantinthedevelopmentofseparation processesthatarefastandscalable,suchasthosethattakeadvantageofthepartition ofanalytestosolid-liquidchromatographyinterfaces.Thesurfactantstructure surroundingSWNTsidewallswasshowntoaltertheretentioncharacteristicsof nanotubesthatpassthroughagarosecolumns.Thesystematicmanipulationof thesesurfactantstructuresenabledtheelucidationoftheretentionmechanism inagarose-basedseparations.TheseresultsshowedtheretentionofSWNTson non-functionalizedagarosegelsisduetoselectiveadsorptionratherthansize-exclusion. Theabilitytocharacterizethesurfactantstructuresalsoshowsthatexposedregionsof theSWNTsidewallisnotresponsiblefortheadsorptionofSWNTsontheagarosegel. Instead,thesurfacedensityandorientationofSDSmoleculesarethemostimportant parameters.Theseresultsimplythatfurthermanipulationandcontrolofthesurfactant shellcouldprovidebetterseparations. 7.2FutureDirections TobetterunderstandtheeffectofthepolarenvironmentaroundSWNTsonthe excitonicprocesses,PLlifetimemeasurementsofSWNTsshouldbeperformedas theirimmediatemicroenvironmentissystematicallychanged.Thesestudies,whichare similartothosedescribedinChapter3,wouldallowtheroleofthelocalizedenvironment onnon-radiativedecaymechanismstobedetermined.Inaddition,theroleofthe surfactantstructureinexcitonicprocessesshouldbecharacterized. Regardingtheagarose-basedseparationsdiscussedinChapter5,twomechanistic issuesremainedtobeaddressed.First,abetterunderstandingofthenatureofthe 143

PAGE 144

interactionsbetweenSDS-coatedSWNTsandtheagarosematrixisneeded.Along thesamelines,furtherinvestigationonthecauseoftheselectiveretentionofsoverm-SWNTsisneeded.Thiscouldprovideaguidetoselectchromatographic mediadifferentfromagarose-basedmaterialstoimprovetheseparationperformance, operabilityandoverallcost. SANScanalsohelpusunderstandtheprocessofswellingandtheresultant structuresinothersurfactant-SWNTsystems.AparticularsystemofinterestisSWNTs suspendedusingSDBS.Ontheotherhand,futurestudiesshouldalsofocusonthe effectofothersurfactant-modicationprocesses,suchaselectrolytetuningcharge screening. 144

PAGE 145

APPENDIXA QUANTUMYIELDMEASUREMENTSANDTHEEFFECTOFBATHSONICATIONON PHOTOLUMINESCENCEAFTERULTRACENTRIFUGATION A.1FlowParameters Theowrate Q ofSWNTsuspensionsthroughthemicrouidicchannelsspanned arangefrom20to4000 L/h.However,experimentaldataisonlyreportedforow ratesbelow2500 L/hsincethemicrouidicchannelwaspronetofailure.TableA-1is acompilationoftheaveragevelocity U m ,averageshearrate,andtheReynoldsnumber characterizingtheowthroughthemicrouidicchannelatdifferentowrates. TableA-1.SummaryoftheSWNTsuspensionowparameterswhileowingthroughthe microuidicchannels. Q L/h Q m 3 /s U m m/sShearRates )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 Re 205.56E-124.63E-038.58E+020.14 1002.78E-112.31E-024.29E+030.69 3008.33E-116.94E-021.29E+042.08 5001.39E-101.16E-012.15E+043.47 7001.94E-101.62E-013.00E+044.86 10002.78E-102.31E-014.29E+046.94 12503.47E-102.89E-015.36E+048.68 15004.17E-103.47E-016.44E+0410.42 20005.56E-104.63E-018.58E+0413.89 25006.94E-105.79E-011.07E+0517.36 30008.33E-106.94E-011.29E+0520.83 35009.72E-108.10E-011.50E+0524.31 40001.11E-099.26E-011.72E+0527.78 FigureA-1aandA-1bshowsthattheuorescenceintensityfromthe,5and ,5SWNTtypesintheinitialSDS-SWNTsuspensiondecreasebyatleastanorder ofmagnitudeatacidicpHvalues,similartothebehaviorof,7and,3species.The smallspecies,,5,intheshearedsuspensionshowsPLintensityvaluesthatareclose toanorderofmagnitudehigherthanintheinitialsuspensionatlowpHvalues. A.2SWNTPLasaFunctionofTime FigureA-2showsthebehaviorof,7and,5speciesasafunctionoftimeinthe initialandshearedsuspensions.Noticetheincreaseinintensityfromthe,7speciesin 145

PAGE 146

FigureA-1.PLintensityofa,5andb,5SWNTspeciesinaninitialandsheared SWNT-SDSsamplesatdifferentpHvalues.Theappliedshearratewas 0.4 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 theinitialsuspension,whileintheshearedsuspensiontheintensitydecreasesgradually duringthe14weeksoftheexperiment. SWNTPLasaFunctionofSonicationTimeSometimessuspensionsyieldedlow PLemissionintensitycomparedtoothersuspensionspreparedbythesameprocedure, asshowninFigure2-5a.ThesedifferencesinPLintensityoccurwithoutsignicant differencestotheabsorbancespectraofthesuspensions,asshowninFigureA-3. ToinvestigatewhetherthelowPLintensitywasduetobundlingofthespeciesin suspension,thesuspensionwithpoorPLemissionwassonicatedforaperiodof430 146

PAGE 147

FigureA-2.PLintensityofa,7andb,5SWNTspeciesinaninitialandsheared SDS-SWNTsamplesatdifferenttimesaftershearingthesuspension.The appliedshearratewas0.4 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 mininabathsonicator.Aliquotsofthesuspensionwerewithdrawnfromthevialat differentpointsintimetomeasurethePLemissionspectra.FigureA-4ashowsthe spectraofthesuspensionatdifferentpointsintime.Itisobservedthatasthesonication timeincreases,thePLemissionintensityfromthespeciesinthesuspensionincreases gradually.Themaximumincreaseinintensityisobservedafter230minofsonication. However,atlongertimesthePLemissionintensityfallsbelowthevalueoftheinitial suspensionforallspecies.Forevenlongertimestheemissionintensityiscompletely quenchedforallspecies.Incontrast,whenthesameinitialsuspensionowsthrough 147

PAGE 148

FigureA-3.E22regionofabsorbancespectrafortwoshearedsuspensions.4 10 4 s )]TJ/F25 7.9701 Tf 6.587 0 Td [(1 thatinitiallyshowedgoodandpoorPLintensities.Thenon-resonant backgroundhasbeensubtracted. themicrochannelwithanaverageshearrateof0.4 10 4 s )]TJ/F25 7.9701 Tf 6.586 0 Td [(1 ,FigureA-4bshowsthat theintensityimprovesconsiderablymorethanobservedwithsonication.IfthePL emissionintensityincreasesaftershearingwasentirelyduetodisaggregation,similar improvementsshouldalsobeobservedwithsonication.Thefactthatshearingshows substantialintensityincreasesoveradditionalsonicationstronglysuggeststhatother phenomenaareresponsiblefortheobservedincreasesinPLintensityaftershearing. Interestingly,thePLemissionintensityplummetsaftersonicationfor230min,indicating thatexcessivesonicationcouldbedetrimentaltotheSDS-SWNTstructure. A.3QuantumYieldCalculation ThisprocedureissimilartothatreportedbyBlackburnetal.tocalculatequantum yieldsofSWNTsinsuspensions. 82 Briey,SWNTquantumyieldswereobtainedby comparisonwithIR26dyesolutionin1,2-dicloroethaneDCE.IR26dyewaspurchased fromExciton.Thedyehasaknownquantumyieldof0.5%inDCE. 82,220 Generally,the quantumyieldofasubstanceisdenedas, 221 PL = Z 1 0 F PL d PL A 148

PAGE 149

FigureA-4.aEmissionspectraofanSDS-SWNTsuspensionwithpoorPLemission Ex. = 784nmmeasuredatdifferenttimesasthesuspensionis ultrasonicated.bComparisonoftheemissionspectraoftheinitialSWNT suspensiontothesuspensionafter230minofsonicationandafterowing throughthemicrochannel. where F istheuorescenceintensityperabsorbedphotonasafunctionofthe wavelengthoftheemittedphoton. F canbewrittenas, F PL = I PL PL I A Ex A where I PL and I A arethephotoluminescenceandabsorbedintensity,and isaconstant whosevaluedependsonthecongurationandoperatingparametersoftheequipment. UsingtheBeer-LambertLaw, I A ex canbewrittenas, I A Ex = I 0 Ex [1 )]TJ/F24 11.9552 Tf 11.955 0 Td [(10 )]TJ/F43 7.9701 Tf 6.587 0 Td [(A ], A where A istheabsorbanceattheexcitationwavelength.Toaccountforinstrument dependence,theunknownquantumyieldofauorophorecanbeobtainedbyusinga referenceuorophoreofknownquantumyield, PL PL R = n 2 R 1 0 F PL d PL n 2 R R 1 0 F PL d PL A 149

PAGE 150

where n aretherefractiveindexesofthesolventsusedforthereferenceandthe uorophoreofunknownquantumyield.Finally,substitutingequationsA-1-A-3into equationA-4,thequantumyieldofaselectedSWNTtypecanbecalculatedwiththe followingexpression: 221 PL SWNT PL R = n 2 n 2 R 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [(10 )]TJ/F43 7.9701 Tf 6.587 0 Td [(A R Ex 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [(10 )]TJ/F43 7.9701 Tf 6.587 0 Td [(A Ex R 1 0 I PL Ex PL d PL R 1 0 I PL R Ex PL d PL A TocalculatethetermsinequationA,thesestepswerefollowed: 1.GiventhefactthatthemeasuredspectraofourSWNTsuspensionsisthesum oftheindividualcontributionsfromeachofthesemiconductingSWNTpresent inthesuspension,thePLspectraweredeconvolutedtoobtainthecontribution ofeachsemiconductingSWNTtothesuspensionspectra.Thespectrawere deconvolutedusingthesoftwareprovidedbyAppliedNanouorescence.The calculatedemissionspectrumofeach n m typethatappearinthespectrawas thenintegratedonawavelengthscale, R 1 0 I PL Ex PL d PL .Themeasurements wereperformedforconcentrationsthatwereinthelinearregimeforabsorbanceas wellasforPL. 2.TheintegralofthePLintensitywasalsocalculatedforthedye.Intheseexperiments, theexcitationwavelength Ex wasthesameforboththeSWNTsuspensionand referencesolution.Themeasurementswereperformedforconcentrationsthat wereinthelinearregimeforabsorbanceaswellasforPL. 3.FindingAalsorequiresthedeconvolutionoftheabsorbancespectraintoeach n m type.Theabsorbancespectraweredeconvolutedusingatwostepprocedure reportedbyNairetal. 222 Therststepwastosubtractthenon-resonance backgroundintheabsorbancespectra.Ithasbeenshownthatthenon-resonant backgroundisduetocarbonaceousimpuritiesthatdonotcontributetothePL spectra.Itisnecessarytosubtractthenon-resonantbackgroundtocalculate thequantumyieldwiththeabsorbancecontributionfromSWNTsonly. 114 The backgroundwassubtractedbyusinganinversepowerlawt, A bkg = a b A whereaandbareempiricaltparameters.Afterthebackgroundwassubtracted, thepeaksintheabsorbancespectrawerettedbyassumingVoigtlineshapesand thepreviouslyassignedpeakpositionsforSWNT n m types.Afterdeconvolution, thevalueofAisobtainedfromthespecic n m absorbanceproleatthe excitationwavelength. 150

PAGE 151

4.Thevalueof A R isobtainedfromtheabsorptionspectrumofthedyesolutionatthe excitationwavelength. Sinceweonlyhadtwoexcitationsourcesofdifferentwavelengths,wecalculated QYonlyforthosenanotubeswhoseoptimumexcitationwavelengthwasclosetothat ofthelasers.Forexample,the E 22 transitionsforthe,5and,7SWNTtypesare closetothe784nmlaser.TheQYresultsareshowninTableS2. TableA-2.Relativequantumyields%forselected n m speciesinsheared suspensionsthatwereinitiallypoorandinitiallygoodi.e.,suspensionsA andBinFigure2-5. InitialSuspensionsQY 1 %InitialSuspensionsQY 1 %QY 2 /QY 1 n m PoorGoodPoorGoodPoorGood ,30.82 0.26 0.54 0.16 0.88 0.26 0.71 0.09 1.071.32 ,50.10 0.05 0.39 0.2 0.93 0.38 0.71 0.12 9.701.83 ,50.02 0.03 0.50 0.08 1.05 0.12 0.88 0.14 65.521.78 ,70.0008 0.001 0.13 0.12 0.66 0.05 0.58 0.12 871.084.36 151

PAGE 152

APPENDIXB SUPPORTINGINFORMATIONFORCHAPTER3:DECONVOLUTIONOFTHE SPECTRAANDSOLVENTCHARACTERIZATIONSHEET B.1DeconvolutionoftheSpectra FigureB-1showsatypicalsimulatedcurveredcomparedtotheexperimental curveblueaswellasthedeconvolutedpeakscorrespondingtoeach n m type. Thesimulatedcurveshowsagoodttotheexperimentalcurve.Thepeaksforeach n m componentareshownunderthesimulatedcurve.Notethattheattachedsolvent characterizationsheetsshowthedeconvolutionforallPLspectra. FigureB-1.DeconvolutionofPLemissionspectraforexcitationata662andb784 nm.Weightingfactorswereassignedtoeachpeakbasedontheamountof overlapwithotherpeaksandtheintensity.Thosepeakshighlightedina showexamplesofthethreeweightingfactorsassociatedwithoverlapwhile bshowsthoseassociatedwithintensity.Thegreen,blue,andpurple curveswereassignedvaluesof1,0.5,and0.25,respectively. B.2SolventCharacterizationSheets Theresultsofeachsolventaresummarizedinthefollowingsheetsforeasy reference.ThesedatasheetsincludethePLemissionfromexcitationat662and 784nm,theabsorbancespectrumforthepuresolvent,comparisonoftheabsorbance andPLemissionspectra,andatableofthepeakpositionandsolvatochromicshiftof each n m type. 152

PAGE 153

FigureB-2.Hexane =1.89 =1.37 D =0 f 2 =0.369 TableB-1.Hexane. Type E 11 eV E 11 ShiftmeV ,31.30167 ,51.26983 ,51.21163 ,21.17549 ,41.12744 ,61.10456 ,11.06033 ,61.05248 ,31.03533 ,50.99746 ,50.99331 ,10.99781 ,70.93830 ,20.90150 153

PAGE 154

FigureB-3.Heptane =1.92 =1.39 D =0 f 2 =0.383 TableB-2.Heptane. Type E 11 eV E 11 ShiftmeV ,31.30266 ,51.26983 ,51.21163 ,21.17648 ,41.12744 ,61.10555 ,11.06033 ,61.05149 ,31.03533 ,50.99746 ,50.99331 ,10.97880 ,70.93830 ,20.90150 154

PAGE 155

FigureB-4.Cyclohexane =2.02 =1.43 D =0 f 2 =0.411 TableB-3.Cyclohexane. Type E 11 eV E 11 ShiftmeV ,31.30266 ,51.26983 ,51.21262 ,21.17648 ,41.12744 ,61.10555 ,11.06033 ,61.05248 ,31.03533 ,50.99746 ,50.99331 ,10.97880 ,70.93830 ,20.90150 155

PAGE 156

FigureB-5.CarbonTetrachloride =2.23 =1.4 D =0 f 2 =0.369 TableB-4.CarbonTetrachloride. Type E 11 eV E 11 ShiftmeV ,31.29969 ,51.26686 ,51.20965 ,21.17549 ,41.12645 ,61.10357 ,11.06033 ,61.05248 ,31.03434 ,50.99449 ,50.99232 ,10.97682 ,70.93731 ,20.90051 156

PAGE 157

FigureB-6.p-Xylene =2.27 =1.50 D =0 f 2 =0.455 TableB-5.p-Xylene. Type E 11 eV E 11 ShiftmeV ,31.29474 ,51.26389 ,51.20569 ,21.16955 ,41.12546 ,61.10060 ,11.05637 ,61.04951 ,31.02939 ,50.99449 ,50.98935 ,10.97781 ,70.93434 ,20.90051 157

PAGE 158

FigureB-7.Benzene =2.28 =1.50 D =0 f 2 =0.369 TableB-6.Benzene. Type E 11 eV E 11 ShiftmeV ,31.29474 ,51.26587 ,51.20669 ,21.17054 ,41.12447 ,61.10159 ,11.05736 ,61.05050 ,31.03137 ,50.99548 ,50.99034 ,10.97682 ,70.93533 ,20.89952 158

PAGE 159

FigureB-8.Toluene =2.39 =1.50 D =0.38 f 2 =0.455 TableB-7.Toluene. Type E 11 eV E 11 ShiftmeV ,31.29474 ,51.26488 ,51.20668 ,21.17054 ,41.12348 ,61.10159 ,11.05637 ,61.05050 ,31.03038 ,50.99449 ,50.98935 ,10.97682 ,70.93533 ,20.90051 159

PAGE 160

FigureB-9.2,6-Dichlorotoluene =3.36 =1.55 D =0 f 2 =0.483 TableB-8.2,6-Dichlorotoluene. Type E 11 eV E 11 ShiftmeV ,31.28781 ,51.25993 ,51.19975 ,21.16361 ,41.14328 ,61.09565 ,11.05142 ,61.04555 ,31.02444 ,51.00736 ,50.98341 ,10.97781 ,70.92939 ,20.89853 160

PAGE 161

APPENDIXC SUPPORTINGINFORMATIONFORCHAPTER5:SETUP,DATAANALYSISAND ADDITIONALDATA C.1ExperimentalSetup ThecolumnshowninFigureC-1ispackeduptothespeciedheightorvolume. Aperistalticpumpensuresacontinuousowofthesampleandeluents.Amixeris usedbeforethepumptohomogenizetheeluentwheneveraconcentrationgradientis used.EluentAwasusuallya1wt%SDSsolution,whileeluentBwasusually2wt%SC solution.Aowadapterisusedtoprovideacontinuousowthroughthecolumnand toavoidbedexpansionduringorafteranexperiment.Thevisibleabsorbanceorthe NIRuorescencespectraoftheefuentaremeasuredcontinuously.Theefuentisthen collectedusingafractioncollector. FigureC-1.Schematicofthesetupusedforthecolumnexperiments. 161

PAGE 162

C.2DataAnalysis Absorbancespectrawerecollectedevery20sduringtheexperiments.The elutioncurvesweregivenintermsoftheabsorbanceat763nmateachtimeA. Theabsorbancevaluesateachtimewerenormalizedbytheabsorbanceoftheinitial sampleA I .Thecumulativemasscurveswereshowntogiveabetterdescriptionofthe elutionpattern.Thecurveswerecalculatedasfollows: 1.Assumetheuidelementsevery20sareuniforminconcentration. 2.Calculatetheamountofuidthatowsin20s. 3.CalculatetheconcentrationofSWNTsineveryfractionusingtheBeer-Lambert lawwheretheextinctioncoefcientwaschosentobe0.043at763nmandthe pathlengthwas1cm. 223 4.Calculatethemassateachfractionm,every20sandaddthemtogetthetotal massrecoveredm T,R Thecalculatedm T,R wasnotalwaysequaltothecalculatedinputmass.Hence,the cumulativemasscurveswereconstructedbynormalizingeachfractionbythetotalmass recovered,m/m T,R .Itwaspossibletodistinguishwhenasamplehadevenmarginal irreversibleretentionduetothecontrastbetweentheSWNTsandtheagarosebeads white.Hence,thecumulativemasscurveswereusedtocorroboratetrendsinthe retentionbehaviorofsuspensionsandnotnecessarilytogiveanaccuratevalueforthe amountofretainedSWNTs. C.3SWNTSuspensionsatDifferentAggregationStates TheamountofbundledSWNTsinthesuspensionwasalteredbycontrollingthe ultracentrifugationtime.AsshowninFigureC-2,theabsorbancespectrachangewith ultracentrifugationtime.Thespectrabecomesharperasmorebundlednanotubes areremovedfromthesuspensionatlongerultracentrifugationtimes.Theelution characteristicsandcorrespondingabsorbancespectraforthesesuspensionsareshown inFiguresC-3. 162

PAGE 163

FigureC-2.Normalizedvis-NIRabsorbanceof1wt%SDS-SWNTsuspensions ultracentrifugedat20,000rpmfortimesof240,180,90,30,and0min. Beforemeasuringthespectra,thesuspensionspreparedat180,90,30,and 0minweredilutedtomatchtheabsorbanceofthe240minsuspensionat 763nm. FigureC-3.Retentionbehaviorof1wt%SDS-SWNTsuspensionspreparedatdifferent aggregationstates.Elutionandcumulativemasscurvesforsuspensions subjectedtoa240,b90,c30,andd0min.Thecorresponding absorbancespectraareshowninehfortheinitialsuspensionandthe efuentattherstP1andsecondP2peaksoftheelutioncurve,which areshownasdashedlinesintheelutioncurves. 163

PAGE 164

FigureC-4.PLemissionspectrafrom1wt%SDS-SWNTsuspensionstitratedwith differentamountsofGA. C.4GumArabic-ModiedSWNTSuspension GumarabicGAwasaddedtotheinitial1wt%SDS-SWNTsuspensionatdifferent concentrations.ThePLspectrafortheseSWNTsuspensionsareshowninFigure C-4.AtlowconcentrationsofGA,thePLemissionintensityincreases.Theseintensity changesareattributedtobetterprotectionoftheSWNTsfromthequenchingeffectsof theaqueoussolvent.AthigherGAconcentrations,thesuspensionstartstoaggregate, resultinginlowerPLintensity.Theelutioncharacteristicsandcorrespondingabsorbance spectraforthesesuspensionsareshowninFiguresC-5. C.5Salt-ModiedSWNTSuspension Thesurfactantstructureoftheinitial1wt%SDS-SWNTsuspensionwasmodied byaddingNaClsolutionsatconcentrationsbetween40and110mM.Previously, Doornandco-workersshowedthatSDS-suspendedSWNTsexperienceablueshift intheuorescencespectrawhensaltisadded. 28 FigureC-6ashowsthespectral changesobservedwhendifferentamountsofNaClareaddedtoSWNTssuspended by1wt%SDS.Whencomparedtotheinitialsuspension,itisclearthatthepeaks 164

PAGE 165

haveblue-shiftedandthattheintensityhasincreasedforsome n m types.Asthe concentrationofNaClincreases,theemissionenergyred-shiftsandPLdecreases, whichisduetotheaggregationofSWNTs.Thisaggregationiscausedbythetotal screeningofthechargesontheSWNTs,whichoccursatdifferentconcentrations foreach n m type.Aspreviouslyreported,largerSWNTsaremoresensitivetothe increaseinNaClconcentration.ThePLintensityoflargerdiameterSWNTsdropsfaster asNaClconcentrationincreases. 68 Theseintensityandemissionenergydifferencesindicatethattheimmediate environmentsurroundingSWNTshaschanged.Asthesesuspensionsarecentrifuged inadensitygradient,itisobservedthattheirbuoyantdensityhasdecreased,as demonstratedpreviously 28 andconrmedinpartsbandcofFigureC-6.The FigureC-5.RetentionbehaviorofSDS-SWNTsuspensionsmodiedwith0.1wt%of GA.Elutionandcumulativemasscurvesfortheainitialandb GA-modiedSWNTsuspensions.Thecorrespondingabsorbancespectra areshownincanddfortheinitialsuspensionandtheefuentattherst P1andsecondP2peaksoftheelutioncurve,whichareshownas dashedlinesintheelutioncurves. 165

PAGE 166

FigureC-6.AlteringthesurfactantstructurearoundSWNTswithsalt.aPLspectrafor NaCltitratedsuspensions.Imagesofthecentrifugationtubesfortheb initialandcsalt-modiedmMSWNTsuspensionsinadensity gradientmedium. changesaremoredramaticform-SWNTs,whichformadistinctredbandinthe centrifugationtube,indicatingthatm-SWNTshavelargerdensitychangesdecreases thans-SWNTs.AsdescribedbyDoornandco-workers,theseresultsareexpected sincethesaltscreensthechargesontheSDSheadgroup,allowingbetterpacking ofthemoleculesontothesurface.Therefore,thedifferenceinPLemissionenergy andintensityalongwiththedecreaseinbuoyantdensityindicateschangestothe conformationandaggregationnumberofSDSmoleculesontheSWNTsurface.The surfactantstructureoftheinitial1wt%SDS-SWNTsuspensionwasmodiedbyadding NaClsolutionsatconcentrationsbetween40and110mM.Theelutioncharacteristics andcorrespondingabsorbancespectraforthesesuspensionsareshowninFigure C-7.Toruleoutchargescreeningeffectsontheagarosebeads,theinitialSDS-SWNT 166

PAGE 167

FigureC-7.Retentionbehaviorofsalt-modiedSWNTsuspensions.Elutionand cumulativemasscurvesfor1wt%SDS-SWNTsmodiedwitha40,b60, c80,andd110mMNaClsolution.Thecorrespondingabsorbance spectraareshowninehfortheinitialsuspensionandtheefuentat therstP1andsecondP2peaksoftheelutioncurve,whichareshown asdashedlinesintheelutioncurves. suspensionwasloadedintothecolumnat75%ofCVtosaturatetheagarosebeads. Undertheseconditions,thecolumnwasclearlyoverloaded.Theremainings-SWNTs retainedwithinthecolumnweresubjectedtoasaltgradientof0to80mMNaClin1 wt%SDSover6CVapprox.62min,asshowninFigureC-8.Onlyasmallfractionof s-SWNTselutedunderthesaltgradient,sotheeluentwaschangedto2wt%SC. C.6Solvent-ModiedSWNTSuspension Asdescribedpreviouslyinchapter3,thehydrophobicregionbetweenthesurfactant andtheSWNTsurfacecanbeswelledwithimmiscibleorganicsolvents.Theanticipated changestothesurfactantstructureareshowninFigureC-9.Thesurfactantstructureof theinitial1wt%SDS-SWNTsuspensionwasmodiedbyaddingtheimmisciblesolvents o-dichlorobenzeneODCBandcarbontetrachloride.Theelutioncharacteristicsand correspondingabsorbancespectraforthesesuspensionsareshowninFiguresC-10 andC-11. 167

PAGE 168

FigureC-8.RetentionbehaviorofSDS-SWNTselutedwith1wt%SDS-NaClsolution gradient.aElutionandcumulativemasscurves.Theeluentprocedureis shownatthetopofthegraph.bAbsorbancespectrafromtheinitial sampleandtheefuentattherstP1andsecondP2peaksoftheelution curve,whichareshownasdashedlinesinparta.Thecolumninternal diameter,heightandvolumewere1.5cm,6cm,and10.5cm 3 ,respectively. C.7SWNTsSuspendedwithDeoxyribonucleicAcidDNA SWNTsweresuspendedwithDNAtoprobetheimportanceofthenanotube surfaceonretentioncharacteristics.AsshowninFigureC-12,theabsorbancespectra forDNA-SWNTsissimilartoSDS-SWNTs.However,thePLemissionintensityof DNA-coatedSWNTsissignicantlylowerthantheintensityofSDS-coatedSWNTs. ThesedifferencesareassociatedwiththeexposedsidewallofDNA-coatedSWNTsthat enablestheexcitontobequenchedbytheaqueousphase. 168

PAGE 169

FigureC-9.Schematicoftheprocessbywhichthesolventswelledsamplesare prepared.Theainitialsuspensionismixedwithanimmiscibleorganic solventtocreateblocalsolventenvironmentsaroundthenanotubes.c Theresidualsolventcanthenberemovedbyevaporationover24h.The cartoonsshowexamplesofhowthesurfactantsurroundingSWNTsmight changeduringthepresenceandsubsequentremovaloftheimmiscible organicsolvent. FigureC-10.RetentionbehaviorofODCB-modiedSWNTsuspensions.Elutionand cumulativemasscurvesfortheainitial,bODCB-swelled,andc ODCB-evaporatedSWNTsuspensions.Thecorrespondingabsorbance spectraareshownindffortheinitialsuspensionandtheefuentat therstP1andsecondP2peaksoftheelutioncurve,whichareshown asdashedlinesintheelutioncurves. 169

PAGE 170

FigureC-11.RetentionbehaviorofCCl 4 -modiedSWNTsuspensions.Elutionand cumulativemasscurvesfortheainitial,bCCl 4 -swelled,andc CCl 4 -evaporatedSWNTsuspensions.Thecorrespondingabsorbance spectraareshownindffortheinitialsuspensionandtheefuentat therstP1andsecondP2peaksoftheelutioncurve,whichareshown asdashedlinesintheelutioncurves. FigureC-12.Comparisonoftheabsorbanceaandphotoluminescencebspectraof SWNTsuspendedwithSDSorDNA.Theabsorbanceofeachsuspension wasmatchedat763nm.NoticethattheDNA-SWNTshavepeaksthatare substantiallyred-shiftedintheabsorbanceandPLspectraaswellas quenchedPLemission. 170

PAGE 171

REFERENCES 1.Saito,R.,G.,D.&S.,D.M. Physicalpropertiesofcarbonnanotubes Imperial CollegePress,London,1998. 2.Baughman,R.,Zakhidov,A.&deHeer,W.Carbonnanotubes-theroutetoward applications. Science 297 ,787. 3.Kim,P.,Shi,L.,Majumdar,A.&McEuen,P.Thermaltransportmeasurementsof individualmultiwallednanotubes. Phys.Rev.Lett. 87 ,215502. 4.Pop,E.,Mann,D.,Wang,Q.,Goodson,K.&Dai,H.Thermalconductanceofan individualsingle-wallcarbonnanotubeaboveroomtemperature. NanoLett. 6 96. 5.Bachtold,A. etal. Scannedprobemicroscopyofelectronictransportincarbon nanotubes. Phys.Rev.Lett. 84 ,6082. 6.Avouris,P.,Chen,Z.&Perebeinos,V.Carbon-basedelectronics. Nat.Nanotechnol. 2 ,605. 7.Durkop,T.,Getty,S.,Cobas,E.&Fuhrer,M.Extraordinarymobilityin semiconductingcarbonnanotubes. NanoLett. 4 ,35. 8.Franklin,A.D.&Chen,Z.Lengthscalingofcarbonnanotubetransistors. Nat. Nanotechnol. 5 ,858. 9.Tans,S.,Verschueren,A.&Dekker,C.Room-temperaturetransistorbasedona singlecarbonnanotube. Nature 393 ,49. 10.Rutherglen,C.&Burke,P.Carbonnanotuberadio. NanoLett. 7 ,3296 11.Mueller,T. etal. Efcientnarrow-bandlightemissionfromasinglecarbon nanotubep-ndiode. Nat.Nanotechnol. 5 ,27. 12.Cao,Q. etal. Medium-scalecarbonnanotubethin-lmintegratedcircuitson exibleplasticsubstrates. Nature 454 ,495. 13.Engel,M. etal. Thinlmnanotubetransistorsbasedonself-assembled,aligned, semiconductingcarbonnanotubearrays. ACSNano 2 ,2445. 14.Wang,C.,Zhang,J.&Zhou,C.Macroelectronicintegratedcircuitsusing high-performanceseparatedcarbonnanotubethin-lmtransistors. ACSNano 4 ,7123. 15.Sgobba,V.&Guldi,D.M.Carbonnanotubeselectronic/electrochemical propertiesandapplicationfornanoelectronicsandphotonics. Chem.Soc.Rev. 38 165. 171

PAGE 172

16.Holt,J.M. etal. ProlongingchargeseparationinP3HT-swntcompositesusing highlyenrichedsemiconductingnanotubes. NanoLett. 10 ,4627. 17.Ferguson,A.J. etal. Photoinducedenergyandchargetransferinp3ht:swnt composites. J.Phys.Chem.Lett. 1 ,2406. 18.Ham,M.-H. etal. Evidenceforhigh-efciencyexcitondissociationat polymer/single-walledcarbonnanotubeinterfacesinplanarnano-heterojunction photovoltaics. AcsNano 4 ,6251. 19.Li,X. etal. Selectivesynthesiscombinedwithchemicalseparationofsingle-walled carbonnanotubesforchiralityselection. J.Am.Chem.Soc. 129 ,15770 20.Strano,M. etal. Electronicstructurecontrolofsingle-walledcarbonnanotube functionalization. Science 301 ,1519. 21.Usrey,M.,Lippmann,E.&Strano,M.Evidenceforatwo-stepmechanismin electronicallyselectivesingle-walledcarbonnanotubereactions. J.Am.Chem. Soc. 127 ,16129. 22.Krupke,R.,Hennrich,F.,vonLohneysen,H.&Kappes,M.Separationofmetallic fromsemiconductingsingle-walledcarbonnanotubes. Science 301 ,344 23.Hong,S.,Jung,S.,Choi,J.,Kim,Y.&Baik,S.Electricaltransportcharacteristics ofsurface-conductance-controlled,dielectrophoreticallyseparatedsingle-walled carbonnanotubes. Langmuir 23 ,4749. 24.Zheng,M.Structure-basedcarbonnanotubesortingbysequence-dependentDNA assembly. Science 302 ,1545. 25.Lustig,S.R.,Jagota,A.,aConstantineKhripin&Zheng,M.Theoryof structure-basedcarbonnanotubeseparationsbyion-exchangechromatographyof DNA/CNThybrids. J.Phys.Chem.B 109 ,2559. 26.Tu,X.,Manohar,S.,Jagota,A.&Zheng,M.Dnasequencemotifsfor structure-specicrecognitionandseparationofcarbonnanotubes. Nature 460 ,250. 27.Arnold,M.S.,Green,A.A.,Hulvat,J.F.,Stupp,S.I.&Hersam,M.C.Sorting carbonnanotubesbyelectronicstructureusingdensitydifferentiation. Nat. Nanotechnol. 1 ,60. 28.Niyogi,S.,Densmore,C.G.&Doorn,S.K.Electrolytetuningofsurfactant interfacialbehaviorforenhanceddensity-basedseparationsofsingle-walled carbonnanotubes. J.Am.Chem.Soc. 131 ,1144. 172

PAGE 173

29.Duque,J.G.,Densmore,C.G.&Doorn,S.K.Saturationofsurfactantstructureat thesingle-walledcarbonnanotubesurface. J.Am.Chem.Soc. 132 ,16165 30.Carvalho,E.J.F.&dosSantos,M.C.Roleofsurfactantsincarbonnanotubes densitygradientseparation. ACSNano 4 ,765. 31.Ghosh,S.,Bachilo,S.M.&Weisman,R.B.Advancedsortingofsingle-walled carbonnanotubesbynonlineardensity-gradientultracentrifugation. Nat.Nanotechnol. 5 ,443. 32.Fagan,J.A. etal. Centrifugallengthseparationofcarbonnanotubes. Langmuir 24 ,13880. 33.Tanaka,T. etal. Simpleandscalablegel-basedseparationofmetallicand semiconductingcarbonnanotubes. NanoLett. 9 ,1497. 34.Tsyboulski,D.A.,Rocha,J.-D.R.,Bachilo,S.M.,Cognet,L.&Weisman,R.B. Structure-dependentuorescenceefcienciesofindividualsingle-walledcarbon nanotubes. NanoLett. 7 ,3080. 35.Ambrose,W. etal. Singlemoleculeuorescencespectroscopyatambient temperature. Chem.Rev. 99 ,2929. 36.Hartschuh,A.,Pedrosa,H.,Novotny,L.&Krauss,T.Simultaneousuorescence andramanscatteringfromsinglecarbonnanotubes. Science 301 ,1354 37.Barone,P.W.,Baik,S.,Heller,D.A.&Strano,M.S.Near-infraredopticalsensors basedonsingle-walledcarbonnanotubes. Nat.Mater. 4 ,86. 38.Heller,D.A. etal. Opticaldetectionofdnaconformationalpolymorphismon single-walledcarbonnanotubes. Science 311 ,508. 39.Jin,H.,Heller,D.A.,Kim,J.-H.&Strano,M.S.Stochasticanalysisofstepwise uorescencequenchingreactionsonsingle-walledcarbonnanotubes:Single moleculesensors. NanoLett. 8 ,4299. 40.Jin,H. etal. Detectionofsingle-moleculeh2o2signallingfromepidermalgrowth factorreceptorusinguorescentsingle-walledcarbonnanotubes. Nat.Nanotechnol. 5 ,302. 41.Heller,D.A. etal. Multimodalopticalsensingandanalytespecicityusing single-walledcarbonnanotubes. Nat.Nanotechnol. 4 ,114. 42.Xu,Z.&Buehler,M.J.Nanoengineeringheattransferperformanceatcarbon nanotubeinterfaces. ACSNano 3 ,2767. 173

PAGE 174

43.Bergin,S.D. etal. Towardssolutionsofsingle-walledcarbonnanotubesin commonsolvents. Adv.Mater. 20 ,1876. 44.Bergin,S.D.,Sun,Z.,Streich,P.,Hamilton,J.&Coleman,J.N.Newsolventsfor nanotubes:Approachingthedispersibilityofsurfactants. J.Phys.Chem.C 114 231. 45.Strano,M.S. etal. Electronicstructurecontrolofsingle-walledcarbonnanotube functionalization. Science 301 ,1519. 46.Wang,R.K. etal. Improvingtheeffectivenessofinterfacialtrappinginremoving single-walledcarbonnanotubebundles. J.Am.Chem.Soc. 130 ,14721 47.Wang,R.K.,Reeves,R.D.&Ziegler,K.J.Interfacialtrappingofsingle-walled carbonnanotubebundles. J.Am.Chem.Soc. 129 ,15124. 48.Heller,D.A. etal. Concomitantlengthanddiameterseparationofsingle-walled carbonnanotubes. J.Am.Chem.Soc. 126 ,14567. 49.Duque,J.G. etal. Diameter-dependentsolubilityofsingle-walledcarbon nanotubes. ACSNano 50.Blanch,A.J.,Lenehan,C.E.&Quinton,J.S.Optimizingsurfactantconcentrations fordispersionofsingle-walledcarbonnanotubesinaqueoussolution. J.Phys. Chem.B 114 ,9805. 51.Wang,R.K.,Chen,W.-C.,Campos,D.K.&Ziegler,K.J.Swellingthemicelle coresurroundingsingle-walledcarbonnanotubeswithwater-immiscibleorganic solvents. J.Am.Chem.Soc. 130 ,16330. 52.Duque,J.G. etal. Stableluminescencefromindividualcarbonnanotubesin acidic,basic,andbiologicalenvironments. J.Am.Chem.Soc. 130 ,2626 53.Tummala,N.R.&Striolo,A.SDSsurfactantsoncarbonnanotubes.Aggregate morphology. ACSNano 3 ,595. 54.Islam,M.F.,Rojas,E.,Bergey,D.M.,Johnson,A.T.&Yodh,A.G.Highweight fractionsurfactantsolubilizationofsingle-wallcarbonnanotubesinwater. Nano Lett. 3 ,269. 55.O'Connell,M.J. etal. Bandgapuorescencefromindividualsingle-walledcarbon nanotubes. Science 297 ,593. 56.Yurekli,K.,Mitchell,C.A.&Krishnamoorti,R.Small-angleneutronscatteringfrom surfactant-assistedaqueousdispersionsofcarbonnanotubes. J.Am.Chem.Soc. 126 ,9902. 174

PAGE 175

57.Xu,Z.,Yang,X.&Yang,Z.Amolecularsimulationprobingofstructureand interactionforsupramolecularsodiumdodecylsulfate/single-wallcarbonnanotube assemblies. NanoLett. 10 ,985. 58.Weisman,R.B.&Bachilo,S.M.Dependenceofopticaltransitionenergieson structureforsingle-walledcarbonnanotubesinaqueoussuspension: AL'an empiricalkatauraplot. NanoLett. 3 ,1235. 59.Baughman,R.H.,Zakhidov,A.A.&de,W.A.,Heer.Carbonnanotubes-theroute towardapplications. Science 297 ,787. 60.Hu,H. etal. Inuenceofthezetapotentialonthedispersabilityandpuricationof single-walledcarbonnanotubes. J.Phys.Chem.B 109 ,11520. 61.Bandyopadhyaya,R.,Nativ-Roth,E.,Regev,O.&Yerushalmi-Rozen,R. Stabilizationofindividualcarbonnanotubesinaqueoussolutions. NanoLett. 2 ,25. 62.Haggenmueller,R. etal. Comparisonofthequalityofaqueousdispersionsof singlewallcarbonnanotubesusingsurfactantsandbiomolecules. Langmuir 24 5070. 63.Moore,V.C. etal. Individuallysuspendedsingle-walledcarbonnanotubesin varioussurfactants. NanoLett. 3 ,1379. 64.Bachilo,S.M. etal. Structure-assignedopticalspectraofsingle-walledcarbon nanotubes. Science 298 ,2361. 65.Weisman,R.B.&Bachilo,S.M.Dependenceofopticaltransitionenergieson structureforsingle-walledcarbonnanotubesinaqueoussuspension:Anempirical katauraplot. NanoLett. 3 ,1235. 66.Chen,F.,Wang,B.,Chen,Y.&Li,L.-J.Towardtheextractionofsinglespecies ofsingle-walledcarbonnanotubesusinguorene-basedpolymers. NanoLett. 7 3013. 67.Hwang,J.-Y. etal. Polymerstructureandsolventeffectsontheselectivedispersion ofsingle-walledcarbonnanotubes. J.Am.Chem.Soc. 130 ,3543. 68.Niyogi,S. etal. Selectiveaggregationofsingle-walledcarbonnanotubesviasalt addition. J.Am.Chem.Soc. 129 ,1898. 69.Dukovic,G. etal. Reversiblesurfaceoxidationandefcientluminescence quenchinginsemiconductorsingle-wallcarbonnanotubes. J.Am.Chem.Soc. 126 ,15269. 70.Okazaki,T. etal. Photoluminescencemappingof"as-grown"single-walledcarbon nanotubes:Acomparisonwithmicelle-encapsulatednanotubesolutions. Nano Lett. 5 ,2618. 175

PAGE 176

71.Cognet,L. etal. Stepwisequenchingofexcitonuorescenceincarbonnanotubes bysingle-moleculereactions. Science 316 ,1465. 72.Crochet,J.,Clemens,M.&Hertel,T.Quantumyieldheterogeneitiesofaqueous single-wallcarbonnanotubesuspensions. J.Am.Chem.Soc. 129 ,8058 73.Qian,H. etal. Visualizingthelocalopticalresponseofsemiconductingcarbon nanotubestodna-wrapping. NanoLett. 8 ,2706. 74.Tan,P. etal. Photoluminescencespectroscopyofcarbonnanotubebundles: evidenceforexcitonenergytransfer. Phys.Rev.Lett. 99 ,137402. 75.Torrens,O.N.,Milkie,D.E.,Zheng,M.&Kikkawa,J.M.Photoluminescencefrom intertubecarriermigrationinsingle-walledcarbonnanotubebundles. NanoLett. 6 2864. 76.Choi,J.H.&Strano,M.S.Solvatochromisminsingle-walledcarbonnanotubes. Appl.Phys.Lett. 90 ,223114. 77.Strano,M.S. etal. Reversible,band-gap-selectiveprotonationofsingle-walled carbonnanotubesinsolution. J.Phys.Chem.B 107 ,6979. 78.Doyle,C.D.,Rocha,J.-D.R.,Weisman,R.B.&Tour,J.M.Structure-dependent reactivityofsemiconductingsingle-walledcarbonnanotubeswith benzenediazoniumsalts. J.Am.Chem.Soc. 130 ,6795. 79.Nish,A.&Nicholas,R.J.Temperatureinducedrestorationofuorescencefrom oxidizedsingle-walledcarbonnanotubesinaqueoussodiumdodecylsulfate solution. Phys.Chem.Chem.Phys. 8 ,3547. 80.Gokus,T. etal. Excitondecaydynamicsinindividualcarbonnanotubesatroom temperature. Appl.Phys.Lett. 92 ,153119. 81.Lefebvre,J.,Austing,D.G.,Bond,J.&Finnie,P.Photoluminescenceimagingof suspendedsingle-walledcarbonnanotubes. NanoLett. 6 ,1603. 82.Blackburn,J.L. etal. Protonationeffectsonthebranchingratioinphotoexcited single-walledcarbonnanotubedispersions. NanoLett. 8 ,1047. 83.Naumov,A.V.,Bachilo,S.M.,Tsyboulski,D.A.&Weisman,R.B.Electriceld quenchingofcarbonnanotubephotoluminescence. NanoLett. 8 ,1527 84.Fagan,J.A. etal. Length-dependentopticaleffectsinsingle-wallcarbon nanotubes. J.Am.Chem.Soc. 129 ,10607. 85.Jones,M. etal. Analysisofphotoluminescencefromsolubilizedsingle-walled carbonnanotubes. Phys.Rev.B 71 ,115426. 176

PAGE 177

86.Heller,D.A.,Barone,P.W.&Strano,M.S.Sonication-inducedchangesin chiraldistribution:Acomplicationintheuseofsingle-walledcarbonnanotube uorescencefordeterminingspeciesdistribution. Carbon 43 ,651. 87.Wallace,E.J.&Sansom,M.S.P.Carbonnanotube/detergentinteractionsvia coarse-grainedmoleculardynamics. NanoLett. 7 ,1923. 88.Mattsson,M. etal. Dielectrophoresis-inducedseparationofmetallicand semiconductingsingle-wallcarbonnanotubesinacontinuousowmicrouidic system. J.Nanosci.Nanotechnol. 7 ,3431. 89.Peng,H.,Alvarez,N.T.,Kittrell,C.,Hauge,R.H.&Schmidt,H.K. Dielectrophoresiseldowfractionationofsingle-walledcarbonnanotubes. J. Am.Chem.Soc. 128 ,8396. 90.Shin,D.H. etal. Continuousextractionofhighlypuremetallicsingle-walledcarbon nanotubesinamicrouidicchannel. NanoLett. 8 ,4380. 91.Casey,J.P.,Bachilo,S.M.,Moran,C.H.&Weisman,R.B.Chirality-resolved lengthanalysisofsingle-walledcarbonnanotubesamplesthroughshear-aligned photoluminescenceanisotropy. ACSNano 2 ,1738. 92.Xia,Y.&Whitesides,G.M.Softlithography. Annu.Rev.Mater.Sci. 28 ,153 93.Lefebvre,J.,Fraser,J.,Finnie,P.&Homma,Y.Photoluminescencefroman individualsingle-walledcarbonnanotube. Phys.Rev.B 69 ,075403. 94.Kim,Y.,Minami,N.&Kazaoui,S.Highlypolarizedabsorptionand photoluminescenceofstretch-alignedsingle-wallcarbonnanotubesdispersed ingelatinlms. Appl.Phys.Lett. 86 ,073103. 95.Fagan,J. etal. Dielectricresponseofalignedsemiconductingsingle-wall nanotubes. Phys.Rev.Lett. 98 ,147402. 96.Fry,D. etal. Anisotropyofshearedcarbon-nanotubesuspensions. Phys.Rev.Lett. 95 ,038304. 97.Doi,M.&Edwards,S.F. Thetheoryofpolymerdynamics OxfordUniversityPress, Oxford,1986. 98.Duggal,R.&Pasquali,M.Dynamicsofindividualsingle-walledcarbonnanotubes inwaterbyreal-timevisualization. Phys.Rev.Lett. 96 ,246104. 99.Tsyboulski,D.A.,Bachilo,S.M.,Kolomeisky,A.B.&Weisman,R.B.Translational androtationaldynamicsofindividualsingle-walledcarbonnanotubesinaqueous suspension. ACSNano 2 ,1770. 177

PAGE 178

100.Arnold,M.S.,Suntivich,J.,Stupp,S.I.&Hersam,M.C.Hydrodynamic characterizationofsurfactantencapsulatedcarbonnanotubesusingananalytical ultracentrifuge. ACSNano 2 ,2291. 101.Richard,C.,Balavoine,F.,Schultz,P.,Ebbesen,T.W.&Mioskowski,C. Supramolecularself-assemblyoflipidderivativesoncarbonnanotubes. Science 300 ,775. 102.Heller,D.A.,Barone,P.W.,Swanson,J.P.,Mayrhofer,R.M.&Strano,M.S.Using ramanspectroscopytoelucidatetheaggregationstateofsingle-walledcarbon nanotubes. J.Phys.Chem.B 108 ,6905. 103.Strano,M.S. etal. Theroleofsurfactantadsorptionduringultrasonicationinthe dispersionofsingle-walledcarbonnanotubes. J.Nanosci.Nanotechnol. 3 ,81 104.Kumatani,A.&Warburton,P.A.Characterizationofthedisaggregationstate ofsingle-walledcarbonnanotubebundlesbydielectrophoresisandRaman spectroscopy. Appl.Phys.Lett. 92 ,243123. 105.Krupke,R.,Hennrich,F.,Kappes,M.M.&von,H.,Loehneysen.Surface conductanceinduceddielectrophoresisofsemiconductingsingle-walledcarbon nanotubes. NanoLett. 4 ,1395. 106.Park,J.&Butler,J.Inhomogeneousdistributionofarigidbreundergoing rectilinearowbetweenparallelwallsathighpcletnumbers. J.FluidMech. 630 267. 107.AlKahwaji,A.&Kellay,H.Observationsofthecollapseofdilutelyotropiclamellar phasesundershearow. Phys.Rev.Lett. 84 ,3073. 108.Hoffmann,H.,Hofmann,S.,Rauscher,A.&Kalus,J. TrendsinColloidand InterfaceScience ,vol.V,chap.Shear-inducedtransitionsinmicellarsolutions,24 Springer-Verlag,1991. 109.Berret,J.-F.,Gamez-Corrales,R.,Serero,Y.,Molino,F.&Lindner,P. Shear-inducedmicellargrowthindilutesurfactantsolutions. Europhys.Lett. 54 ,605. 110.McDonald,T.J.,Blackburn,J.L.,Metzger,W.K.,Rumbles,G.&Heben,M.J. Chiral-selectiveprotectionofsingle-walledcarbonnanotubephotoluminescenceby surfactantselection. J.Phys.Chem.C 111 ,17894. 111.McDonald,T.J.,Engtrakul,C.,Jones,M.,Rumbles,G.&Heben,M.J. Kineticsofplquenchingduringsingle-walledcarbonnanotuberebundlingand diameter-dependentsurfactantinteractions. J.Phys.Chem.B 110 ,25339 178

PAGE 179

112.Shvartzman-Cohen,R. etal. Selectivedispersionofsingle-walledcarbon nanotubesinthepresenceofpolymers:theroleofmolecularandcolloidallength scales. J.Am.Chem.Soc. 126 ,14850. 113.Carlson,L.J.,Maccagnano,S.E.,Zheng,M.,Silcox,J.&Krauss,T.D. Fluorescenceefciencyofindividualcarbonnanotubes. NanoLett. 7 ,3698 114.Nish,A.,Hwang,J.-Y.,Doig,J.&Nicholas,R.J.Highlyselectivedispersionof single-walledcarbonnanotubesusingaromaticpolymers. Nat.Nanotechnol. 2 640. 115.Araujo,P.T.&Jorio,A.Theroleofenvironmentaleffectsontheopticaltransition energiesandradialbreathingmodefrequencyofsinglewallcarbonnanotubes. Phys.StatusSolidiB 245 ,2201. 116.Araujo,P.T. etal. Natureoftheconstantfactorintherelationbetweenradial breathingmodefrequencyandtubediameterforsingle-wallcarbonnanotubes. PhysicalReviewB 77 ,241403. 117.Jorio,A.,Maciel,I.O.,Araujo,P.T.,Pesce,P.B.C.&Pimenta,M.A.The fundamentalaspectsofcarbonnanotubemetrology. Phys.StatusSolidiB 244 ,4011. 118.Kiowski,O. etal. Photoluminescencemicroscopyofcarbonnanotubesgrownby chemicalvapordeposition:Inuenceofexternaldielectricscreeningonoptical transitionenergies. Phys.Rev.B 75 ,075421. 119.Longhurst,M.&Quirke,N.Theenvironmentaleffectontheradialbreathingmode ofcarbonnanotubesinwater. J.Chem.Phys. 124 ,234708. 120.Miyauchi,Y. etal. Dependenceofexcitontransitionenergyofsingle-walledcarbon nanotubesonsurroundingdielectricmaterials. Chem.Phys.Lett. 442 ,394 121.Ohno,Y. etal. Excitonictransitionenergiesinsingle-walledcarbonnanotubes: dependenceonenvironmentaldielectricconstant. Phys.StatusSolidiB 244 4002. 122.Walsh,A.G. etal. Screeningofexcitonsinsingle,suspendedcarbonnanotubes. NanoLett. 7 ,1485. 123.Duque,J.G.,Pasquali,M.,Cognet,L.&Lounis,B.Environmentaland synthesis-dependentluminescencepropertiesofindividualsingle-walledcarbon nanotubes. ACSNano 3 ,2153. 124.Hertel,T. etal. Spectroscopyofsingle-anddouble-wallcarbonnanotubesin differentenvironments. NanoLett. 5 ,511. 179

PAGE 180

125.Bahr,J.L.,Mickelson,E.T.,Bronikowski,M.J.,Smalley,R.E.&Tour,J.M. Dissolutionofsmalldiametersingle-wallcarbonnanotubesinorganicsolvents. Chem.Commun. 193. 126.Li,L.J.,Nicholas,R.J.,Deacon,R.S.&Shields,P.A.Chiralityassignmentof single-walledcarbonnanotubeswithstrain. Phys.Rev.Lett. 93 ,156104. 127.Benedict,L.X.,Louie,S.G.&Cohen,M.L.Staticpolarizabilitiesofsingle-wall carbonmolecules. Phys.Rev.B 52 ,8541. 128.Brothers,E.N.,Scuseria,G.E.&Kudin,K.N.Longitudinalpolarizabilityofcarbon nanotubes. J.Phys.Chem.B 110 ,12860. 129.Kozinsky,B.&Marzari,N.Staticdielectricpropertiesofcarbonnanotubesfrom rstpriciples. Phys.Rev.Lett. 96 ,166801. 130.Guo,G.Y.,Chu,K.C.,Wang,D.-S.&Duan,C.-G.Staticpolarizabilityofcarbon nanotubes:abinitioindependent-particlecalculations. Comput.Mater.Sci. 30 269. 131.Brown,M.S.,Shan,J.W.,Lin,C.&Zimmermann,F.M.Electricalpolarizabilityof carbonnanotubesinliquidsuspension. Appl.Phys.Lett. 90 ,203108. 132.Lide,D.R. CRCHandbookofChemistryandPhysics.80thEdition CRCPress, BocaRaton,Florida,1999,80thedn. 133.Calderbank,K.E.,Le,R.J.W.,Fevre&Pierens,R.K.Molecularpolarizability. ortho-substitutedbenzylchlorides. J.Chem.Soc.B 968. 134.Vaughan,W.E. DigestofLiteratureonDielectrics NationalAcademyofSciences, Washington,DC,1975. 135.Lowry,T.H.&Richardson,K.S. MechanismandTheoryinOrganicChemistry Benjamin-CummingsPublishingCompany,1987,3rdedn. 136.Suppan,P.Invitedreview.solvatochromicshifts.theinuenceofthemediumon theenergyofelectronicstates. J.Photochem.Photobiol.,A 50 ,293. 137.Suppan,P.&Ghoneim,N. Solvatochromism TheRoyalSocietyofChemistry, Cambridge,UK,1997. 138.Smith,S.&Lasdon,L.Solvinglargesparsenonlinearprogramsusinggrg. J. Comput. 4 ,1. 139.Ziegler,K.J.,Lasdon,L.,Chlistunoff,J.&Johnston,K.P.Optimizationmodels fordeterminingnitricacidequilibriainsupercriticalwater. Comput.Chem. 23 421. 140.Davis,V.A. etal. Truesolutionsofsingle-walledcarbonnanotubesforassembly intomacroscopicmaterials. Nat.Nanotechnol. 4 ,830. 180

PAGE 181

141.Yu,A.,Su,C.-C.L.,Roes,I.,Fan,B.&Haddon,R.C.Gram-scalepreparation ofsurfactant-free,carboxylicacidgroupsfunctionalized,individualsingle-walled carbonnanotubesinaqueoussolution. Langmuir 26 ,1221.PMID: 19916485. 142.Zheng,M. etal. DNA-assisteddispersionandseparationofcarbonnanotubes. Nat.Mater. 2 ,338. 143.Bakota,E.L.,Aulisa,L.,Tsyboulski,D.A.,Weisman,R.B.&Hartgerink,J.D. Multidomainpeptidesassingle-walledcarbonnanotubesurfactantsincellculture. Biomacromolecules 10 ,2201. 144.Wenseleers,W. etal. Efcientisolationandsolubilizationofpristinesingle-walled nanotubesinbilesaltmicelles. Adv.Funct.Mater. 14 ,1105. 145.Ju,S.-Y.,Kopcha,W.P.&Papadimitrakopoulos,F.Brightlyuorescent single-walledcarbonnanotubesviaanoxygen-excludingsurfactantorganization. Science 323 ,1319. 146.Duque,J.G.,Densmore,C.G.&Doorn,S.K.Saturationofsurfactantstructureat thesingle-walledcarbonnanotubesurface. J.Am.Chem.Soc. 132 ,16165 147.Chen,W.-C.,Wang,R.K.&Ziegler,K.J.Coatingindividualsingle-walledcarbon nanotubeswithnylon6,10throughemulsionpolymerization. ACSAppl.Mater. Interfaces 1 ,1821. 148.Roe,R.-J. MethodsofX-rayandneutronscatteringinpolymerscience Oxford UniversityPress,NewYork,2000. 149.Higgins,J.S.&Benot,H.C. PolymersandNeutronScattering OxfordUniversity Press,NewYork,1996. 150.Fagan,J.,Landi,B.,Mandelbaum,I.&Simpson,J.Comparativemeasuresof single-wallcarbonnanotubedispersion. J.Phys.Chem.B 110 ,23801. 151.Wang,H. etal. Dispersingsingle-walledcarbonnanotubeswithsurfactants:A smallangleneutronscatteringstudy. NanoLett. 4 ,1789. 152.Dror,Y.,Pyckhout-Hintzen,W.&Cohen,Y.Conformationofpolymersdispersing single-walledcarbonnanotubesinwater:Asmall-angleneutronscatteringstudy. Macromolecules 38 ,7828. 153.Granite,M.,Radulescu,A.,Pyckhout-Hintzen,W.&Cohen,Y.Interactions betweenblockcopolymersandsingle-walledcarbonnanotubesinaqueous solutions:Asmall-angleneutronscatteringstudy. Langmuir 27 ,751. 181

PAGE 182

154.Hough,L.A.,Islam,M.F.,Hammouda,B.,aA.G.Yodh&Heiney,P.A.Structure ofsemidilutesingle-wallcarbonnanotubesuspensionsandgels. NanoLett. 6 313. 155.Zhou,W. etal. Smallangleneutronscatteringfromsingle-wallcarbonnanotube suspensions:evidenceforisolatedrigidrodsandrodnetworks. Chem.Phys.Lett. 384 ,185. 156.Kim,T.-H.,Doe,C.,Kline,S.&Choi,S.-M.Water-redispersibleisolated single-walledcarbonnanotubesfabricatedbyin-situpolymerizationofmicelles. Adv.Mater. 19 ,929. 157.Pethrick,R.&Dawkins,J.V. Moderntechniquesforpolymercharacterisation J. Wiley,NewYork,1999. 158.Prevost,S.&Gradzielski,M.SANSinvestigationofthemicrostructuresin catanionicmixturesofSDS/DTACandtheeffectofvariousaddedsalts. J.Colloid InterfaceSci. 337 ,472. 159.Bergstrom,M.&Pedersen,J.Structureofpuresdsanddtabmicellesinbrine determinedbysmall-angleneutronscatteringsans. Phys.Chem.Chem.Phys. 1 4437. 160.Hayter,J.B.&Penfold,J.Ananalyticstructurefactorformacroionsolutions. Mol. Phys. 42 ,109. 161.Hayter,J.B.&Penfold,J.Self-consistentstructuralanddynamicstudyof concentratedmicellesolutions. J.Chem.Soc,FaradayTrans.1 77 ,1851 162.Israelachvili,J. Intermolecular&SurfaceForces AcademicPress,Amsterdam, 1991,secondedn. 163.Lianos,P.&Zana,R.Fluorescenceprobestudiesoftheeffectofconcentrationon thestateofaggregationofsurfactantsinaqueoussolution. J.ColloidInterfaceSci. 84 ,100. 164.Hammouda,B.Clusteringinpolarmedia. J.Chem.Phys. 133 ,084901. 165.Hammouda,B.,,Ho,D.L.&Kline,S.Insightintoclusteringinpolyethyleneoxide solutions. Macromolecules 37 ,6932. 166.Moshammer,K.,Hennrich,F.&Kappes,M.M.Selectivesuspensioninaqueous sodiumdodecylsulfateaccordingtoelectronicstructuretypeallowssimple separationofmetallicfromsemiconductingsingle-walledcarbonnanotubes. Nano Res. 2 ,599. 167.Bauer,B.J. etal. Measurementofsingle-wallnanotubedispersionbysize exclusionchromatography. J.Phys.Chem.C 111 ,17914. 182

PAGE 183

168.Mukerjee,P.&Cardinal,J.R.Benzenederivativesandnaphthalenesolubilizedin micelles.polarityofmicroenvironments,locationanddistributioninmicelles,and correlationwithsurfaceactivityinhydrocarbon-watersystems. J.Phys.Chem. 82 1620. 169.Wasylishen,R.E. etal. Nmrstudiesofhydrocarbonssolubilizedinaqueous micellarsolutions. Can.J.Chem. 69 ,822. 170.Liu,G.G.,Roy,D.&Rosen,M.J.Asimplemethodtoestimatethesurfactant micellear-waterdistributioncoefcientsofaromatichydrocarbons. Langmuir 16 3595. 171.Sprunger,L.,WilliamEAcree,J.,&Abraham,M.H.Linearfreeenergyrelationship correlationofthedistributionofsolutesbetweenwaterandsodiumdodecylsulfate SDSmicellesandbetweengasandSDSmicelless. J.Chem.Inf.Model. 47 1808. 172.Rosen,M.J. SurfactantsandInterfacialPhenomena Wiley-Interscience,New York,2004. 173.Kumar,S.,Bansal,D.&udDin,K.Micellargrowthinthepresenceofsalts andaromatichydrocarbons:Inuenceofthenatureofthesalt. Langmuir 15 4960. 174.Patrick,H.N.,Warr,G.G.,Manne,S.&Aksay,I.A.Self-assemblystructuresof nonionicsurfactantsatgraphite/solutioninterfaces. Langmuir 13 ,4349 175.Hersam,M.C.Progresstowardsmonodispersesingle-walledcarbonnanotubes. Nat.Nanotechnol. 3 ,387. 176.Liu,J.&Hersam,M.C.Recentdevelopmentsincarbonnanotubesortingand selectivegrowth. MRSBull. 35 ,315. 177.Tanaka,T.,Urabe,Y.,Nishide,D.&Kataura,H.Continuousseparationofmetallic andsemiconductingcarbonnanotubesusingagarosegel. Apppl.Phys.Express 2 125002. 178.Liu,H.,Feng,Y.,Tanaka,T.,Urabe,Y.&Kataura,H.Diameter-selective metal/semiconductorseparationofsingle-wallcarbonnanotubesbyagarose gel. J.Phys.Chem.C 114 ,9270. 179.Sun,Z. etal. Quantitativeevaluationofsurfactant-stabilizedsingle-walledcarbon nanotubes:Dispersionqualityanditscorrelationwithzetapotential. J.Phys. Chem.C 112 ,10692. 180.Naumov,A.V. etal. Quantifyingthesemiconductingfractioninsingle-walled carbonnanotubesamplesthroughcomparativeatomicforceand photoluminescencemicroscopies. NanoLett. 9 ,3203. 183

PAGE 184

181.Striegel,A.M.,Yau,W.W.,Kirkland,J.J.&Bly,D.D. Modernsize-exclusionliquid chromatography:practiceofgelpermeationandgelltrationchromatography Wiley,Hoboken,N.J.,2009,2ndedn. 182.Tan,Y.,&Resasco,D.E.Dispersionofsingle-walledcarbonnanotubesofnarrow diameterdistribution. J.Phys.Chem.B 109 ,14454. 183.Liu,T.,Luo,S.,Xiao,Z.,Zhang,C.&Wang,B.Preparativeultracentrifugemethod forcharacterizationofcarbonnanotubedispersions. J.Phys.Chem.C 112 19193. 184.Dubin,P.ed. Aqueoussize-exclusionchromatography Elsevier,Amsterdam, 1988. 185.Scopes,R.K. Proteinpurication:priciplesandpractice Springer-Verlag,New York,1994,3rdedn. 186.Islam,A.M.,Phillips,G.,Sljivoa,A.,Snowdena,M.&Williams,P.Areviewof recentdevelopmentsontheregulatory,structuralandfunctionalaspectsofgum arabic. FoodHydrocolloids 11 ,493. 187.Silvera-Batista,C.A.,Weinberg,P.,Butler,J.E.&Ziegler,K.J.Long-term improvementstophotoluminescenceanddispersionstabilitybyowingsds-swnt suspensionsthroughmicrouidicchannels. J.Am.Chem.Soc. 131 ,12721 188.Cathcart,H. etal. OrderedDNAwrappingswitchesonluminescencein single-wallednanotubedispersions. J.Am.Chem.Soc. 130 ,12734 189.Jeng,E.S. etal. Detectionofdnahybridizationusingthenear-infraredband-gap uorescenceofsingle-walledcarbonnanotubes. NanoLett. 6 ,371. 190.Tanaka,T.,Jin,H.,Miyata,Y.&Kataura,H.High-yieldseparationofmetallicand semiconductingsingle-wallcarbonnanotubesbyagarosegelelectrophoresis. Apppl.Phys.Express 1 ,114001. 191.Tanaka,T. etal. Metal/semiconductorseparationofsingle-wallcarbonnanotubes byselectiveadsorptionanddesorptionforagarosegel. Phys.StatusSolidiB 247 2867. 192.Li,H. etal. Understandingtheelectrophoreticseparationofsingle-walledcarbon nanotubesassistedbythionineasaprobe. J.Phys.Chem.C 114 ,19234 193.Tummala,N.R.&Striolo,A.Curvatureeffectsontheadsorptionofaqueous sodium-dodecyl-sulfatesurfactantsoncarbonaceoussubstrates:Structural featuresandcounteriondynamics. Phys.Rev.E 80 ,021408. 184

PAGE 185

194.Nishide,D.,Liu,H.,Tanaka,T.&Kataura,H.Sortingsingle-wallcarbonnanotubes combininggelchromatographyanddensity-gradientultracentrifugation. Phys. StatusSolidiB 247 ,2746. 195.Jones,T.B. ElectromechanicsofParticles CambridgeUniversityPress,NewYork, 1995. 196.Dittrich,P.&Schwille,P.Anintegratedmicrouidicsystemforreaction, high-sensitivitydetection,andsortingofuorescentcellsandparticles. Anal. Chem. 75 ,5767. 197.Holmes,D.,Morgan,H.&Green,N.Highthroughputparticleanalysis:Combining dielectrophoreticparticlefocussingwithconfocalopticaldetection. Biosens. Bioelectron. 21 ,1621. 198.Holmes,D.,Sandison,M.E.,Green,N.G.&Morgan,H.On-chiphigh-speed sortingofmicron-sizedparticlesforhigh-throughputanalysis. IEEPNanobiotechnol. 152 ,129. 199.Morgan,H.,Holmes,D.&Green,N.Highspeedsimultaneoussingleparticle impedanceanduorescenceanalysisonachip. Curr.Appl.Phys. 6 ,367 200.Fiedler,S.,Shirley,S.,Schnelle,T.&Fuhr,G.Dielectrophoreticsortingofparticles andcellsinamicrosystem. Anal.Chem. 70 ,1909. 201.Krupke,R.,Hennrich,F.,Kappes,M.&Lohneysen,H.Surfaceconductance induceddielectrophoresisofsemiconductingsingle-walledcarbonnanotubes. NanoLett. 4 ,1395. 202.Kim,Y. etal. Dielectrophoresisofsurfaceconductancemodulatedsingle-walled carbonnanotubesusingcatanionicsurfactants. J.Phys.Chem.B 110 ,1541 203.Dimaki,M.&Boggild,P.Dielectrophoresisofcarbonnanotubesusing microelectrodes:anumericalstudy. Nanotechnology 15 ,1095. 204.Mattsson,M. etal. Dielectrophoresis-inducedseparationofmetallicand semiconductingsingle-wallcarbonnanotubesinacontinuousowmicrouidic system. J.Nanosci.Nanotechnol. 7 ,3431. 205.Shin,D.H. etal. Continuousextractionofhighlypuremetallicsingle-walledcarbon nanotubesinamicrouidicchannel. NanoLett. 8 ,4380. 206.Mendes,M.J.,Schmidt,H.K.&Pasquali,M.Browniandynamicssimulationsof single-wallcarbonnanotubeseparationbytypeusingdielectrophoresis. J.Phys. Chem.B 112 ,7467. 185

PAGE 186

207.Wang,L.,Flanagan,L.&Lee,A.P.Side-wallverticalelectrodesforlateraleld microuidicapplications. J.Microelectromech.S. 16 ,454. 208.Wang,L.,Flanagan,L.A.,Jeon,N.L.,Monuki,E.&Lee,A.P.Dielectrophoresis switchingwithverticalsidewallelectrodesformicrouidicowcytometry. LabChip 7 ,1114. 209.Satoh,A. IntroductiontoMolecular-MicrosimulationofColloidalDispersions Elsevier,Boston,2003. 210.Batchelor,G.K.Slender-bodytheoryforparticlesofarbitrarycross-sectionin stokesow. J.FluidMech. 44 ,419. 211.Cobb,P.&Butler,J.Simulationsofconcentratedsuspensionsofrigidbers: Relationshipbetweenshort-timediffusivitiesandthelong-timerotationaldiffusion. J.Chem.Phys. 123 ,054908. 212.Phelan,F.R.&Bauer,B.J.Simulationofnanotubeseparationineld-ow fractionationfff. Chem.Eng.Sci. 62 ,4620. 213.Blatt,S. etal. Inuenceofstructuralanddielectricanisotropyonthe dielectrophoresisofsingle-walledcarbonnanotubes. NanoLett. 7 ,1960 214.Ermak,D.&McCammon,J.Browniandynamicswithhydrodynamicinteractions. J. Chem.Phys. 69 ,1352. 215.Benedict,L.,Louie,S.&Cohen,M.Staticpolarizabilitiesofsingle-wallcarbon nanotubes. Phys.Rev.B 52 ,8541. 216.Kozinsky,B.&Marzari,N.Staticdielectricpropertiesofcarbonnanotubesfrom rstprinciples. Phys.Rev.Lett. 96 ,166801. 217.Fagan,J. etal. Dielectricresponseofalignedsemiconductingsingle-wall nanotubes. Phys.Rev.Lett. 98 ,147402. 218.Lu,W.,,Wang,D.&Chen,L.Near-staticdielectricpolarizationofindividualcarbon nanotubes. NanoLett. 7 ,2729. 219.Morgan,H.&Green,N.G. ACElectrokinetics:colloidsandnanoparticles ResearchStudiesPress,England,2003. 220.Wehrenberg,B.L.,Wang,C.&Guyot-Sionnest,P.Interbandandintrabandoptical studiesofPbSecolloidalquantumdots. J.Phys.Chem.B 106 ,10634 221.Valeur,B. MolecularFluorescence:PrinciplesandApplications Wiley-VCH, Weinheim,2002. 186

PAGE 187

222.Nair,N.,Usrey,M.L.,Kim,W.-J.,Braatz,R.D.&Strano,M.S.Estimationof then,mconcentrationdistributionofsingle-walledcarbonnanotubesfrom photoabsorptionspectra. Anal.Chem. 78 ,7689. 223.Parra-Vasquez,A.N.G. etal. Simplelengthdeterminationofsingle-walledcarbon nanotubesbyviscositymeasurementsindilutesuspensions. Macromolecules 40 4043. 187

PAGE 188

BIOGRAPHICALSKETCH CarlosA.SilveraBatistawasborninatropicalcityinthenorthcoastofColombia calledCartagena.There,hereceivedhishighschoolandmostofhisundergraduate education.Hepursuedabachelor'sdegreeinchemicalengineeringattheUniversidad deSanBuenaventura,Cartagena,forfouryears.Theopportunitytogettrained toahigherlevelcameearlierthanexpectedwhenhewasabletoconcludehis undergraduateandbegingraduatestudiesattheCityCollegeofNewYorkCCNY. OnceatCCNY,hehadthefortunetoparticipateinresearchactivitiesasaLouisStokes AllianceforMinorityParticipationLSAMPscholar.LSAMPallowedhimtospendtime mostlyonschoolrelatedactivities.Aspartoftheprogram,hehadthegreatopportunity ofworkingwithProf.Kretzschmarofthechemicalengineeringdepartmentasan LSAMPandlaterasaBridgetothedoctoratescholar. 188