<%BANNER%>

Investigation and Mitigation of Abiotic Factors Affecting Reliability of Microelectrode Arrays in Chronic Neural Implants

MISSING IMAGE

Material Information

Title:
Investigation and Mitigation of Abiotic Factors Affecting Reliability of Microelectrode Arrays in Chronic Neural Implants
Physical Description:
1 online resource (8 p.)
Language:
english
Creator:
Sankar, Viswanath
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Electrical and Computer Engineering
Committee Chair:
Nishida, Toshikazu
Committee Members:
Yoon, Yong Kyu
Xie, Huikai
Taylor, Curtis
Sanchez, Justin C

Subjects

Subjects / Keywords:
corrosion -- fabrication -- fib -- impedance -- insulation -- materials -- mems -- microelectrodes -- neural -- reliability
Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre:
Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Brain-Machine Interfaces are designed with the intention of improving the quality of life of individuals with motor disabilities such as paralysis by providing an alternative communication pathway via the control of prosthetic devices. One of the major challenges that the BMI research community is facing now is to develop a fully implantable system that is capable of recording high quality neural signals with high spatial resolution for the maximum number of years (preferably the lifetime of the patient). However the current electrodes fail after a few years of implantation owing to several biological and non-biological causes. A systematic study is needed to understand the degree of impact of the biotic and abiotic factors on the failure mode of electrode arrays. Such an understanding will help in the development of more robust and efficient implants which could be reliably used for chronic applications. The focus of this research is to investigate different abiotic factors such as micromotion-induced strain, insulation delamination, and corrosion of the recording site that play a major role in the failure of chronic microelectrodes. The first part of this work investigates the engineering of a highly compliant cable that can provide front-end strain relief to the tissue, and thereby mitigate tissue immune response due to micromotion. A serpentine shaped electrode was designed as a possible solution for front-end strain relief. The cable compliance was calculated using analytical and numerical methods, and prototypes were fabricated and tested. The test results showed that the serpentine cables are up to 10 times more compliant than the straight cables. In the second part, the association between the electrode surface morphology variations and the electrode impedance was evaluated using numerical simulations and models were developed using COMSOL Multiphysics finite element modeling (FEM) package and validated using in-vitro experiments. New materials such as benzocyclobutene and hafnium oxide, which are good candidates to be more robust than polyimide or parylene-C, were investigated for probe insulation. Microwire samples were also surface modified using Focused Ion Beam milling process at the recording surface and tested under in-vitro conditions.
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 Viswanath Sankar.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Nishida, Toshikazu.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2015-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045857:00001

MISSING IMAGE

Material Information

Title:
Investigation and Mitigation of Abiotic Factors Affecting Reliability of Microelectrode Arrays in Chronic Neural Implants
Physical Description:
1 online resource (8 p.)
Language:
english
Creator:
Sankar, Viswanath
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Electrical and Computer Engineering
Committee Chair:
Nishida, Toshikazu
Committee Members:
Yoon, Yong Kyu
Xie, Huikai
Taylor, Curtis
Sanchez, Justin C

Subjects

Subjects / Keywords:
corrosion -- fabrication -- fib -- impedance -- insulation -- materials -- mems -- microelectrodes -- neural -- reliability
Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre:
Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Brain-Machine Interfaces are designed with the intention of improving the quality of life of individuals with motor disabilities such as paralysis by providing an alternative communication pathway via the control of prosthetic devices. One of the major challenges that the BMI research community is facing now is to develop a fully implantable system that is capable of recording high quality neural signals with high spatial resolution for the maximum number of years (preferably the lifetime of the patient). However the current electrodes fail after a few years of implantation owing to several biological and non-biological causes. A systematic study is needed to understand the degree of impact of the biotic and abiotic factors on the failure mode of electrode arrays. Such an understanding will help in the development of more robust and efficient implants which could be reliably used for chronic applications. The focus of this research is to investigate different abiotic factors such as micromotion-induced strain, insulation delamination, and corrosion of the recording site that play a major role in the failure of chronic microelectrodes. The first part of this work investigates the engineering of a highly compliant cable that can provide front-end strain relief to the tissue, and thereby mitigate tissue immune response due to micromotion. A serpentine shaped electrode was designed as a possible solution for front-end strain relief. The cable compliance was calculated using analytical and numerical methods, and prototypes were fabricated and tested. The test results showed that the serpentine cables are up to 10 times more compliant than the straight cables. In the second part, the association between the electrode surface morphology variations and the electrode impedance was evaluated using numerical simulations and models were developed using COMSOL Multiphysics finite element modeling (FEM) package and validated using in-vitro experiments. New materials such as benzocyclobutene and hafnium oxide, which are good candidates to be more robust than polyimide or parylene-C, were investigated for probe insulation. Microwire samples were also surface modified using Focused Ion Beam milling process at the recording surface and tested under in-vitro conditions.
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 Viswanath Sankar.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Nishida, Toshikazu.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2015-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045857:00001


This item has the following downloads:


Full Text



PAGE 1

INVESTIGATIONANDMITIGATIONOFABIOTICFACTORSAFFECTINGRELIABILITYOFMICROELECTRODEARRAYSINCHRONICNEURALIMPLANTSByVISWANATHSANKARADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2013

PAGE 2

c2013ViswanathSankar 2

PAGE 3

Tomymother,fatherandfamily 3

PAGE 4

ACKNOWLEDGMENTS Tostartwith,Iwouldliketothankmyadvisor,Dr.ToshikazuNishida,forprovidingmeanopportunitytopursuemyPhDresearchunderhimandguidingmeallthroughmygraduatestudies.IwouldalsoliketothankmyPhDcommitteemembersDr.JustinSanchez,Dr.HuikaiXie,Dr.Y.K.Yoon,andDr.CurtisTaylorfortheiradviceandtechnicalinputsforthesuccessfulcompletionofthisproject.IwouldliketomakeaspecialacknowledgmenttoourcollaboratorsattheUniversityofMiami,Dr.JustinSanchezandDr.AbhishekPrasad.Theyhaveplayedintegralrolesinthisprojectrightfromitsinceptionandprovidedconstanttechnicalsupport.MydeepestthankstoDr.ErinPatrickforherinsightfulinputsrightfromthebeginningofmygraduateresearch.NumerousintensediscussionsthatIhadwithherhelpedmetotaketheprojectforwardintherightdirection.MyspecialthankstoDr.RobertDiemeforhelpingwiththeScanningElectronMicroscope(SEM)imagingoftheelectrodes.AlsoIwouldliketoextendmyacknowledgmentstoallofmyformerandpresentcolleaguesattheInterdisciplinaryMicrosystemsGroup(IMG)labsandattheUniversityofFlorida-AmitGupta,ShancyAugustine,PengZhao,PatrickLomenzoandallotherIMGmembers.Theywerewonderfulcolleaguesanditwasagreatexperienceworkingwiththem.SpecialthankstoDr.Y.K.Yoonandhisstudents,ArianandPitFeeJaoforprovidingtechnicalhelpwiththecoatingofBenzoclycobutene(BCB)andSU-8nanobers.SincerethankstoDr.CurtisTaylorandhisstudents,EdwardMcCumiskeyandNagidBrownandDr.HenrySodanoandhisformerstudentDr.GregoryEhlertforhelpingwiththecomplianceexperiments.ThankstoDr.NicholasRudawskiforhishelpwithdefectengineeringoftungstenmicrowiresusingFocusIonBeam(FIB)tool.SpecialthankstoDr.SaeedMoghaddamandhisstudentQanitfortheirhelpwiththecoatingofhafniumoxide(HfO2)onplatinum-iridium(Pt-Ir)microwires.ThankstoDr.AndrewRinzlerforprovidinguswithBCBsamples. 4

PAGE 5

MyacknowledgementstothestaffattheNanoscienceResearchFacility(NRF)andattheMajorAnalyticalInstrumentationCenter(MAIC)attheUniversityofFlorida(UF)fortheirsuggestionsrelatedtothefabricationandcharacterizationoftheelectrodes.ThankstothestaffatMaterialsEvaluationandTestingLaboratory(METLAB),SouthDakotaStateUniversity(SDSU)fortheirhelpwithsurfaceroughnessanalysisofelectrodes.SpecialthankstotheElectricalandComputerEngineering(ECE)departmentstaffatUFfortheirassistance.Iwouldliketothankthefundingsourcesofthisproject-NationalScienceFoundation(NSF),NationalInstituteofHealth(NIH)andDefenseAdvancedResearchProjectsAgency(DARPA).Lastbutnotleast,Iwouldliketoextendaveryspecialthankstomyfamilyandfriendsfortheirloveandencouragement.Withoutthesupportandsacriceofmymother,SanthaSankarandmyfather,SankarRamachandran,thiswouldnothavebeenpossible.Theconstantencouragementfrommybrother,Rammohan,sister-in-law,Shaliniandniece,Roshniwasinspiring.Finally,Iwouldliketothankallmyteachersfrommykindergartentomygraduateschoolwhomadethisdissertationareality. 5

PAGE 6

TABLEOFCONTENTS page ACKNOWLEDGMENTS .................................. 4 LISTOFTABLES ...................................... 10 LISTOFFIGURES ..................................... 11 ABSTRACT ......................................... 15 CHAPTER 1INTRODUCTORYREMARKS ............................ 17 1.1OverviewandMotivation ............................ 17 1.1.1Brain-MachineInterfaces(BMI) .................... 18 1.1.2KeyRequirementsforEffectiveBrain-MachineInterfaces ...... 20 1.1.3ChronicNeuralImplantChallenges-BioticandAbioticFactors .. 21 1.2ResearchGoalsandContributions ...................... 22 1.3DissertationOrganization ........................... 24 2BACKGROUND ................................... 26 2.1Introduction ................................... 26 2.2TheProblemofInterfacingtotheBrain .................... 26 2.2.1AnatomyandPhysiologyofBrain ................... 27 2.2.2Electrode-TissueInterfaceandInteraction .............. 30 2.2.3IntroductiontoNeuralElectrodesandFactorsAffectingReliabilityofNeuralElectrodes .......................... 32 2.3LiteratureReviewofBMISystemsandMicroelectrodeArrays ....... 34 2.3.1LiteratureReviewofBMISystems .................. 34 2.3.2LiteratureReviewofMicroelectrodeArrays .............. 36 2.3.3UFFlexibleMicroelectrodeArrays .................. 41 2.4AbioticFactorsAffectingReliabilityofMicroelectrodeArraysinChronicNeuralImplants ................................. 44 2.4.1LiteratureReviewoftheReportsPublishedonMicromotionInducedStrainEffects .............................. 44 2.4.2LiteratureReviewofReportsPublishedonEffectsofElectrodeGeometry,Size,Insulation,andSurfaceMorphologyVariationsonChronicSignalQuality ....................... 46 2.5Summary .................................... 47 3HIGHLYCOMPLIANTMODULAR2-DMICROELECTRODEARRAYSFORFRONT-ENDSTRAINRELIEF ........................... 48 3.1Introduction ................................... 48 3.2FlexibleCableandRigidModuleDesign ................... 48 6

PAGE 7

3.3AnalyticalandNumericalModelingofCableCompliance .......... 49 3.3.1AnalyticalAnalysis ........................... 50 3.3.2NumericalAnalysis ........................... 54 3.4FlexibleCableandRigidModuleFabrication ................. 54 3.5ExperimentalMeasurementoftheCableCompliance ........... 56 3.5.1ExperimentalMeasurementoftheIn-Plane(X-axis)CableCompliance ............................... 56 3.5.2ExperimentalMeasurementoftheOut-of-Plane(Y-axis)CableCompliance ............................... 60 3.5.3ExperimentalMeasurementoftheOut-of-Plane(Z-axis)CableCompliance ............................... 62 3.6ResultsandDiscussion ............................ 65 3.7Summary .................................... 67 4BACKGROUNDONELECTRODEIMPEDANCEANDINFLUENCEOFELECTRODESURFACEVARIATIONSONCHRONICSIGNALRECORDINGRELIABILITY ..................................... 69 4.1Introduction ................................... 69 4.2PhysicsofElectrodeImpedance ....................... 70 4.2.1Electrode-ElectrolyteInterface ..................... 70 4.2.2ChargeTransportinElectrode-ElectrolyteInterface ......... 73 4.2.3Electrode-ElectrolyteInterfaceinBiologicalEnvironment ...... 75 4.3ReviewofResultsfromPriorIn-vivoandIn-vitroStudies .......... 76 4.3.1ChronicPerformanceofCommonlyUsedInsulationMaterials ... 76 4.3.2ReviewofSEMImagesofimplantedTucker-DavisTechnologies(TDT)andUFTungstenMicroelectrodes ............... 77 4.3.3ReviewofIn-vivoImpedanceMeasurements ............ 80 4.3.4AnalysisofElectrodeSurfaceRoughnessusingKeyencerLaserScanningMicroscope ......................... 81 4.4ImpactofElectrodeSurfaceModicationonSignalQuality ........ 82 4.4.1EffectsofRecordingSiteCorrosiononSignalQuality ........ 82 4.4.2ImprovedRecordingUsingConductivePolymerandCarbonNanotubeCoatedNeuralElectrodes ................. 83 4.5Summary .................................... 84 5COMSOLFINITEELEMENTMODELINGSTUDYOFTHEEFFECTOFELECTRODESURFACEVARIATIONSONIMPEDANCE ............ 85 5.1Introduction ................................... 85 5.2ModelforElectrodeSurfaceMorphologybasedonLong-TermIn-vivoCharacterizationandIn-vitroCorrosionStudies ............... 85 5.3COSMOL3DModel .............................. 87 5.3.1ModelParameters ........................... 88 5.3.2Case1:EffectofRecordingSiteCorrosionandInsulationDelaminationonImpedance ...................... 90 7

PAGE 8

5.3.3Case2:EffectofRecordingSiteCorrosionandInsulationCrackingonImpedance .............................. 91 5.4ResultsandDiscussion ............................ 92 5.5Summary .................................... 93 6IN-VITROEXPERIMENTALEVALUATIONOFTHEEFFECTOFELECTRODESURFACEVARIATIONSONIMPEDANCE .................... 96 6.1Introduction ................................... 96 6.2ExperimentalDesign .............................. 96 6.3MicrowireSurfaceModicationsusingFocusedIonBeamandCorrosion 97 6.3.1EngineeringMetalRecessionusingControlledCorrosion ..... 98 6.3.2EngineeringInsulationDamagesusingFocussedIonBeam(FIB) 99 6.3.3ElectronMicroscopyImagingofSurfaceModiedMicrowires ... 99 6.4ImpedanceMeasurementusingNanoZrImpedanceMeter ........ 101 6.4.1CalibrationofNanoZrInstrument ................... 104 6.4.2MicrowireSamplePreparation ..................... 105 6.4.3ImpedanceMeasurementProcedure ................. 105 6.5ResultsandDiscussion ............................ 106 6.5.1In-vitroImpedanceMeasurementResults .............. 107 6.5.1.1ImpedanceMeasurementonPhosphateBufferedSaline(PBS)CorrodedMicrowires ................. 107 6.5.1.2ImpedanceMeasurementonFIBInsulationModiedMicrowires .......................... 108 6.5.2ComparisonwithIn-vivoResults ................... 109 6.6Summary .................................... 129 7INVESTIGATIONOFMATERIALSANDENGINEERINGMETHODSTOBUILDMICROELECTRODESWITHIMPROVEDSTABILITYANDREDUCEDIMPEDANCE ..................................... 131 7.1Introduction ................................... 131 7.2MaterialChoicesforProbeInsulationandImpedanceReduction ..... 131 7.3DesignConsiderations ............................. 134 7.3.1AnalyticalCalculationofOptimalInsulationThickness ....... 134 7.3.2NumericalAnalysisofOptimalSU-8Thickness ........... 136 7.4FabricationProcess .............................. 137 7.5CharacterizationofBCBInsulation ...................... 141 7.5.1DurabilityEvaluationusingSoakTest ................. 141 7.5.2SurfaceMorphologyEvaluationusingSEMImages ......... 141 7.6SurfaceMorphologyEvaluationofSU-8nanobers ............ 143 7.7FIBSurfaceModicationforReducedImpedance .............. 143 7.8ResultsandDiscussion ............................ 150 7.9Summary .................................... 152 8

PAGE 9

8SUMMARYANDFUTUREWORK ......................... 154 8.1ResearchSummary .............................. 154 8.2RecommendationsforFutureWork ...................... 156 REFERENCES ....................................... 158 BIOGRAPHICALSKETCH ................................ 173 9

PAGE 10

LISTOFTABLES Table page 2-1Bioticandabioticfactorsassociatedwiththefailureofchronicrecordingcapabilitiesofmicroelectrodearrays. .............................. 33 2-2OverviewmatrixofdifferentBMIsystems. ..................... 36 3-1Dimensionsofthestraightandserpentinecablesusedintheanalyticalandnumericalanalysis. .................................. 53 3-2ListofYoung'smodulusvaluesofpolyimideobtainedfromliterature. ...... 53 3-3Matrixcomparingthestraightcablecomplianceestimatedthroughanalytical,numericalandexperimentalanalysis. ....................... 67 3-4Matrixcomparingtheserpentinecablecomplianceestimatedthroughanalytical,numericalandexperimentalanalysis. ....................... 67 5-1ConductivityvaluesusedforCOMSOLniteelementmodeling. ......... 89 7-1Tablecomparingthekeymaterialpropertiesoftheproposedandtheexistingelectrodeinsulationmaterials. ............................ 133 7-2ConductivityvaluesusedforCOMSOLniteelementmodeling. ......... 136 10

PAGE 11

LISTOFFIGURES Figure page 1-1Qualityofrecordedsignalduringdifferentphasesofelectrodelifetime[ 1 ]c2011IEEE. ...................................... 22 2-1Schematicrepresentationofaneuron. ....................... 29 2-2Schematicrepresentationofintracellularactionpotential. ............ 30 2-3Schematicrepresentationofdifferentphasesofelectrode-tissuereaction. ... 32 2-4Photographsofdifferenttypesofmicrowireelectrodearrays. .......... 38 2-5Photographsofsiliconmicromachinedelectrodearrays. ............. 40 2-6UFgeneration1microelectrodearray[ 2 ]c2006IEEE. ............. 42 2-7UFgeneration2microelectrodearray[ 3 ]c2008IEEE. ............. 42 2-8UFgeneration2amplierintegratedmicroelectrodearray[ 4 ]c2010IEEE. .. 43 3-1Illustrationofthemodularelectrodedesign. .................... 49 3-2Schematicandfree-bodydrawingsofelectrodecables. ............. 51 3-3Processowfor2Dtransverseexiblecablemodule. ............... 55 3-4Processowforrigidelectronics/connectormodule. ............... 57 3-5Photographsoffabricateddevices. ......................... 58 3-6SchematicandphotographofcablemountedonInstronrmechanicaltestingsystem. ........................................ 59 3-7MeasuredX-axiscompliancevaluesfromstraightandserpentinecables. ... 60 3-8Photographshowingthestraightcablemountedverticallyonaglassslideandsecuredwithmagnetpiecesandsteelnutsoneithersides. ........... 61 3-9SchematicdiagramshowingtheverticalloadingoftheserpentinecableinY-axisbythenanoindentertip. ........................... 62 3-10MeasuredY-axiscompliancevaluesfromstraightandserpentinecables. ... 63 3-11Photographsofmounteddevices. ......................... 64 3-12SchematicdiagramshowingtheverticalloadingofthecableintheZ-axisbythenanoindentertip. ................................. 65 3-13MeasuredZ-axiscompliancevaluesfromstraightandserpentinecables. ... 66 11

PAGE 12

4-1Electrode-electrolyteinterfaceatequilibrium[ 5 ]. ................. 71 4-2Electricalequivalentcircuitfornonfaradaicelectrode-electrolyteinterface. ... 72 4-3Randleselectricalequivalentcircuitforelectrode-electrolytefaradaicinterface. 72 4-4Electricalequivalentcircuitfornon-idealfaradaicinterfacewithCPE. ...... 73 4-5Electricalequivalentcircuitforelectrode-electrolyteinterfaceinabiologicalenvironment. ..................................... 75 4-6SEMimagesandneuronalyieldplotoftungstenmicroelectrodeswithpolyimideinsulation. ....................................... 79 4-7SEMimagesofplatinum-iridiumelectrodeswithparylene-Cinsulation. ..... 80 4-8ImagesofelectrodeR9obtainedfromKeyencerlaserscanningmicroscope. 82 5-1Schematicillustrationoftheproposedhypothesisexplainingthein-vivoimpedancetrend. ......................................... 88 5-2IllustrationoftheCOMSOLniteelement3Dmodeloftheelectrode-electrolyteinterface. ....................................... 89 5-3IllustrationoftheboundaryconditionsusedintheCOMSOLniteelementmodel. ......................................... 90 5-4IllustrationofCase1electrodesurfacemodicationoverimplantduration. ... 91 5-5IllustrationofCase2electrodesurfacemodicationoverimplantduration. ... 92 5-6COMSOL3Dsimulatedresults. ........................... 94 6-1FlowchartofthestepsinvolvedinIn-vitroimpedancemeasurementstudyofcorrodedwires. .................................... 97 6-2FlowchartofthestepsinvolvedinIn-vitroimpedancemeasurementstudyofinsulationcrackedwires. ............................... 98 6-3SEMimagesoftungstenmicrowiresamplenumber2. .............. 100 6-4SEMimagesoftungstenmicrowiresamplenumber4. .............. 101 6-5SEMimagesoftungstenmicrowiresamplenumber3. .............. 102 6-6SEMimagesoftungstenmicrowiresamplenumber5. .............. 103 6-7PhotographoftheNanoZrNZ-CALcalibrationadaptor.FigureadaptedfromNanoZrusermanual[ 6 ]. .............................. 104 6-8In-vitroimpedancemeasurementexperimentsetup. ............... 107 12

PAGE 13

6-9Averageimpedancemeasuredfromsamplenumber2beforeandafterimmersioninPBS. ........................................ 108 6-10Averageimpedancemeasuredfromsamplenumber4beforeandafterimmersioninPBS. ........................................ 109 6-11Averageimpedancemeasuredfromsamplenumber3beforeandafterFIBinsulationmodication. ................................ 110 6-12Averageimpedancemeasuredfromsamplenumber5beforeandafterFIBinsulationmodication. ................................ 111 6-13Percentagechangeintheaveragein-vivoimpedanceforallwiresofelectrodeR9plottedagainsttheimplanteddurationandttedwithaGaussiancurve. .. 112 6-14Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforwires1-4ofelectrodeR9. ......... 114 6-15Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforwires5-8ofelectrodeR9. ......... 115 6-16Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforwires9-12ofelectrodeR9. ......... 116 6-17Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforwires13-16ofelectrodeR9. ........ 117 6-18FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire1. ............................... 119 6-19FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire3. ............................... 120 6-20FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire7. ............................... 121 6-21Frequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofcorrodedtungstenwiresample2. ................. 122 6-22FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofFIBmodiedtungstenwiresample3. ............... 123 6-23RealandimaginarypartsoftheimpedanceofR9wire1at1kHzwithrespecttoimplantduration. .................................. 124 6-24RealandimaginarypartsoftheimpedanceofR9wire3at1kHzwithrespecttoimplantduration. .................................. 125 6-25RealandimaginarypartsoftheimpedanceofR9wire7at1kHzwithrespecttoimplantduration. .................................. 126 13

PAGE 14

6-26Realandimaginarypartsoftheimpedanceofcorrosionstudysample2at1kHzwithrespecttodifferentsurfaceconditions. .................. 127 6-27Realandimaginarypartsoftheimpedanceofinsulationcrackingstudysample3at1kHzwithrespecttodifferentsurfaceconditions. .............. 128 7-1IllustrationshowingBCB-HfO2insulatedPt-Irmicrowiredesign. ......... 134 7-2Schematicshowingthechargedistributionandelectriceldpatternofthecylindricalcapacitancemodel. ................................. 135 7-3PercentagereductioninimpedancefordifferentthicknessesofanuniformlmofSU-8coatingonPt-IrwiresascomparedtononSU-8coatedtungstenwires. ......................................... 137 7-4SchematicshowingthefabricationprocessowforBCB-HfO2insulatedPt-IrwirescoatedwithcarbonizedSU-8nanobers. .................. 138 7-5SEMimagesofhighviscosityBCBandlowviscosityBCBcoatedPt-Irmicrowires. ...................................... 140 7-6PhotographsofSU-8electrospinningprocess. .................. 142 7-7SEMimagesofpolyimideinsulatedtungstenmicrowiresamples1-3takenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2. ..... 144 7-8SEMimagesofpolyimideinsulatedtungstenmicrowiresamples4-6takenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2. ..... 145 7-9SEMimagesofBCBinsulatedPt-Irmicrowiresamples1-3takenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2. ............ 146 7-10SEMimagesofBCBinsulatedPt-Irmicrowiresamples4-6takenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2. ............ 147 7-11SEMimagesofPt-IrmicrowirescoatedwithcarbonizedSU-8nanobers. ... 148 7-12SEMimagesoftungstenmicrowirebeforeandafterFIBsurfaceroughnessmodication. ..................................... 149 7-13SEMimagesofBCBinsulatedPt-IrmicrowirebeforeandafterFIBsurfaceroughnessmodication. ............................... 149 7-14FrequencyspectrumoftheaverageimpedancemeasuredfromtungstenmicrowiresamplebeforeandafterFIBsurfacemodication. ................. 151 7-15FrequencyspectrumoftheaverageimpedancemeasuredfromPt-IrmicrowiresamplebeforeandafterFIBsurfacemodication. ................. 152 14

PAGE 15

AbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulllmentoftheRequirementsfortheDegreeofDoctorofPhilosophyINVESTIGATIONANDMITIGATIONOFABIOTICFACTORSAFFECTINGRELIABILITYOFMICROELECTRODEARRAYSINCHRONICNEURALIMPLANTSByViswanathSankarAugust2013Chair:ToshikazuNishidaMajor:ElectricalandComputerEngineering Brain-MachineInterfacesaredesignedwiththeintentionofimprovingthequalityoflifeofindividualswithmotordisabilitiessuchasparalysisbyprovidinganalternativecommunicationpathwayviathecontrolofprostheticdevices.OneofthemajorchallengesthattheBMIresearchcommunityisfacingnowistodevelopafullyimplantablesystemthatiscapableofrecordinghighqualityneuralsignalswithhighspatialresolutionforthemaximumnumberofyears(preferablythelifetimeofthepatient).Howeverthecurrentelectrodesfailafterafewyearsofimplantationowingtoseveralbiologicalandnon-biologicalcauses.Asystematicstudyisneededtounderstandthedegreeofimpactofthebioticandabioticfactorsonthefailuremodeofelectrodearrays.Suchanunderstandingwillhelpinthedevelopmentofmorerobustandefcientimplantswhichcouldbereliablyusedforchronicapplications. Thefocusofthisresearchistoinvestigatedifferentabioticfactorssuchasmicromotion-inducedstrain,insulationdelamination,andcorrosionoftherecordingsitethatplayamajorroleinthefailureofchronicmicroelectrodes.Therstpartofthisworkinvestigatestheengineeringofahighlycompliantcablethatcanprovidefront-endstrainrelieftothetissue,andtherebymitigatetissueimmuneresponseduetomicromotion.Aserpentineshapedelectrodewasdesignedasapossiblesolutionforfront-endstrainrelief.Thecablecompliancewascalculatedusinganalyticaland 15

PAGE 16

numericalmethods,andprototypeswerefabricatedandtested.Thetestresultsshowedthattheserpentinecablesareupto10timesmorecompliantthanthestraightcables. Inthesecondpart,theassociationbetweentheelectrodesurfacemorphologyvariationsandtheelectrodeimpedancewasevaluatedusingnumericalsimulationsandmodelsweredevelopedusingCOMSOLMultiphysicsniteelementmodeling(FEM)packageandvalidatedusingin-vitroexperiments.Newmaterialssuchasbenzocyclobuteneandhafniumoxide,whicharegoodcandidatestobemorerobustthanpolyimideorparylene-C,wereinvestigatedforprobeinsulation.MicrowiresampleswerealsosurfacemodiedusingFocusedIonBeammillingprocessattherecordingsurfaceandtestedunderin-vitroconditions. 16

PAGE 17

CHAPTER1INTRODUCTORYREMARKS 1.1OverviewandMotivation Withnearly100billionneuronsandcloseto100trillionconnections,thehumanbrainandthenervoussystemisoneofnature'sengineeringmarvels.Encompassingsuchanintricateandcomplexnetwork,anyinjurycausedtothenervoussystemwillhaveahugeimpactonthequalityoflifeoftheindividual.Neurologicaldisordersaffectingthecentralnervoussystem(brain)andperipheralnervoussystem(spinalcordandothernervesinthebody)resultinseriousandpermanentdamagetothebody,suchasparalysisofmotorfunctionality.TheNationalInstituteofNeurologicalDisordersandStroke(NINDS)hasidentiedmorethan600neurologicaldisorders[ 7 ].Thelistincludessomeofthemostcommonmovementdisorderssuchasepilepsy,stroke,spinalcordinjury,andParkinson'sdisease(PD).Stroke,whichcausesabout29%ofparalysiscases[ 8 ],isthethirdleadingcauseofdeathintheUnitedStates(U.S.),andeachyearroughlyabout795,000peoplesufferfromstroke[ 9 ].Spinalcordinjury(SCI),whichcontributesabout23%ofparalysiscases[ 8 ],constitutesabout12,000newcaseseveryyear[ 10 ]andithasbeenestimatedfor2012,about270,000peopleintheU.S.willhaveSCI[ 10 ].About3millionpeopleintheUnitedStatesareaffectedwithepilepsyandapproximately150,000newcasesarebeingdiagnosedeveryyear[ 11 ].Alsoitisestimatedthatnearly$15.5billionisspenteveryyearforepilepsymedicalexpenses[ 11 ].ThenumberofpeopleaffectedwithParkinsonsdisease(PD)in2005intheU.S.wasapproximately349,000[ 12 ]andin2010,itwasroughly1.1million[ 13 ].Itisprojectedthatthenumberwouldriseto1.8million(approx.)in2030and2.5million(approx.)in2050[ 13 ]. Itisevidentfromthesefactsandguresthattheimpactofneurologicaldisordersinthephysical,social,andtheeconomicallifeoftheaffectedindividualisenormous.Anyprogressaimedatregainingthelostfunctionalityandimprovingthequalityoflife 17

PAGE 18

ofthesepatientswouldbeseenasamajorcontributiontomankind.Drivenbysuchaphilanthropicmotivation,neuralprosthetictechnologiesarebeingdevelopedtowardsrecoveringthelostfunctionalityofthepatientsaffectedwithneurologicaldisordersbyprovidinganalternativecommunicationpathwaybetweenthenervoussystemandanexternalsyntheticdevice,suchasaprostheticlimb.Neuroprosthesesestablishcommunicationbetweenthenervoussystemandtheexternaldeviceeitherbyprovidingelectricalstimulationtotheneuronsorbyrecordingelectricalsignalsfromthenervecells.Further,basedontheirtargetlocations,theprostheticsaresub-classiedassensoryprostheticsandmotorprosthetics. Cochlearimplants[ 14 15 ]andretinalimplants[ 16 ]areclassicexamplesofsensorystimulatingneuralprosthetics.Otherexamplesofstimulatingneuralprosthesesincludedeep-brainstimulationtherapiescarriedoutforParkinson'sdisease[ 17 ],epilepsy[ 18 19 ],anddepression[ 20 ].Signalsareelectricallytransmittedfromthesensoryprostheticstodesiredregionsofthepatient'sbraininordertoachievetheexpectedimprovementinthesensoryfunctionortoalleviatetheirpain. Motorrecordingneuralprostheticsincludeprostheticsassociatedwiththeperipheralnervoussystem(PNS)andprostheticsassociatedwiththecentralnervoussystem(CNS).PNSprostheticsexamplesincludeanarticialroboticarmthatisinterfacedwiththeperipheralnerveslocatedneartheshoulderofanamputee[ 21 22 ].CNSprostheticsarecommonlyreferredtoasBrain-MachineInterface(BMI)orBrain-ComputerInterface(BCI).BothBMIandBCIservetointerprettheindividual'sthoughtsthroughrecordingneuralactivityandexecutingthedesiredactionontheexternalarticialdevice[ 23 25 ]. 1.1.1Brain-MachineInterfaces(BMI) BMIsorBCIshavealonghistorystartingwiththeinceptionofelectroencephalography(EEG)byHansBergerin1929[ 26 ].AfterBergerdemonstratedthepossibilityofrecordingelectricalactivityfromhumanbraincells,neuroscientists,engineersand 18

PAGE 19

neurosurgeonsstartedshowinginterestinthedevelopmentofacommunicationsystemthatwillpermitthebraintotalkwiththesurroundingswithoutneedingtousethetraditionalchannelsofperipheralnervesandmuscles.ThisledtothebirthofBMIsystems. Wolpawet.al.,havedenedtheBrain-ComputerInterfaceasacommunicationsystemthatdoesnotdependonthebrain'snormaloutputpathwaysofperipheralnervesandmuscles[ 27 ].Tohelprealizethealternatecommunicationpathway,BMIsystemsincludethefollowingfourbuildingblocks[ 28 ]: 1. arecordingunittoacquiretheelectricalactivityfromtheneurons, 2. aprocessingunittotranslatetherawelectricalsignaltomeaningfulinformation, 3. anoutputunit,wheretherecoveredintentionsareexecutedasactions,and 4. anoperatingprotocolthatclosestheloopbyprovidingfeedbacktotheuser. DevelopmentofBMIsystemsbeganin1980,whentheearlyreportonthepreclinicalstudydemonstratingthefunctionalityofsimplecommunicationsystemsonprimateswaspublishedbySchmidt[ 29 ].Schmidt'sstudyinvestigatedthelong-termimplantationofmicroelectrodesonmonkeysandexaminedthelikelihoodoftrainingtheimplantedmonkeystomodifytheneuronalringpatternstocontrolanexternaldevice[ 29 ].In1999,apreclinicalstudybyChapinet.al.[ 30 ],demonstratingthesynchronizedcontrolofaroboticarmthroughcorticalrecordingsinaratmodelpromotedtheinterestsinBMIfurther.Sincethen,severalpreclinicalandclinicalstudieshavebeenconductedonrodents[ 31 ],primates[ 32 36 ],andhuman[ 37 39 ]subjects.RecentstudiesonBMIs,whichdemonstratedthecapabilityoftheinterfacestoenablepatientswithtetraplegiatomoveacursoronacomputerscreenandreachandgraspobjectsusingaroboticarm[ 37 39 ],arestandingexamplesoftheadvancementmadeinthiseld.However,thereisalongwaytogoinordertorealizeanidealBMI. 19

PAGE 20

1.1.2KeyRequirementsforEffectiveBrain-MachineInterfaces TherearethreekeyrequirementstobefullledtorealizeanidealBMIsystem[ 23 38 40 ].Theultimateinterfaceshouldenabletheusertohavecompletemobilityandallowthemtocontroltheirarticialprostheseswithmultipledegreesoffreedomreliablythroughouttheirlifetime.Unrestrictedmovementcanbeachievedonlywithaself-sustainable,fully-implantablewirelessBMIsystem.Thisimpliesrealizinganinterfacethatintegratestherecordingmicroelectrodearray,electroniccircuitry(i.e.,amplierandencoder),wirelesstelemetry(i.e.,transmitterelectronicsandantenna),andpowermodule(i.e.,rechargeablebatterypack)inanimplantableformfactor.Severaleffortsarebeingpursuedbyvariousgroupstorealizeandimprovesuchaninterface[ 41 44 ]. Awiderangeofexibilityinoperatingthearticialprosthesescouldbeaccomplishedonlywithalargesampleofhighresolutionactionpotentialsobtainedfromtheneurons.Thisrequiresmicroelectrodearrayswithhighspatialresolution.Neuralimplantswithhighchannelcountelectrodesarebeingdevelopedtoaddressthisrequirement.Aclassicexampleisthe100channelmicroelectrodearraydevelopedbytheUniversityofUtah[ 45 ]. ReliablechronicrecordingcapabilitieswithuncompromisingsignalqualityforalongperiodoftimeisthethirdkeyrequirementforanidealBMIsystem.However,interfacesdevelopedsofarhavenotabletomeetthisrequirement,astheyfailafterafewyearsofsuccessfulrecordingandfunctioning.Thisiscurrentlyoneofthemajorchallengesfortheresearchcommunityandseveralreasonsareconsideredaspotentialcontributingfailuremodesoftherecordingsystem.Differentfactorsassociatedwithchronicfailureofneuralrecordingsystemsarediscussedbrieyinthefollowingsectionandelaboratedinchapter 2 20

PAGE 21

1.1.3ChronicNeuralImplantChallenges-BioticandAbioticFactors Severalstudiesintherecentpasthaveobservedthattheperformanceofneuralelectrodesimplantedchronicallyinthesubjectsexperienceatemporaldegradationoftheelectrodequalityandrecordingmetricssuchassignaltonoiseratio,noiseoor,peakamplitude,neuronalyield,andtheabilitytomeasureconsistentactionpotentialsfromthesameneuronfortheimplantduration.Potentialfactorscontributingtothefailureofchronicimplantsareputforthfromtwodifferentparadigms-bioticandabiotic[ 1 ].Thebioticmodelpertainstothephysiologyandtheanatomyoftheanimalorhumansubject,anditincludeseffectssuchasinammatoryresponse,blood-brainbarrier(BBB)disruption,astrogliosisinitiation,synapticchanges,andmicrogliaandmacrophagesrecruitment.Theabioticmodelincludesthefactorsrelatedtothephysicalmodicationsintheelectrodesuchaschangesingeometry,insulationfailure,instabilityinrecordingsite(includingcorrosionandothermaterialfailure),andstrainduetomicromotion.Commonmetricsmeasuredandevaluatedinthebioticmodelarethebiomarkersofinjuryandmorphometry,whilethoseintheabioticmodelareimpedanceandsignaltonoiseratio(SNR). Allthroughthedurationoftheimplantedelectrode,boththebioticandtheabioticeffectsappeartobeoccurringsimultaneouslyandtheirindependentinuenceonthedegradationofthesignalrecordingqualityisyettobeexploredinasystematicway[ 1 ].However,ithasbeenestablishedthatthecollectiveeffectsofboththefactorsaffecttheelectrophysiologicalsignalqualityoftheimplantinphases.Fourkeyphaseshavebeenidentiedthatareassociatedwithelectrodefailure[ 1 ].Theyare: 1. Acutephase,whichlastsforafewhoursafterimplantation.Duringthisphase,thesignalqualityishighandthereissomedamagetotissueduetoinsertiontraumaandonsetofwoundresponse. 2. Recoveryphase,whichextendsforuptotwoweeksafterimplantation.Thersthalfofthisphaseexperiencesasubstantialdropinthesignalqualityduetoactiveinammatoryresponse,whilethesecondhalfofthisphaseseesarapidincreaseintherecordingqualityasthereissomereductionintheintialwoundresponse. 21

PAGE 22

Figure1-1. Qualityofrecordedsignalduringdifferentphasesofelectrodelifetime[ 1 ]c2011IEEE. 3. Chronicphase,whichistheactiverecordingperiodandextendsforafewmonths.Theimplantexperiencesconstantforeignbodyresponseandotherbioticeffectsaswellasundergoesvariousabioticphysicalmodications.Duringthisphase,thereisagradualdecreaseinthesignalquality. 4. Failurephaseisthelastphaseoftheimplant,whereallthebioticandabioticeffectshaveaccumulatedandleadtothecompletefailureoftheelectrode.Duringthisphase,theelectrodefailstorecordelectricalactivityfromtheneurons. Figure 1-1 showsthefourphasesoftheelectrodelifetimeandthequalityoftherecordedsignalduringeachphase. 1.2ResearchGoalsandContributions Thisdissertationworkisaimedataddressingsomeofthekeyabioticissuescontributingtothefailureofneuralimplantsinchronicrecordingapplications.Therstissueaddressedinthisworkisanengineeringsolutiontothestraininneuralimplantsduetomicromotion.Ahighlycompliantserpentineshaped2Dmicroelectrodepolyimidecabledesignwasdesigned,fabricatedandcharacterizedforfront-endstrainrelief.The 22

PAGE 23

designalsoincludedarigidsiliconmoduleservingasaplatformforconnectors,groundscrewandotherelectronics.PrototypesofboththecablesandtherigidmodulesweremicrofabricatedusingMEMSfabricationprocesses.Theoreticalandnumericalmodelsweredevelopedtovalidatethecomplianceofthenewhighlycompliantserpentinecabledesign.Further,inordertocorroboratethemodels,thecomplianceofthefabricatedcableprototypesaremeasuredusingmechanicalstressexperimentsinallthreeaxes.Theexperimentalresultsshowedanincreaseincablecomplianceofupto10timesinthenewserpentinedesigncomparedwithstraightcablesofsamedimensions.Theseresultsimplythatwithanincreasedcompliancethenewserpentinecableswillbeabletoprovidestrainreliefforrecordingmicrowiresandhencealleviatethestraininducedtissueresponse.Areducedtissueresponseisexpectedtoimprovethelongevityoftherecordingelectrodesandthequalityofthechronicrecordings. TheDARPAfundedHistologyforInterfaceStabilityoverTime(DARPA-HIST)project,acollaborativeeffortbetweentheUniversityofFloridaandtheUniversityofMiami,isaimedatunderstandingthefailuremechanismsforin-vivoneuralimplants,bydevelopingaresearchplanthatcouplesthestudyofbioticandabioticmetricsinasystematicway.Thisresearchwork,beingapartofthatprojectwasalsofocussedoninvestigatingtheeffectofelectrodesurfacevariationsonelectrodeimpedancetowardsunderstandingtheimpactofelectrodeabioticfactorsonchronicsignalrecordingreliability.NumericalmodelsweredevelopedusingCOMSOLniteelementmodelingsoftwaretostudytheimpactofrecordingsitecorrosionandinsulationdelaminationandcrackingonelectrodeimpedance.Themodelswerevalidatedwithin-vitroexperimentsthatincludedimpedancemeasurementsondefectengineeredtungstenmicrowiresinsalineenvironment.Theresultswerecomparedwithin-vivomeasurementsandanin-depthanalysiswasmadetounderstandtheeffectofdifferentfactorsonelectrodeimpedance.Thenumericalresultsalongwithexperimentalanalysisprovidedgoodinsightonthephysicsbehindthetemporalvariationofelectrodeimpedance.Theresults 23

PAGE 24

alsoledtoaphysicalmodelthatcanpossiblyserveasascienticbasistounderstandtheinuenceofelectrodesurfacevariationonimpedance. Moreover,thedesignforarobustmicroelectrodewithlowimpedancerecordingsiteisintroducedinthiswork.Thedesignincludesplatinum-iridiumastherecordingmicrowireandbenzocyclobutene(BCB)andhafniumoxide(HfO2)asinsulationmaterialsandnormalincidenceFocussedIonBeam(FIB)toreduceelectrodeimpedance.Prototypeswerefabricatedandsampleswerecharacterizedforrobustnessandreducedimpedance.Preliminaryresultsshowgoodstabilityunderaqueousenvironmentandmuchreducedimpedanceincomparisonwithexistingelectrodematerials,suggestingthatthenewdesignelectrodesmayhavebetterchronicperformance. 1.3DissertationOrganization Thisdissertationdocumentincludeseightchapters.Theremainingchaptersofthisdocumentareorganizedasfollows.Chapter 2 givesanindepthreviewoftheexistingBMIandneuralelectrodetechnologies,aswellaselaboratesontheabioticfactorsaffectingthereliabilityofneuralrecordinginchronicimplants.Theanatomyandthephysiologyofthebrain,thebiophysicsbehindtheelectrode-tissueinterface,andthebiologyoftheelectrode-tissueimmuneresponsearealsodescribed.Furthermore,theexperiencegainedthroughbuildinggeneration1andgeneration2microelectrodesatUFisbrieydiscussedalongwiththedesign,fabricationandtestresultsofmonolithicexiblemicroelectrodearraysintegratedwithpre-amplieronapyrexbase,whichisanupgradeofthegeneration2microelectrodearray.Chapter 3 discussesthedesign,fabrication,andmechanicaltestingofthehighlycompliantserpentineshapedmicroelectrodecables.Alsoitgivesadetailedexplanationoftheanalyticalandnumericalmodelsaswellasdiscussestheexperimentalresults.Chapter 4 providesabackgroundoftheelectrode-electrolyteimpedanceanddiscussestheimpactofelectrodesurfacevariationonelectrodeimpedance.Theseincludethephysicsofelectrode-electrolyteimpedance,chronicperformanceofcommonlyused 24

PAGE 25

insulationmaterials,observationsfromSEMimages,in-vivoimpedancemeasurements,andtheeffectsofrecordingsitecorrosionandsurfacemodicationonsignalquality.Chapter 5 presentsamodelforelectrodesurfacemorphologybasedonlong-termin-vivocharacterizationandin-vitrocorrosionstudies.Also,COMSOL3Dniteelementanalysisofthetwocommoncasesofelectrodesurfacevariationsandtheirinuenceonimpedanceisdiscussed.Thesimulationmethodsandresultsaredescribedindetailandacomparisonofthenumericalresultswithin-vivoimpedancemeasurementsisgiven.Chapter 6 presentsin-vitroexperimentalmethodsandresultsforevaluatingtheeffectofelectrodesurfacevariationonimpedance.Chapter 7 describesthedesign,fabrication,andcharacterizationofHfO2-BCBinsulatedmicroelectrodeswithwithFocussedIonBeam(FIB)surfacemodicationontherecordingsiteforreducedimpedance.Chapter 8 summarizestheresultsofthisworkandadvisesaplanforfuturemicroelectrodereliabilitystudy. 25

PAGE 26

CHAPTER2BACKGROUND 2.1Introduction Thischapterprovidesthebackgroundforneuralrecordingelectrodesandtheabioticfactorsaffectingthereliabilityofchronicneuralimplants.TheproblemofinterfacinganelectrodewiththebraintissueisdiscussedinSection 2.2 .Inthatsection,anoverviewoftheanatomyandthephysiologyofthebrainaregivenrst,anditisfollowedbyadiscussionofthephysicsoftheelectrode-tissueinterface.Thelastpartofthatsectionintroducesdifferentkindsofneuralrecordingelectrodesandthefactorsaffectingthereliabilityofneuralelectrodes.TheliteratureofBMIsystemsandmicroelectrodearraysaswellasthehistoryofexiblemicroelectrodearraydevelopmentatUFisreviewedinSection 2.3 .AdiscussionoftheabioticfactorsaffectingthereliabilityofchronicneuralimplantsisgiveninSection 2.4 .Therstpartofthatsectiondiscussesmicroelectrode-tissueimmuneresponse,andthesecondpartofthesectionprovidesaliteraturereviewoftheabioticfactorsaffectingchronicimplantreliability.FinallythechapterissummarizedinSection 2.5 2.2TheProblemofInterfacingtotheBrain Microelectrodearraysarebuiltusingmetaloralloymicrowiresandthindielectricorpolymerlms.Whentheyareimplantedintocorticaltissue,aninterfaceisestablishedbetweenthebiologicalmediumandthenon-biologicalforeignbody.Inordertobestevaluatethefailuremodesofachronicallyimplantedelectrodearray,itiscriticaltounderstandthesciencegoverningtheinterfaceincludingthemorphologyandbehaviorofthebiologicalcellsthatconstitutethecorticalenvironmentbothpriortoandpostimplant,thephysicsbehindthechargetransferbetweentheelectrode-electrolyteinterface,andthebiologyoftheelectrode-tissueimmuneresponseduringtheimplantedperiod.Thissectiondiscussesthesefundamentaltheoriesandmodelsandservesasabackgroundforthefollowingchapters. 26

PAGE 27

2.2.1AnatomyandPhysiologyofBrain Thecellsinthebrainandintheotherorgansofthenervoussystemcanbebroadlyclassiedintotwocategories:theneuronandtheneurogliaorglia.Theneuronornervecellisthefundamentalunitofelectricalactivityofthenervoussystem.Althoughafullygrownadulthumanbrainisestimatedtohave100billionneurons,theycontributetoonly25%ofthecellpopulationofthebrain[ 46 ].Neuroglialcellsorgliaarethesupportingcellularstructuresthatconstitutenearly75%ofthenervoustissue.Whiletheglialcellsdonotparticipateintheelectricalactivityofthebrain,theyplayavitalroleinprovidingamechanicalframeworkfortheneuronalcircuit,regulatingthebrainionambience,interfacingandmaintainingthereleaseofneurotransmitterduringsignaling,andcontrollingneuronalringrateandaidingintherecuperationoftissueinjury[ 46 48 ].Thethreetypesofglialcellsareastrocytes,oligodendrocytes,andmicroglialcells. Astrocytes,whicharespeciconlytothecentralnervoussystem(CNS),constitutenearly30-65%oftheCNSglialcellpopulation.Theyhavelongandbranchedoutextensionsthatgivethemastarorasteriskshapedappearance.Theirprimaryfunctionistosustainanoptimalchemicalenvironmentforneuralelectricalactivity.Intheirnormalstate,astrocyteshave8-10nmdiameterglialbrillaryacidprotein(GFAP)lamentsandroundnucleus[ 46 49 ].OligodendrocytesarealsoCNSexclusive,andtheyareresponsibleforformingamyelinsheatharoundtheneuronalaxons,whichdeterminesthespeedofthetransmissionoftheneuralactionpotentials.Microgliarepresentabout5-10%oftheglialcellpopulation,andtheyplaytheroleofscavengersinthebrain'swoundhealingresponse[ 46 49 ].Theyproducecell-signalingproteinmoleculescalledcytokines,whichactupontheinammationanddeterminesthesurvivalorthedeathofacell.Intheirrestingstate,microgliacellshaveahighlybranchedstructurewithnomacrophagereceptors. NeuronandActionPotential:Theprimarycellularstructuresofatypicalneuronincludethecellbodyorsoma,andtheextendedfeaturescalledtheneurites.The 27

PAGE 28

neuritesareclassiedintodendritesandaxon.Dendritesarehighlybranchedstructuresandtheyreceiveelectricalsignalsfromotherneurons,whereastheaxonisalongslenderstructurethatterminatesintomultiplebranchesandtransmitssignaltootherneurons.Axonisusuallyinsulatedwithadielectricsheathcalledmyelinsheathalongitsaxis.However,therearefewgapsthatremainuninsulatedandtheyarereferredtoasNodesofRanvier.TheRanviernodeshelpinpromotingtheactionpotentialasitistransmittedthroughtheaxon.Theaxonofoneneuronisconnectedtothedendriteofanotherneuronataspecializedsitecalledsynapsethroughwhichthesignalispropagated.Eachneuronisconnectedtoseveralotherneuronsthroughthesesynapses.Thenumberofsynapticconnectionsvariesfromoneneurontoanotherdependingontheirfunctionality,however,theyaretypicallyontheorderof10,000[ 50 ].AschematicillustrationofapyramidalneuronisshowninFigure 2-1 .Differentmechanismsarefollowedtotransmittheinformationfromoneneurontoanother.Thegeneralprocessingschemeatthesynapseinvolvestheuseofthereceivedsignaltoalterthestateofthereceivingneuron,andtotriggerthegenerationofanelectricpulse,calledtheactionpotentialinthereceivingneuron. Neuronalcellmembraneisembeddedwithspecialtypesofproteinscalledionchannels,thatformporesthroughwhichonlycertainionscanenterorleavethecell,therebymakingthecellmembraneselectivelypermeable.Commonionsinvolvedintheneuronalsignalingprocessaresodium(Na+)andpotassium(K+).Theionchannels,whichintheirrestingstatearecalledleakagechannels,canbeoperatedeitherbymodulatingtheassociatedneurotransmitters(suchasGlutamate,GABAorDopamine)orbymodulatingthemembranepotential.Theformerarereferredtoasneurotransmitter-gatedionchannelsandthelatterarereferredtoasvoltage-gatedionchannels.Therestingpotentialoftheneuron,whichisapproximately-65mV[ 50 ],isalteredbytheincomingneurotransmittersinitiatingthedepolarizationphaseofanactionpotential.Immediately,thevoltagedependentsodiumchannelsopenandallowaninow 28

PAGE 29

Figure2-1. Schematicrepresentationofaneuron. ofNa+ionsintothecelltobalancetheconcentrationgradient.ThisinuxofNa+ionswillraisethemembranepotentialclosertothesodiumrestingpotential(i.e.around+65mV).Roughlyabout1mslater,thesodiumchannelscloseduetoablockadebyaprotein,andavoltage-gatedpotassiumchannelopensup,enablingtheefuxofmoreK+ions.TheoutowofK+ionswillbringdownthemembranepotentialclosertothepotassiumrestingpotential(i.e.,around-80mV),leadingtoahyperpolarizationphase.Aftersometime,thepotentialdifferencebetweenthehyperpolarizationandtherestingpotentialsclosesthepotassiumchannelsandmovesthemembranepotentialsbackto 29

PAGE 30

Figure2-2. Schematicrepresentationofintracellularactionpotential. itsnormalrestingpotential.AnillustrationofatypicalactionpotentialwaveformisgiveninFigure 2-2 2.2.2Electrode-TissueInterfaceandInteraction Themomentanelectrodeisinsertedintothebrain,thebiologyofthetissuebeginstochangeinresponsetothenew'foreignbody'.Thetissueresponsebeginsasaninitialinsertion-inducedwoundresponseandcontinuestodevelopastheelectrodesremaininthebrainchronically.Thewoundhealingresponsecomprisesoffourphases[ 51 ]: 1. Hemostasis:Thisphasebeginsattheinstantofelectrodeinsertionandlastsforafewhours.Thesurgicalinsertionofthemicroelectrodecausesatraumaticbraininjurythatdamagesthetissueandbloodvesselstoalargeextent.Theruptureoftheblood-brainbarriercausesaninuxofblood-bornecellsandserumproteinsintothewoundsiteandformsabarrierclot[ 52 ].Hemostasisactivatesplateletswhichinturninitiatesthereleaseofcell-signalingcytokinemolecules[ 53 55 ].Thisleadstothenextstepofhealingnamely,inammation. 2. Inammation:Theinammationphasebeginsafewhoursaftertheinsertionoftheelectrodeandlastsforafewdays.Duringthisphase,themicrogliacells(ortheresidentmacrophagesofthebrain)areactivated[ 51 ].Activatedmicrogliacellshaveamoeboidstructureandupregulatetheproductionoflyticenzymesthatarecapableofdegradingcells[ 49 ].Microgliacellsalsoproduceavarietyof 30

PAGE 31

pro-inammatoryandneurotoxicchemicalsneartheinjurysiteincludingsomecytokinessuchasTGF-,MCP-1,IL-1,andIL-6.Axonalbreakdownoccursduringthisphase.Astrocytecellsalsobecomeactivatedduringthisphase.ReactiveastrocytesarecharacterizedbyupregulationofGFAPlaments[ 49 ],andtheyaidininterveningtheinammationbyproducingchemicalsthatarebothpro-inammatoryandanti-inammatory[ 51 ].Cytokinesproducedduringthisphasepromotethegrowthofbroblastcells(cellsthatcansynthesizeproteins)andtakethewoundhealingtothenextphase,therepairphase. 3. Repair:Thisphaseextendsforfewweeks.Gliosisortheproliferationofastrocytesbecomesprominentduringthisphase,andthisbeginstheformationoftheglialscar.Continuousproliferationofreactiveastrocytesresultsintheproductionofinhibitorymoleculesandhencepreventstheregenerationofdamagedaxons.Itiscomprehendedthatthepurposeoftheglialscaristoprotectneuronfunctionalitybyrestoringthedamagedblood-brainbarrierandminimizingtheinammatoryresponse[ 56 ]andalsotoavoidneuronsfromformingincompatiblenetworkconnectionspostinjury[ 57 ].Theextentoftheglialscarisassociatedwiththedegreeofdamagetotheblood-brainbarrierandthesurroundingtissue[ 56 ]. 4. Remodeling:Whilethewoundhealinginnon-CNStissuerepairsandremodelstheinjuredtissue,CNSwoundhealingdoesnotinvolvetherepairandtheremodelingofinjuredneurons.Duringthisphase,whichlastsformonths,thephenomenonoffrustratedphagocytosisisnoted.Itisthestatewherethereactivemacrophagesandmicrogliacellscontinuetosecretedegradativechemicalsasapersistentefforttodecomposetheelectrode.Thisfurtherpromotestheproliferationofreactiveastrocytesandtheglialscarbecomesdenseraroundthewound. AschematicrepresentationofthedifferentphasesoftissuereactionwithimplantedmicroelectrodeisshowninFigure 2-3 .Asfortheabioticmodications,theelectrodeexperiencesphysicalchangesinitsrecordingsiteandinsulation.Ithasbeenobservedthatthecommonorganicpolymerinsulationmaterialssuchaspolyimideandparylene-Creactwiththeaqueousenvironmentandsufferfromsurfacedelamination.Similarly,themetalsusedfortherecordingsitecorrodeinthesalineenvironmentcreatedbythecerebro-spinaluid(CSF).Differentmetalshavedifferentcorrosionrates.Ithasbeenestimatedthatatungstenmicrowirehasacorrosionrateof300-700mperyearinphosphatebufferedsaline(PBS)environmentandacorrosionrateof100mperyearinin-vivoenvironment[ 58 ],whileplatinumorplatinum-iridiumwireshavereducedornegligiblecorrosionrate. 31

PAGE 32

Figure2-3. Schematicrepresentationofdifferentphasesofelectrode-tissuereaction. 2.2.3IntroductiontoNeuralElectrodesandFactorsAffectingReliabilityofNeuralElectrodes Thisresearchisfocusedonsingleneuronextra-cellularactionpotentialrecordingmicroelectrodes.However,inordertounderstandtheprosandconsofthisrecordingmechanism,itisnecessarytooverviewthedifferentrecordingmethodsusedtocaptureelectrophysiologicalsignalsfromthebrain.Theotherrecordingmechanismsthatarecommonlyusedare:electroencephalography(EEG),electrocorticography(ECoG),andlocaleldpotentials(LFPs)[ 59 ].Allofthesetypesofelectrodesrecordextracellularneuralsignalsfromthecortexwithdifferentspatialandtemporalresolutionsaswellasdifferentlevelsofinvasiveness.TheEEGelectrodesarethenon-invasiveelectrodesandareplacedonthescalpwhichis2-3cmawayfromthesurfaceofthecortex.Becauseofitsnon-invasiveness,EEGelectrodeshavebeenwidelyusedinhumanBCIstudies.However,theseelectrodesoperateinaspatialdomainof3cmandarelimitedtoafrequencyof70Hz[ 59 ].ECoGelectrodesareinvasiveandareplacedonthesurface 32

PAGE 33

ofthecortex.Thespatialaveragingofneuralactivityoftheseelectrodesisoverasmallregionof0.5cm,andtheoperatingfrequencyrangeishigherthanthatofEEG.WhilethetypicalbandwidthofECoGelectrodesisbetween10Hzand200Hz[ 60 ],arecentstudyhasreportedtherecordingofmeaningfulsignalsfromECoGelectrodesupto6KHz[ 61 ].Microelectrodearraysthatareimplantedintothecortexcanbeusedtorecordsingleunitactivityfromtheneuronsbetweenthefrequenciesof500Hzto6KHz.However,thesameelectrodearrayscanbeusedtorecordlowfrequencysignals,lessthan250Hzcalledlocaleldpotentials(LFPs).Whilesingleunitarrayshaveaspatialresolutionof0.2mm,LFParraysextendupto1mm[ 59 ].Sincethemicroelectrodearrayshavethesmallestspatialaveragingandcanrecordsignalsfromdiscreteneurons,severalBMIapplicationsusesingleunitintracorticalmicroelectrodearraysforobtaininghighresolutionelectrophysiologicalsignals.Thisresearchalsoconcentratesontheperformanceandfailuremodesoftheintracorticalmicroelectrodearraysinchronicapplications. Asseenfromsection 1.1.3 ,severalbiologicalandnon-biologicalfactorscontributetothefailureofmicroelectrodearraysinchronicconditions.Thedifferentsourcesofbiologicalfactorscontributingtothefailureofreliableneuralrecordingwereintroducedinsection 2.2.2 .Table 2-1 liststhedifferentbioticandabioticfactorsconnectedwiththefailureofreliablechronicrecordinginmicroelectrodearrays.Theabioticfactorswillbediscussedindetailinsection 2.4 Table2-1. Bioticandabioticfactorsassociatedwiththefailureofchronicrecordingcapabilitiesofmicroelectrodearrays. BioticfactorsAbioticfactors InammatoryresponseVariationsinelectrodegeometryConstantblood-brainbarrier(BBB)disruptionStrainduetobrainmicromotionChangesintheneuralsynapseInsulationdelaminationandcrackingAstrogliosisinitiationRecordingsitecorrosionRecruitmentofmacrophagesandmicroglialcellsVariationsinelectrodematerials 33

PAGE 34

2.3LiteratureReviewofBMISystemsandMicroelectrodeArrays ThissectionprovidesasummaryofdifferentBrain-MachineInterfacesystemsandrecordingmicroelectrodearraysdesignedanddevelopedfromthe70'sand80'stothepresent.Alsoabriefdiscussionofthedifferentgenerationsofexiblerecordingmicroelectrodesdesigned,fabricatedandcharacterizedattheUniversityofFloridaisgivenattheendofthissection. 2.3.1LiteratureReviewofBMISystems Itisseeninsubsection 1.1.1 thatthehistoryofBrain-MachineInterfacesystemdevelopmentdatesbackto1980[ 29 ],andsincethenseveralstudieshavebeenconductedonrat,monkey,andhumanmodels[ 30 36 ].ThissubsectionprovidesasummaryofdifferentBMIsystemsdevelopedinthelastfewyearsandhighlightsthepreclinicalandclinicalstudiesconductedusingthosesystems. SomeoftheimportantBMIsystemshavebeendevelopedandtestedbytheresearchgroupsattheUniversityofMichigan,UniversityofUtah,StanfordUniversity,BrownUniversity,DukeUniversity,GeorgiaInstituteofTechnology,andJohnHopkinsUniversity.ThegroupattheUniversityofMichiganhasdevelopeda64channelwirelessimplantablemicrosystemforchronicneuralrecordingapplications[ 41 62 ].Long-establishedMichigansiliconarrayswereusedastherecordingprobes.A60dbgainamplierwasusedtoamplifythesignalsoverabandwidthof100Hz-10KHz.TheampliedsignalsweretransmittedwirelesslythroughaloopantennaandthesystemwaspoweredbyaRFinductivecoil.Componentswerewirebondedtoasiliconplatform.TheHermessystemdevelopedbythegroupatStanfordUniversityisaLi-ionbatteryoperatedwirelesslowpowerneuralrecordingsystem.ThegrouphasreportedfourversionsoftheHermessystem-HermesB[ 63 ],HermesC[ 64 ],HermesD[ 42 ],andHermesE[ 65 ].Thoughtheversionshadminorvariationsintheircomponentdesign,theyallusedtheUtahsiliconelectrodearrayforrecordingsignals,andlithiumionbatteryforpoweringthesystem.Thewirelesstransmissionscheme 34

PAGE 35

differedacrossthedifferentversions.AllthecomponentsweresolderedtoaPCB.TheHermessystemsweretestedinprimates. ResearchersatBrownUniversityhavereporteddifferentversionsoftheirimplantablemicrosystemsforBMIapplications.Oneoftheirearlierversions[ 43 66 ]includeda16-channelUtaharrayforrecordingsignalsandahybridRF/IR(VerticalCavitySurfaceEmittingLaser-VCSELdiode)wirelesstelemetryfordatatransfer.ThesystemwaspoweredbyaRFcoil.Componentswereipchipbondedtoasiliconsubstrate.Thesystemwastestedinrats.Theirnextversion[ 67 ]includeda100channelUtah(Blackrock)electrodearraywithVCSELdiodefordatatransferandRFcoilforpoweringthesystem.Thesystemwastestedinpigsandmonkeys.TheresearchgroupsatBrownUniversityhavealsosuccessfullyconductedsomeclinicaltrialsoftheirBrainGateneuralinterfacesystemsonhumans.In2011,theyreportedsuccessfuldemonstrationofneuralcontrolofcursortrajectorybyatetraplegichumanpatientimplantedwiththeirBMIsystem[ 38 ].In2012,theydemonstratedneurallycontrolledroboticarmmovementforreachingandgraspingobjectsonhumanwithtetraplegiaimplantedwiththeirsystem[ 39 ]. DukeUniversityresearchersreportedthesuccessfuldemonstrationofneuralrecordingusingtheirimplantable96channeldataacquisitionsysteminasheep[ 44 ].Theyused50mdiametertungstenmicrowiresinsulatedwithpolyimidethatwashandassembledandconnectedtoanOmneticsconnectortorealizetheirmicroelectrodearray.Datawaswirelesslytransmittedviaastainlesssteeldipoleantenna,andthesystemwaspoweredbyRFcoil.AgroupatUCLAreportedtheirneuralrecordingsystemin2002[ 68 ],whichincludedelectrodesmadeoftungstenlamentandpoweredbyaRFinductivecoil.AresearchgroupatGeorgiaTechreportedthedesignandfunctionalityoftheir32-channelneuralrecordingsystem[ 69 70 ].TheirdesignincludedaSMDinductorasthetransmitterantennaandbatteryorinductivecoilforpoweringthesystem.AresearchgroupatJohnHopkinsUniversityrecentlyreportedthe 35

PAGE 36

demonstrationofsingleunitrecordingusingtheirwirelessrecordingsystem[ 71 ].Thesystemincludedacommercial16channelelectrodearray,warp16array(Neuralynx),integratedwithachipantennaasawirelesstransmitterandarechargeablebatteryasthepowersupply.Theytestedtheirsysteminprimates.Table 2-2 givesanoverviewmatrixofdifferentBMIsystems. Table2-2. OverviewmatrixofdifferentBMIsystems. AfliationRecordingelectrodesPowersourceWirelessscheme/AntennaElectronicsinte-grationscheme UniversityofMichigan[ 41 62 ]MichigansiliconelectrodesRFinductivecoilLoopantennaWirebondedtosiliconplatform StanfordUniver-sity[ 42 63 65 ]Utah/BlackrocksiliconelectrodearrayLithiumionbatteryMicrostrippatchantenna(Her-mesDandE);Stubantenna(HermesC)SolderedtoPCB BrownUniver-sity[ 43 66 67 ]Utah/BlackrocksiliconelectrodearrayRFinductivecoilOpticalteleme-tryusingaVCSELdiode;RF/IRtelemetryFlipchipbondedwithepoxyonsilicon DukeUniversity[ 44 ]Tungstenmi-crowirearrayRFinductivecoilDipoleantennaofastainlesssteelwireFlipchipbond-ing UniversityofCaliforniaatLosAngeles[ 68 ]Tungstenla-mentelectrodeRFinductivecoilSolderedtoPCB GeorgiaInsti-tuteofTechnol-ogy[ 69 70 ]Battery/Induc-tivecoilSMDinductorasatransmitterantennaSolderedtoPCB;Flipchipbonding JohnHopkinsUniversity[ 71 ]Warp16elec-trodearray(Neuralynx)RechargeablebatteryChipantenna 2.3.2LiteratureReviewofMicroelectrodeArrays Thissubsectionreviewsthedifferenttypesofsinglemicrowireelectrodes,microwirearrays,siliconmicromachinedarrays,andpolymermicromachinedelectrodearrays. 36

PAGE 37

SingleMicrowireElectrodes:Effortsforbuildingsinglewireelectrodesbeganasearlyas1950whenGrundfestet.al.reportedthefabricationofstainlesssteelmicro-needleelectrodesusingelectrolyticetching[ 72 ].TheearliestreportonsinglemicrowireelectrodeforsingleunitrecordingwaspublishedbyHubelin1957[ 73 ].Tungstenmicrowiresof125mdiameterinsulatedwithlacquerandsharpenedatthetipto0.4mwereimplantedintocatcerebralcortexandextracellularspikeswererecorded.Followingthis,afewreportswerepublishedonthefabricationofglasscoatedplatinummicrowireelectrodesforelectrophysiologicalrecording[ 74 75 ].SalcmanandBakreportedsuccessfulchronicrecordingofsingleunitactivityusingglassinsulatedplatinummicrowireelectrodes[ 76 ]andparylenecoatedgold-iridiummicrowireelectrodes[ 77 ]obtainedfromthevisualcortexofcat.Theirdesignalsopermittedthemicrowireelectrodestobeconnectedtoanultraexiblemetalliccablethatprovidedreliefforstrainduetobrainmicromotion.However,thelimitationofasinglewiremicroelectrodetoobtainelectrophysiologicalsignalsfromalargeneuronensemblepreventedfurtherinterestinthisdesign. MicrowireElectrodeArrays:Williamset.al.publishedareportin1999[ 78 ]reportingtheprocedureofassemblingmicrowireelectrodearraysusingdiscretetungstenmicrowires.Acustommadejigwasusedasaplatformtoholdthreerowsofequallyspaced35mdiametertungstenmicrowirecoatedwith7mthickpolyimideinsulation.Eachrowhad11microwires.Thespacingbetweentherowswas400mandbetweenthewireswas250m.Commercialelectrodearrayssimilartothisdesignapproacharecurrentlybeingmanufacturedbytwocompanies-Tucker-DavisTechnologiesInc.andMicroProbesInc. InTucker-DavisTechnologiesmicrowireelectrodearrays[ 79 ],polyimideinsulatedtungstenmicrowiresaretherecordingprobes.Thesizeandnumberoftheprobesarecustomizable.ThelasercutdiskmicrowiresareconnectedtoaprintedcircuitboardwhichisattachedtoastandardOmneticsorZIFconnector.MicroProbesInc.microwire 37

PAGE 38

electrodearrays[ 80 ]includeneedleshapedplatinum-iridiummicrowiresasthecoreconductorandparylene-Castheprobeinsulationmaterial.Thesize,positioningandthenumberofprobesiscustomizable.Theyalsofabricateverticallypositioned2Darrays.ElectrodearraysdesignedbyWilliamset.al.andrepresentativesofTDTandMicroProbearraysareshowninFigure 2-4 .Oneofthedisadvantagesofthemicrowirearraysisthetimeconsumingandlaborioushandassemblyofindividualwirestotheelectrodearray. A B C D E Figure2-4. Photographsofdifferenttypesofmicrowireelectrodearrays.A)MicrowireelectrodearraydevelopedbyWilliamset.al.,[ 78 ]c1999Elsevier,B)Tucker-DavisTechnologiesZIFconnectormicrowirearray[ 79 ]cTucker-DavisTechnologiesInc.,C)Tucker-DavisTechnologiesOmneticsconnectormicrowirearray[ 79 ]cTucker-DavisTechnologiesInc.,D)MicroProbesInc.1Dmicroelectrodearray[ 80 ]cMicroProbesInc.,E)MicroProbesInc.2Doatingmicroelectrodearray[ 80 ]cMicroProbesInc. 38

PAGE 39

SiliconMicromachinedElectrodeArrays:Electrodearraysfabricatedusingsiliconmicromachiningtechniqueshavetheadvantagesofreducingthepainstakinghandassemblingstepofdistinctmicrowiresintoanarraylayoutandprovidingmoreaccuratecontrolofelectrodegeometry.Alsosiliconcanserveasanexcellentplatformforintegrationwithassociatedelectronicinterfacesforrealizingafully-implantableself-sustainedsystem.TheUniversityofMichiganandtheUniversityofUtaharethepioneersindevelopingsiliconmicromachinedelectrodearrays.Thenextfewparagraphswillbrieydescribedifferenteffortstakenbytheresearchgroupsattheseuniversitiesindevelopingsiliconmicroelectrodesandwillalsoprovideasummaryofotherdesignsdevelopedbyotherresearchgroups. TherstdesignoftheMichiganprobeswaspublishedin1970[ 81 ].TheelectrodewasaplanararrayoftaperedsiliconbeamsmicromachinedusingMEMStechnology.Goldlinesactedasconductivetraces,whileSiO2servedastheinsulator.Areviseddesignwaspublishedin1985[ 82 ],withsomechangesingeometryandfabricationprocess.Designchangesincludedsinglesiliconshankwithgoldrecordingsitesonthetopsurface.Fabricationchangesincludedborondopingforbettergeometrydenition.Si3N4wasincludedasaninsulator.Thethirdandfourthversionsincludedfunctionalelectroniccomponenetssuchasamplier,multiplexerandbuffer[ 83 84 ].Astraightsiliconribboncablewasincludedinthenextversiontointerfacetheelectrodewiththeback-endelectronics[ 85 ].In1994,thedesignprogressedfrom2Dto3D[ 86 ].Thedesignbondedmultiple2Dsiliconarraysperpendiculartoasiliconbaseandintegratedwithasiliconcable.Aphotographofthe3DMichiganelectrodearrayisshowninFigure 2-5 A.NeuroNexusisaUniversityofMichiganspin-offcompanythatcommerciallymanufacturestheMichiganelectrodearrays. TheUtahelectrodearraywasrstreportedin1991[ 88 ].Thefabricationincludedmicromachiningofa1.7mmthickSiwaferintoathreedimensional10x10arrayof1.5mmthickcolumnsseparatedfromeachotherby400m.Thecolumnshadatapered 39

PAGE 40

A B Figure2-5. Photographsofsiliconmicromachinedelectrodearrays.A)3DMichiganmicroelectrodearray[ 87 ]c2004IEEEandB)UtahMicroelectrodearray[ 45 ]c1996Elsevier. endgivingabedofnailsappearancetotheelectrodearray.ThenextversionofthedesignincludedcoatingthetipwithPt/Ti/W/Ptmetalsandpolyimideinsulationalongtheshank[ 89 ].Amoreadvanceddesignincludedagradualvariationinthelengthoftheneedles,givingaslantappearancetothearray[ 90 ].AphotographoftheUtahelectrodearrayisshowninFigure 2-5 B.CyberkineticsisaUniversityofUtahspin-offcompanythatcommerciallymanufacturestheUtahmiroelectrodearrays.MichiganandUtaharraysaremostcommonlyusedinmanyBMIsystems. OthersiliconelectrodearraysincludeadesignfromagroupinSweden[ 91 92 ]andatArizonaStateUniversity[ 93 ].TheSwedenelectrodedesignwassimilartotheMichiganarray.Thedifferencewasthesubstrate,i.e.,SOIwasusedasthesubstrateinsteadofSi.TheelectrodesdevelopedbyArizonaStateUniversityusedtheSUMMiTVprocessonpolysilion. PolymerMicromachinedElectrodeArrays:Sincepolymershaveamodulusofelasticitymuchlessthanrigidsiliconormetalwires,andbettermatchtheextremelylowmodulusbraintissue,researchersstartedshowinginterestinbuildingpolymerbasedmicroelectrodearrays.AresearchgroupatArizonaStateUniversitydevelopedapolyimidebasedmicroelectrodearrayafewyearsback[ 94 ].Laterparylene-Cbased 40

PAGE 41

electrodeswithmicrouidicsfordrugdelivery[ 95 ]andopenarchitectureformitigatedtissueresponse[ 96 ]weredevelopedandreported.Severalelectrodedesignshavebeenreportedusingpolyimideandparylene-Casthestructuralmaterial[ 97 99 ].Onelimitationforthisdesignistheextremelylowstiffnessofpolyimideorparylene-Cwhichcanresultinbucklingduringimplantation.AgroupinTokyohasreportedbio-degradablepolymer(PEG)coatedexibleelectrodestoreduceinsertionchallenges[ 100 ]. 2.3.3UFFlexibleMicroelectrodeArrays TheUniversityofFloridaexiblemicroelectrodearrayseffortsincludethedesign,fabrication,andcharacterizationoffourgenerationofmicroelectrodes.Generation1arrays,developedandtestedin2006[ 2 ],includedelectroplatednickelastherigidprobematerialandpolyimideastheexiblecablematerial.Theprobeshankswereinsulatedwithparylene-C,andtheexposednickelattherecordingsiteswereplatedwithgoldforimprovedbiocompatibility.AstandardOmneticsconnectorwasusedtointerfacetheelectrodearraywiththeexternalsystem.Aphotographofthegeneration1microelectrodearrayisshowninFigure 2-6 .Afabricatedprototypeofthearraywasimplantedintoarat,andelectrophysiologicalrecordingsweremadeduringsurgery.HoweverthefailureofthebackendOmneticsconnectionpreventedfurtherrecordingofsignalsaftersurgery.Animportantobservationattheendofin-vivotestingwasthesignicantcorrosionoftherecordingsite.Hence,inthefuturedesigns,thecombinationofnickelandgoldwasavoided,andtheprocessofdepositingtheprobemetalusingelectroplatingwasreplacedwithhybridassembly. Generation2microelectrodearraysusedpre-fabricatedtungstenmicrowiresasrecordingelectrodesandpolyimideastheexiblecable[ 3 ].Goldwasusedfordeningtheconductivetraces.Thetungstenmicrowireswerehandassembledtothemicrofabricatedpolyimidesubstrate.Inaddition,thisgenerationelectrodesalsoincludedtwostainlesssteelnutsthatservedasfastenersaftersurgicalimplantation.Similartogeneration1,anOmneticsconnectorwasusedtoconnecttheelectrodes 41

PAGE 42

Figure2-6. UFgeneration1microelectrodearray[ 2 ]c2006IEEE. Figure2-7. UFgeneration2microelectrodearray[ 3 ]c2008IEEE. withtheexternalsystem.Aphotographofthegeneration2microelectrodearrayisshowninFigure 2-7 .Fabricateddevicesweresubjectedtoacutein-vivotesting,andelectrophysiologicalmeasurementswererecordedsuccessfully.Thoughhybridpackagingofmicrowireswasabigfabricationchallenge,successfulin-vivorecordingusingthegeneration2electrodeguidedthewayforgeneration3activemicroelectrodearrays. 42

PAGE 43

Figure2-8. UFgeneration2amplierintegratedmicroelectrodearray[ 4 ]c2010IEEE. UFgeneration3microelectrodearraysincludedalowpowerpre-amplierintegratedtotheelectrodesubstrate[ 4 ].Theelectrodesubstratematerialwaspolyimide,withgoldinterconnectlines,andtherecordingmicrowireswerepre-fabricatedtungstenwiresof50mdiametersimilartogeneration2electrodes.Thegeometryoftheelectrodesubstratewasmodiedforbetterpositioningontheratskull.Insteadoftwonuts,onlyonexednutwasadded,andaexibleslotwasincludedforthesurgeon'simprovedplacementoptions.AnOmneticsconnectorwasusedtointerfacewiththeexternalsystem.Theamplierdiewasattachedtothepolyimidesubstrateusingaip-chipbondingprocess.Photographofgeneration3microelectrodearrayisshowninFigure 2-8 .Thefabricatedamplier-microelectrodesystemwassubjectedtochronicin-vivotesting,andampliedsingleunitneuralactionpotentialsweresuccessfullyrecordedforover42days.After42days,thedevicefailedduetodelaminationofpolyimideneartheconnectorandsubsequentshortingofconductivetraces.Also,post-implantSEMimagesrevealedcorrosionofthetungstenrecordingsite.Lessonslearnedduringthefabricationofgeneration3devicesservedastheguidelineforgeneration4pyrexsupportedamplier-microelectrodesystem. 43

PAGE 44

Integrationofamplierdieontotheexiblepolyimidesubstratewasamajorchallengeduringthefabricationandassemblyofgeneration3devices.Theexiblesubstratewasnotperfectlyattoallowuniformattachmentofthediewiththebondpads.Whengeneration4deviceswerefabricated,apyrexglasspiecewasattachedtoimprovetherigidityandtheatnessofthepartofthesubstratethatholdstheamplierdie.Thecablepartofthesubstratestillremainedexible.ItshouldbenotedthatallfourgenerationsofUFmicroelectrodesweredevelopedasmonolithicsubstrates.Thiskindofsinglesubstratedesignhaditslimitationsinfabricationandoftenresultedinlowyield.Inordertorectifytheissuesamodularapproachwasapproachedforlatermicroelectrodedesigns.DesignandfabricationofmodularelectrodeswillbedescribedinChapter 3 2.4AbioticFactorsAffectingReliabilityofMicroelectrodeArraysinChronicNeuralImplants Thissectionhighlightsvariousstudiesconductedontheimpactofvariationindifferentabioticfactorsontissueresponseandchronicsignalrecordingqualityoftheelectrode.Therstsubsectionprovidesaliteraturereviewofmicromotioninducedstraineffects,andthesecondsectionsummarizesdifferentstudiesconductedontheeffectofvariationsofprobegeometry,size,insulation,andsurfacemorphologyonrecordingperformance. 2.4.1LiteratureReviewoftheReportsPublishedonMicromotionInducedStrainEffects Strainduetomicromotionisoneofthesignicantcontributorsoffailuremechanismsforlong-termneuralimplants.StudiesconductedbyBiranet.al.[ 101 ]andKimet.al.[ 102 ]reportthatthestraininducedimmuneresponsecausedbytherigidtetheringoftheelectrodetotheskullshowedanincreaseinmicroglialactivityintheimplantedtissueascomparedtountetheredelectrodes.Thisincreasedtissueresponseandcontinuousproliferationofmicroglialcellsaroundtheelectrodecanbedetrimentaltothewellbeingoftheneuronsinthevicinity.HistologicalstudiesconductedbyBiranet.al.,[ 103 ]report 44

PAGE 45

thatupregulationofmicroglialbiomarkerED1wasaccompaniedbyreductioninneuronsandnervecellberssurroundingtheimplant.Thissuggestsacorrelationbetweenincreasedtissueresponseandreducedsignalreliability.Thephysicsofstrainduetomicromotionisbrieyexplainednext. Whentheelectrodesubstrateissecuredtotheskullduringimplantation,itresultsinarigidtetheringofoneendoftheelectrode,whiletheotherendoftheelectrode,thetip,isfreetorotateinthecortex.Thebrainalwaysexperiencesdisplacementontheorderoffewmicronsdrivenbyphysiological,behavioralandmechanicalsources.Ithasbeenobservedthatthebrainmicromotioninanesthetizedratsduetorespiratorypulsationisontheorderof10-30m,andduetovascularpulsationisabout2-4m[ 104 ].Themicromotionofthebrainwithrespecttotheskull(relativemicromotion)canexertanaxialforceonthecorticaltissue.Thestrainduetotheforceactingontherigidback-endistransferredalongtheprobeshankanddisplacestheelectrodetipwithinthebraintissue,promotingmoreupregulationofmicroglialcells. Anefforttoreducefront-endstrainwasrstinitiatedbySalcmanandBakin1973[ 76 ]whentheyproposedtheirsinglewireelectrodedesign,whichincludedalongslendergoldwireasacablebetweentheprobeandtheexternalconnection.Theirdesignlimitedtheelectrodedisplacementto10mfor1mmrelativemicromotionofthebrain.Howeverasmicrowireelectrodearraysbecamemoresuitableforneuralrecordingthansinglewireelectrodes,thisdesignapproachbecameunfeasible. AfewnumericalstudiesdoneinthepastshowthatanelectrodewithlowYoung'smodulusmaterialorredenedgeometryforhighcompliancecanprovidemuchneededfront-endstrainrelief.MechanicalmodelingoftetheringinducedstrainforsiliconandpolymerelectrodesdonebyLeeet.al.showedthatarigidlytetheredsiliconshanktransferssignicantstraintosurroundingbraintissueandfavorsdisplacementoftip[ 105 ].Finiteelementanalysisofapolyimidearraywithrespecttoarigidsilicon 45

PAGE 46

microelectrodearrayofsamedimensionsconductedbySubbaroyanet.al.showedafront-endstrainreliefof65-94%[ 106 107 ]. Intermsofdesignmodicationforfront-endstrainrelief,theserpentineshapedsiliconcabledevelopedbytheUniversityofMichiganshowed50%lessstressthanastraightsiliconcableofthesamedimension[ 108 ].Morerecently,neuralprobesandcableswithdifferentmeanderinggeometriesarebeingreportedwithimprovedcomplianceandbetterstressrelief[ 109 110 ]. 2.4.2LiteratureReviewofReportsPublishedonEffectsofElectrodeGeometry,Size,Insulation,andSurfaceMorphologyVariationsonChronicSignalQuality Studieshavebeenconductedinthepasttounderstandtheeffectofelectrodesizeandgeometryvariationsonthechronicperformanceoftheimplant.Seymouret.al.investigatedtheinuenceofelectrodegeometryandsizeonchronictissueresponse[ 96 ].Theyproposedanopenarchitecturedesignfortheprobeshankinsteadoftheconventionalsolidstructure.Theirdesignincludedalattice-likestructurefortherecordingsitewithsub-cellularfeaturesizes.Designswithdifferentsupportingarmthicknesses(4m,10mand30m)werefabricatedandtestedin-vivoalongwithasolidshank.After4weeksofimplantation,theprobesweretestedforimmunohistology.Itwasobservedthatthethinnerfeaturesoftheopenarchitecturedesignhadaboutone-thirdhigherneuronalattachmentandreducedmicrogilalencapsulationwhencomparedtothesolidshank.RecentlyThelinet.al.studiedthecombinedandtheindependenteffectofimplantsizeandxationmodeonchronictissueresponse[ 111 ].Stainlesssteelwiresofsizes50mand200mwereimplantedintoratcortexfor6-12weeks.Someofthemweretetheredwhilesomeremaineduntethered.Post-implantimmunostainingindicatedincreasedglialresponseonlargerdiameterimplantascomparedtosmallersizeimplants.Similarly,tetheredimplantsshowedhigherimmuneresponsethanuntethered. 46

PAGE 47

Comparativelylittlestudyhasbeenconductedontheinuenceofinsulationandsurfacemodicationvariationsonchronicelectroderecordingperformance.AstudyconductedbyWardet.al.[ 112 ]comparedthein-vivobehaviorofdifferentmicroelectrodearrayswithvariedprobegeometries,recordingsitematerial,andinsulationmaterial.Severalcommerciallyobtainedelectrodeswereimplantedinanimalmodelsfor31days,andmetricsincludingimpedance,signaltonoiseratio,andrecordingstabilitywereevaluatedandanalyzed.Itwasobservedthatthemean1KHzimpedancevariedsignicantlyacrossalltheimplantedelectrodearrays,whereastheSNRdidnotshowmuchvariation.Sincemultiplevariablesareinvolvedinthisstudy,itisnotstraightforwardtoidentifyonespecicmechanismforthevariation. AdetailedstudyontheeffectofinsulationvariationandrecordingsitesurfacevariationonelectrodeimpedanceisprovidedinChapters 4 and 5 2.5Summary NeuralimplantsandBrain-MachineInterfacesaretrulyresultsofmulti-disciplinaryefforts.Tohaveacompletecomprehensionoftheissuesaffectingthereliabilityofneuralelectrodesinchronicapplicationsandtodevelopbetterelectrodesthatcanpotentiallyfunctionwithextendeddurability,itbecomesessentialtounderstandthefundamentalsfromdifferentperspectives.Thischapteraimsatprovidingthatinformationbypresentingareviewofthenecessarybiologicalbackground,asummaryofthecurrentstateoftheartinBMIsystemsandneuralelectrodes,andbydiscussingvariousstudiesconductedontheabioticfactorsinuencingelectrodeperformance.Otherbackgroundinformationrelatedtoelectrode-electrolyteimpedance,andinsulationandrecordingsitesurfacemodicationswillbereviewedinChapter 4 47

PAGE 48

CHAPTER3HIGHLYCOMPLIANTMODULAR2-DMICROELECTRODEARRAYSFORFRONT-ENDSTRAINRELIEF 3.1Introduction Thischapterdescribesthedesign,fabrication,andtestingofhighlycompliantserpentineshapedpolyimidemodularelectrodesdevelopedattheUniversityofFlorida.ModularelectrodesarethefourthgenerationpolyimideelectrodesdevelopedattheUniversityofFloridaafterthegeneration1exiblepolymersubstrateelectrodes[ 3 ],thegeneration2amplierintegratedmicroelectrodearrays[ 4 ],andthegeneration3pyrexsupportedamplierintegratedmicroelectrodearrays.Theseelectrodeshavetwomodules,(i)arigidsiliconplatformservingasastagefortheconnectorandfutureelectronics,and(ii)aserpentineshapedexiblepolyimidecablethatinterfacestherecordingtungstenmicrowirearrayandtherigidmodule.Itisdemonstratedthattheserpentineshapedpolyimidecableismorecompliantthanastraightpolyimidecableandthusprovidebetterfront-endstrainrelief. Thischapterisorganizedasfollows.Section 3.2 describesthedesignofthecableandrigidmodulesanddiscussesthereasoningbehindthedesign.Section 3.3 explainstheanalyticalandthenumericalmodelingofthecablecompliance.Section 3.4 describesthefabricationmethodemployedforbuildingthecablesandtherigidmodules.Section 3.5 describestheexperimentalmethodusedforquantitativelymeasuringthecomplianceofthecables.Section 3.6 summarizestheexperimentresultsandcomparestheresultswithrespecttotheanalyticalandnumericalmodels. 3.2FlexibleCableandRigidModuleDesign Themodularelectrodesincludetwomodules,arigidsiliconmoduleservingasaplatformfortheelectronicsandconnectorandaexiblepolyimidemoduleservingasthefront-endcablebetweenthemicrowiresandtheelectronics.Themodulesarefabricatedindependentlyandnallybondedtogetherusingconductivesilverepoxypaste,andsecuredandhermeticallysealedwithunderllepoxy.Thiskindofmodularapproach 48

PAGE 49

Figure3-1. Illustrationofthemodularelectrodedesign. allowsparallelfabricationofthemodules,therebyreducingtheprocessingtimeandincreasingtheyield.Figure 3-1 showstheconceptualdrawingofthemodularelectrodedesign. TheexiblepolyimidecablemoduleisdesignedtohaveaserpentinestructureasshowninFigure 3-1 .Themeandersinthenewdesignresultsinanincreaseintheeffectivelengthandadecreaseintheeffectivewidthcomparedtoastraightelectrodeofthesameoverallsize,therebyprovidinghighercomplianceandbetterstrainreliefthanitsstraightcounterpart.Atthesametime,theoverallformfactorofthecableisstillmaintainedthesametofacilitateimplantation.Furthermore,thenewgeometryenablesplacingtherecordingmicrowiresinatwodimensionaltransversefashion,with9electrodesbeingplacedina3x3array. 3.3AnalyticalandNumericalModelingofCableCompliance Thissectionprovidesadetaileddescriptionoftheanalyticalandnumericalanalysisusedtocalculatethecablecompliance.TheassumptionsandboundaryconditionsconsideredforthetheoreticalcalculationandtheABAQUSsimulationareexplained.Also,adetailedanalysisofthetheoreticalandmodelinguncertaintyisprovided. 49

PAGE 50

3.3.1AnalyticalAnalysis Afterimplantation,oneendoftheelectrodecableisattachedrmlytotheskullresultinginarigidtethering,whiletheotherendcontainingtherecordingmicrowireisinsertedintothecortex,asshowninFigure 3-2 A.Thenaturalrespirationandvascularpulsationalongwithheadmovementoftheanimalresultinthemicromotionofthebrainwithrespecttotheskull.Thebrainexhibitsmotioninallthethreeaxes.ThetranslationalmovementofbrainwithrespecttotheskullintheXandYaxesexertsaforcebothradialandtangentialtotheelectrodecable[ 106 ].Inaddition,thetranslationalmovementofthebrainintheaxisorthogonaltoplaneoftheelectrode(Z-axis)contributestotheshearingofthecable.However,forthisanalysis,threeassumptionsweremade: 1. therotationalmotionofthebrainandhencethetorsionoftheelectrodeisnotconsidered, 2. theforceactingontheelectrodeisassumedtobeapointloadconcentratedatthecortexendoftheelectrodecable,and 3. thecableisxedintheverticaldirectionatthesurfaceoftheskull. Guidedbytheseboundaryconditionsandassumptions,abeammodelwasdevelopedfortheelectrodecable.Accordingtothebeammodel,thecableisconsideredasaclamped-guidedbeam,whichisxedattheskullandfreetomoveinX,YandZ-axesalongthecortex,andaconcentratedforceisactingattheguidedend.Figures 3-2 Band 3-2 Cshowthefreebodydiagramsofthestraightcableandserpentinecablerespectively. Closed-formtheoreticalexpressionsforclassicserpentinespringsdevelopedbyBarillaroet.al.[ 113 ]areusedforcalculatingthespringconstantsandcompliancesoftheserpentinecableinX,YandZdirections.ThestiffnessequationsinX,YandZ-axesfortheserpentinecablearegivenby, KCx=(N+1)l3o 6EIzo+(N+1)l2olp 2EIzp)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) 50

PAGE 51

Figure3-2. Schematicandfree-bodydrawingsofelectrodecables.A)Schematicdrawingoftheelectrodeimplantedintothebrain,B)Free-bodydiagramofthestraightcable,andC)Free-bodydiagramoftheserpentinecable. whereNisthenumberofmeanders,loisthelengthofthemeanderelementperpendiculartothexandzaxes(m),lpisthelengthofthemeanderelementparalleltothex-axis,Izoisthemomentofinertiawithreferencetothez-axisofthemeanderelementsectionperpendiculartoxandzaxes(m4),andIzpisthemomentofinertiawithreferencetothez-axisofthemeanderelementsectionparalleltox-axis(m4). KCy=KCyzKCzyKCy KCyzKCzy)]TJ /F7 11.955 Tf 11.95 0 Td[(KCyKCz,(3) where, KCy=(2(N+2)lp)3 3EIzp+(8N3+36N2+55N+27)l2plo 3EIzo)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) KCyz=KCzy=2(N2+3N+4)lplo EIzo+2(N+2)2l2p EIzp)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) KCz=2(N+2)lp EIzp+2(N+1)lo EIzo)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) 51

PAGE 52

and KCz=KCzyKCyzKCz KCzyKCyz)]TJ /F7 11.955 Tf 11.96 0 Td[(KCzKCy,(3) where, KCz=(2(N+2)lp)3 3EIyp+(N+1)(lo)3 6EIyo+(N+1)(lo)2lp GJp+(8N3+36N2+55N+27)l2plo 3GJo)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) KCyz=KCzy=2(N2+3N+4)lplo GJo+2(N+2)2l2p EIyp)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) KCy=2(N+2)lp EIyp+2(N+1)lo GJo)]TJ /F3 7.97 Tf 6.59 0 Td[(1,(3) whereIyoisthemomentofinertiawithreferencetothey-axisofthemeanderelementsectionperpendiculartoxandzaxes(m4),andIypisthemomentofinertiawithreferencetothey-axisofthemeanderelementsectionparalleltox-axis(m4),Joisthecross-sectionaltorsionfactorofthemeanderelementperpendiculartoxandzaxes(m4),Jpisthecross-sectionaltorsionfactorofthemeanderelementparalleltoxaxis(m4),andGistheshearmodulus(Pa). Thespringconstantsforthestraightcableswerecalculatedusingthestandardstiffnessequationsforaclamped-guidedbeamobtainedfrom[ 114 ].ThestiffnessequationsinX,YandZ-axesforthestraightcablearegivenby Kx=Ehw L,(3) Ky=Ehw3 L3,(3) and 52

PAGE 53

Kx=Eh3w L3(3) DimensionsusedforAnalyticalandNumericalAnalysis:Table 3-1 givesthedimensionsofthestraightandtheserpentinecablesusedintheanalyticalandnumericalestimationofcompliance. Table3-1. Dimensionsofthestraightandserpentinecablesusedintheanalyticalandnumericalanalysis. DimensionStraightcableSerpentinecable Length6mm6mmWidth2.7mm2.5mmHeight40m40mNumberofmeanders2Lengthofmeander1mmWidthofmeander1mm MaterialUncertainty:Fromtheliteraturereview,itwasfoundthattheYoung'smodulusofpolyimiderangesfrom2.8GPato15GPadependingontheformulation.Table 3-2 givesalistofYoung'smodulusvaluesforpolyimideobtainedfromliterature. Table3-2. ListofYoung'smodulusvaluesofpolyimideobtainedfromliterature. ReferenceYoung'smodulusofpolyimide [ 94 ]2.793GPa[ 115 ]3.0GPa[ 116 ]7.5GPa[ 117 ]8.5GPa[ 116 ]8-15GPa CompliancewascalculatedforallthevalueslistedinTable 3-2 forboththestraightandtheserpentinecables.Meanandstandarddeviationswereobtainedforthecalculatedcompliancevaluestoaccountfortheerrorduetovariationinmaterialproperty.ThecalculatedcompliancevaluesforthestraightcableinX,YandZaxesare,(1.120.67)10)]TJ /F3 7.97 Tf 6.58 0 Td[(6(m/N),(5.543.29)10)]TJ /F3 7.97 Tf 6.58 0 Td[(5(m/N),and0.2520.149(m/N) 53

PAGE 54

respectively.Similarly,thecalculatedcompliancevaluesfortheserpentinecableinX,YandZaxesare,(1.891.12)10)]TJ /F3 7.97 Tf 6.59 0 Td[(3(m/N),(2.701.6)10)]TJ /F3 7.97 Tf 6.59 0 Td[(2(m/N),and26.615.78(m/N)respectively.TheresultsaresummarizedinTable 3-3 andTable 3-4 3.3.2NumericalAnalysis Followingtheanalyticalanalysis,anumericalanalysiswasperformed.AniteelementmodelwasdevelopedtosolveforthecompliancesofstraightandserpentinebeamsusingtheABAQUSsimulationtool[ 118 ].2DShellelementwasusedfortheanalysis.Themeshshapeisquadrilateralelementforstraightcablesandquadrilateralandtriangularelementsforserpentinecables.ThemeshsizeiswithintheABAQUSdefaultsizerange(min:1.3e-6andmax:0.0013),andthemeshnumberis50elementsforstraightcablesand833elementsforserpentinecables.ThetestforconvergenceemploystheABAQUSdefaultconvergencecriteriavaluesfornonlinearproblems[ 118 ].Theboundaryconditionsusedwereencastre(norotationalortranslationalmotioninanyaxis)forxedendandnorotationalmotion,onlytranslationaldisplacementinallaxesfortheguidedend.SimulationswereperformedforalltheYoung'smodulusvaluesofpolyimidegiveninTable 3-2 andtheresultswereaveragedandtheerrorwascalculated.TheresultsarepresentedinTable 3-3 andTable 3-4 3.4FlexibleCableandRigidModuleFabrication Newprocessowsweredevelopedforfabricatingthemodularelectrodesand2Dserpentinecables.Processingandpackagingstepsinvolvedinthefabricationofthe2DcablesareshowninFigure 3-3 .Allthestepsinvolvedintheprocessingaredoneonarigid4inchsiliconwafer.Theprocessbeginswiththespindepositionof20mthicklayerofpolyimideontopofasacricialaluminumlm.Next,athinlm(2000A)ofgoldissputterdepositedandlithographicallypatternedtodenetheconductivetraces.Thetopinsulationisprovidedbyspinningalayerof20mthickpolyimideoverpatternedgold.Thetoppolyimideisplasmaetchedusingthereactiveionetcher(RIE)toexposegoldbondpadssurroundingtheviaholes.Alsothethroughholesareobtained 54

PAGE 55

Figure3-3. Processowfor2Dtransverseexiblecablemodule. byetchingoffthebottompolyimideunderneaththem.Finally,thedeviceisreleasedfromtheSiwaferbyetchingthesacricialaluminumlayerthroughananodicdissolutionprocess.Tungstenmicrowiresmaybemanuallyassembledtotheholesina2Darrayandsecuredwithepoxy. TheprocessingandpackagingstepsinvolvedinthefabricationoftherigidelectronicsandconnectormoduleareshowninFigure 3-4 .Allthestepsarecarriedoutona4inchsiliconwafer.First,athinlmofSiO2isdepositedonsilicon.Similartotheprocessowoftheexiblemodule,athinlm(2000A)ofgoldissputterdepositedontheoxidelayerandlithographicallypatternedtodenetheconductivetraces.The 55

PAGE 56

topinsulationisprovidedbyspinningalayerof20mthickpolyimideoverpatternedgold.ThetoppolyimideisplasmaetchedusingRIEtoexposegoldbondpadsfortheconnectorandtheelectronics.Theholeforthegroundscrewandtheslotholeareobtainedbyetchingoffthebottompolyimideandoxideunderneaththem.Finally,thesiliconunderneaththescrewhole,slothole,andthesurroundingthedeviceoutlineisetchedoffusingthedeepreactiveionetcher(DRIE)toseparatethemodule.Oncethemodulesareseparatedintosinglepieces,theconnectorandthenutareattachedusingconductivesilverepoxy.Theexiblecableandmodulescanbepackagedtogetherusingconductivesilverepoxyandsealedwithunderllepoxy. PhotographsofthefabricatedserpentinecablepriortoattachmentofthemicrowiresandpackagedrigidmoduleareshowninFigures 3-5 Aand 3-5 Brespectively. 3.5ExperimentalMeasurementoftheCableCompliance Thissectiondescribestheexperimentsconductedtomeasurethecomplianceofthestraightandserpentinecablesamplesinallthreeaxes.Acomprehensivereportoftheexperimentalmethodandananalysisofthemeasurementuncertaintyareprovided. 3.5.1ExperimentalMeasurementoftheIn-Plane(X-axis)CableCompliance Thein-plane(orX-axis)stiffnessofthemicro-fabricatedstraightandserpentinecableelectrodeswasexperimentallymeasuredusingtheInstronr5900seriesmechanicaltestingsystem.Thecablesweresubjectedtoin-planetensilestress,andtheextensionintheaxialdirectionwasmeasuredtoevaluatethestiffnessintheX-axis.ThenamingoftheaxescanbefoundinFigure 3-2 Thecablesweresuspendedbetweentwoverticallypositionedclipsthatareconnectedtoloadcells,whichrunacrosstwoloadframes.Thetopframeismovablewhilethebottomremainsxed.Theextensioncontroltestingmethodwasusedtomeasurethecomplianceofthecables.Inthistestingmethod,thecableissubjectedtoaknownextension(fromx0tox1instepsofx)foragivenperiodoftime.FromHooke'slaw,F=kx,thechangeindisplacementinducesachangeintheforce.The 56

PAGE 57

Figure3-4. Processowforrigidelectronics/connectormodule. correspondingchangeintheforceFismeasured.Themeasuredchangeinforceisplottedagainstthechangeindisplacement,andtheslopeoftheresultingcurveiscalculatedtondthestiffnessofthecable. Inordertoensuretheconsistencyofthecablelengthwiththeanalyticalandnumericalcalculations,thecablesweremountedonpapertabswithholesinthecenterusinghotglueorcrystalbond.Thesizeoftheholescorrespondedtothelengthofthecable.Thepapertabswerethenattachedtothetwoverticalclips.Onceattached,thepaperwascutatthecentertopreventanyadditionalloadingduetothepaper.Thisset 57

PAGE 58

Figure3-5. Photographsoffabricateddevices.A)Photographofthemicrofabricatedserpentinepolyimidecablepriortotheattachmentofthetungstenmicrowiresand(B)PhotographshowingthefabricatedsiliconrigidmodulepackagedwiththeOmneticsconnectorandthegroundscrewcomparedagainstaonecentcoin. uphasmorecontrolonthegaugelengthofthecable.Figure 3-6 Ashowstheschematicofthepapermountedcablesuspendedbetweenthetwoverticalclips,andFigure 3-6 Bshowsthephotographofthecablemountedonthepapertab. Themeasurementdetailsaredescribednext.Thestraightcableswereextendedfrom0to50m.Themaximumloadlimitsetonthestraightcableswas300mN.Theserpentinecableswereextendedfrom0to40m,andthemaximumloadlimitonthemwas100mN.Onesampleineachofstraightandserpentinecablesweremeasuredforstiffnessandtentrialswereperformedoneachsampleforconsistency. 58

PAGE 59

Figure3-6. SchematicandphotographofcablemountedonInstronrmechanicaltestingsystem.A)SchematicdiagramofthepapermountedcablesuspendedbetweenthetwoverticalclipsconnectedtotheloadcellsandB)PhotographofthecablemountedonthepapertabconnectedtotheclipsoftheInstronrmechanicaltestingsystem.Insetshowsthecloserviewofthemountedcable. CompliancemeasuredfromallthetentrialsforthestraightandserpentinecablesampleisgraphicallyrepresentedinFigure 3-7 Uncertaintyanalysisoftheexperimentalerror:Theinstrumenterrorwasincludedforin-planeexperimentsbycalculatingthevariabilityintheloadmeasurementforeachmeasuredreadingbasedontheaccuracyvalues(5%)obtainedfromtheinstrumentmanual[ 119 ].Theinstrumenterrorwasincludedforout-of-planeexperimentsbycalculatingthedriftintransducerdisplacementfortheloadingduration.Thevalueusedfortransducerdisplacementdriftwasobtainedfromtheinstrumentmanualas0.05nm/sec[ 120 ]. Themeasuredvaluesfromdifferenttrialsforeachsamplewereanalyzedforexperimentaluncertaintyandthecondenceintervalwasconstructed.Twotailedt-test 59

PAGE 60

Figure3-7. MeasuredX-axiscompliancevaluesfromstraightandserpentinecables.Loadmeasurementvariabilitywasnearlynegligible(twoordersofmagnitudeless)forstraightandserpentinecables. wasusedforcondenceintervalcalculationssincethenumberoftrials(samples)islessthan30. Thecondenceintervalforat-distributionisgivenby, C.I.= xt=2,n)]TJ /F3 7.97 Tf 6.58 0 Td[(1s p n,(3) where x=mean,=signicancelevel(0.05for95%C.I.),n-1=numberofdegreesoffreedom,n=samplesize,ands=samplestandarddeviation,whichisgivenby, s=vuut 1 n)]TJ /F12 11.955 Tf 11.95 0 Td[(1nXi=1(xi)]TJ ET q .478 w 279.79 -543.11 m 286.44 -543.11 l S Q BT /F7 11.955 Tf 279.79 -550.43 Td[(x)2,(3) 3.5.2ExperimentalMeasurementoftheOut-of-Plane(Y-axis)CableCompliance Theout-of-plane(orY-axis)stiffnessofthemicro-fabricatedstraightandserpentinecableelectrodeswasexperimentallymeasuredusingtheHysitronrTriboindenter.The 60

PAGE 61

cableswereloadedbyloweringtheindentertipdownforapresetloadlimitandadepthlimit,andthechangeinthedisplacementwasmeasuredasafunctionofthechangeintheloadtoevaluatethestiffness.Aconicaltipofradius20.1mwasusedforloadingthesamples. SamplePreparation:InordertofacilitateverticalloadingofthecablesalongitsthinY-axis,thecablesweremountedverticallyonaglassslideandsecuredwithtwomagnetpiecesonandsteelnutsoneitherside.Thissetupensuredverticalstandingofthecablewithoutmuchmovement.Itwasalsoconrmedthatthenutsandthemagnetsremainsteadyduringtheloadingofthetip.Figure 3-8 showsaphotographofthestraightcablemountedverticallyonaglassslideswithsupportingmagnetsandnuts.Onesampleofstraightcableandonesampleofserpentinecablewerepreparedfortheexperiment. Figure3-8. Photographshowingthestraightcablemountedverticallyonaglassslideandsecuredwithmagnetpiecesandsteelnutsoneithersides. ExperimentSetup:ThemountedsamplewasintroducedintotheTriboindenterchamberandplaceddirectlyundertheconicaltip.Thesystemwasallowedtothermallystabilizefor10minutes.First,airindentcalibrationwasdone.Next,theinitialposition(orthezeropoint)wascalibratedbyslowlyloweringthetipandestablishingacontactwiththecableataforce<2N.Aftercalibration,thecablewasloadedbyloweringthetipfurtherdownforapresetloadlimitandadepthlimit.Themaximumloadlimitwassetat 61

PAGE 62

25N,andthemaximumdepthwassetat4700nm.Thechangeinthedisplacementwasmeasuredasafunctionofthechangeintheload,andtheslopeoftheresultingplotwascalculatedtoobtainthecompliancevalue.Fivetrialswereperformedoneachsampleforconsistencyandstatistics.Figure 3-9 showstheschematicoftheverticalloadingoftheserpentinecablebytheconicalnanoindentertipalongwithaphotograph.Uncertaintyanalysisofthemeasurementerrorwasdoneusingthesamemethoddescribedforin-planeexperiment.CompliancemeasuredfromallthevetrialsforthestraightandserpentinecablesamplesisgraphicallyrepresentedinFigure 3-10 Figure3-9. SchematicdiagramshowingtheverticalloadingoftheserpentinecableinY-axisbythenanoindentertip.Insetshowingthephotographofthecableandthenanoindentertip. 3.5.3ExperimentalMeasurementoftheOut-of-Plane(Z-axis)CableCompliance Theout-of-plane(orZ-axis)stiffnessofthemicro-fabricatedstraightandserpentinecableelectrodeswasexperimentallymeasuredusingtheHysitronrTriboindenter.Thecableswereloadedbyloweringtheindentertipdownforapresetloadlimitandadepth 62

PAGE 63

Figure3-10. MeasuredY-axiscompliancevaluesfromstraightandserpentinecables.Theerrorbarsrepresentthevariabilityduetotransducerdisplacementdriftduringtheloadingperiod. limit,andthechangeinthedisplacementwasmeasuredasafunctionofthechangeintheloadtoevaluatethestiffness.Forminimizingexperimentalerrors,thesameconicaltipofradius20.1musedintheY-axismeasurementwasusedforloadingthesamples. SamplePreparation:Priortotheexperiment,theelectrodesampleswerepreparedtofacilitatetheloadingofthenanoindentertipontheedgeofthecableendwithoutanydamagetotheelectrodeandthetip.Theelectrodesubstratesweremountedonaglassslideandrmlysecuredwithhotglueatthebackend.Thecableendofthesubstratewasextendedbeyondtheedgeoftheglassslidetoallowfreemovementuponloading.Athinmicroscopeslidewasaddedontopofthebackendtoensureatnessofthesubstrate.Figure 3-11 Ashowsthephotographofthestraightcablemountedonaglassslideandsecuredwiththehotglueandmicroscopeslide.Themountedsampleswerethenplacedonthebottomplateofthenanoindenter.Sincethecableswillbeloadedverticallybythenanoindentertip,itisrequiredtohaveenough 63

PAGE 64

clearanceintheplaneofloading.TwosmallpiecesofmagnetswereplacedunderneaththeglassslidetoincreasetheclearanceofthesampleintheZ-axis.Figure 3-11 Bshowsthephotographofthesampleplacedonthebottomplatewithmagnets.Onesampleofstraightcableandonesampleofserpentinecablewerepreparedfortheexperiment.Aftercompletingtheexperiment,thesampleswereremovedfromtheglassslidesbyheatingtheglueandseparatingtheslides. A B Figure3-11. Photographsofmounteddevices.A)PhotographshowingthestraightelectrodemountedonaglassslideandsecuredwithhotglueandathinmicroscopeslideandB)Photographshowingtheserpentineelectrodemountedonthebottomplateofthenanoindenter. ExperimentSetup:ThesamplemountedonthebottomplatewasintroducedintothechamberoftheHysitronrTriboindenter,andplaceddirectlyundertheconicaltip.Atrst,airindentcalibrationwasdonesimilartoY-axisloading.Later,thezeropointwascalibratedbyslowlyloweringthetipandestablishingacontactwiththecableataforce<2N.Aftercalibration,thecablewasloadedbyloweringthetipfurtherdownforapresetloadlimitandadepthlimit.Themaximumloadlimitwassetat25Nandthemaximumdepthwassetat4700nm.Thechangeinthedisplacementwasmeasuredasafunctionofthechangeintheload,andtheslopeoftheresultingplotwascalculatedtoobtainthecompliancevalue. 64

PAGE 65

Figure3-12. SchematicdiagramshowingtheverticalloadingofthecableintheZ-axisbythenanoindentertip.Insetshowsthezoomedinphotographofthetipandthecable. Theexperimentwasconductedatthreedifferentpointsinthecable-theleftmosttip,therightmosttip,andthecenterpoint.Fivetrialswereperformedoneachmeasuringpointforconsistencyandstatistics.Thedataobtainedfortherightandlefttipsshowedhighnon-linearityforserpentinecable.Henceonlythedataobtainedathecenterpoint,whichwerelinear,wereusedforanalysis.Figure 3-12 showstheschematicoftheverticalloadingofthecableintheZ-axisandthephotographofthenanoindentertipandthecable.Uncertaintyanalysisofthemeasurementerrorwasdoneusingthesamemethoddescribedforin-planeexperiment.CompliancemeasuredfromallthevetrialsforthestraightandserpentinecablesamplesisgraphicallyrepresentedinFigure 3-13 .Theresultsarediscussedinthefollowingsection. 3.6ResultsandDiscussion Theresultsobtainedfromtheanalytical,numerical,andtheexperimentalcompliancemeasurementarediscussedinthissection.Table 3-3 givesasummaryofthestraightcablecompliancedeterminedanalytically,numericallyandexperimentallyforallthree 65

PAGE 66

Figure3-13. MeasuredZ-axiscompliancevaluesfromstraightandserpentinecables.Transducerdisplacementdriftduringtheloadingperiodwasnearlynegligible(twoordersofmagnitudeless)forstraightand(fourordersofmagnitudeless)serpentinecables. axes,andTable 3-4 givesasummaryoftheserpentinecablecomplianceestimatedusingthethreemethodsforallthreeaxes.Itcanbeobserved(fromtheexperimentalresults)thatthecompliancevaluesoftheserpentineshapedcablesisatleastoneorderofmagnitudehigherthanthecomplianceofthestraightcablesofthesamedimensions.Thehighercomplianceorexibilityofthenewserpentineshapedcablesisexpectedtolessenthefront-endstrainoftheelectrodeonthetissue.Mitigatedfront-endstrainisexpectedtoreducethetissueimmuneresponseandimprovethereliabilityoftheimplant'ssignalrecordingquality. ItcanbenotedfromTables 3-3 and 3-4 thatthereissomediscrepancybetweentheanalyticalandthenumericalresultsandtheexperimentallymeasuredvalues.Thisdiscrepancycouldhaveresultedfromtheassumptionsmadewhiledevelopingthebeammodelsforanalyticalandnumericalanalysesandfromthelimitationsinthegeometryof 66

PAGE 67

Table3-3. Matrixcomparingthestraightcablecomplianceestimatedthroughanalytical,numericalandexperimentalanalysis. TypeofanalysisX-axiscompliance(m/N)Y-axiscompliance(m/N)Z-axiscompliance(m/N) Analytical(1.120.67)10)]TJ /F3 7.97 Tf 6.59 0 Td[(5(5.543.29)10)]TJ /F3 7.97 Tf 6.59 0 Td[(50.2520.149 Numerical(1.120.67)10)]TJ /F3 7.97 Tf 6.59 0 Td[(5(2.571.52)10)]TJ /F3 7.97 Tf 6.59 0 Td[(40.250.149 Experimental(5.690.48)10)]TJ /F3 7.97 Tf 6.59 0 Td[(5(7.431.07)10)]TJ /F3 7.97 Tf 6.59 0 Td[(30.1880.06 Table3-4. Matrixcomparingtheserpentinecablecomplianceestimatedthroughanalytical,numericalandexperimentalanalysis. TypeofanalysisX-axiscompliance(m/N)Y-axiscompliance(m/N)Z-axiscompliance(m/N) Analytical(1.891.12)10)]TJ /F3 7.97 Tf 6.59 0 Td[(3(2.71.6)10)]TJ /F3 7.97 Tf 6.58 0 Td[(226.615.78 Numerical(1.921.14)10)]TJ /F3 7.97 Tf 6.59 0 Td[(3(9.85.8)10)]TJ /F3 7.97 Tf 6.58 0 Td[(31.170.69 Experimental(5.340.19)10)]TJ /F3 7.97 Tf 6.59 0 Td[(4(6.81.8)10)]TJ /F3 7.97 Tf 6.58 0 Td[(21.540.56 theniteelementsolver.Aclamped-guidedbeammodelwasassumedfortheanalyticalandthenumericalanalyses.Theboundaryconditionsofthismodelallowtheguidedendtodeectnormaltoitsaxis,whilerestrictingitsrotationalmotion.Howeverinpractice,therewillbesomerotationalmotiondisplayedbythebeamwhichwillcontributetotheoverallbeamdeection.Furthermore,inthecaseoftheserpentinestructure,themeanderswillhaveexuraldegreesoffreedomwhichwillbedifferentfromtherigidbodydegreesoffreedom,andtherewillbeanadditionaleffectofbeamtwistingseenintheserpentinecables. 3.7Summary Strainduetomicromotionisoneoftheimportantabioticcausesdrivingthefailureoflong-termneuralimplants.Therstpartofthisresearchworkisfocusedonaddressingthisissue,bycontributingadesignsolutionforfront-endstrainrelief.Thischaptergivesadetailedaccountofthedesignapproachadoptedforbuildingahighlycompliantelectrodecable,byexplainingtherationalebehindthechosenserpentinedesignandexplainingthepotentialstrainreliefusingabeammodel.Theanalyticalandthenumericalanalysisaredescribed.Fabricationprocessforthecablesandthe 67

PAGE 68

experimentalmeasurementmethodsaredescribed.Finallyadiscussionoftheresultsisprovided.Resultsobtainedfromthisstudyimplythattheserpentinecabledesignwillsignicantlyimprovethefront-endstrainrelief,andtherebyisexpectedtohavemitigatedtissueresponseandimprovedsignalreliability. 68

PAGE 69

CHAPTER4BACKGROUNDONELECTRODEIMPEDANCEANDINFLUENCEOFELECTRODESURFACEVARIATIONSONCHRONICSIGNALRECORDINGRELIABILITY 4.1Introduction Impedanceisoneoftheimportantabioticmetricsthatdeterminestherecordingcompetencyofanelectrode.Variationsintheelectrodeimpedanceaffectthequalityoftherecordedsignalandcandeterioratethesignal.Generally,recordingmicroelectrodesaredesignedforanoptimalimpedancevaluethatwillenabletheelectrodetocaptureactionpotentialswiththebestsignaltonoiseratio.Itisalwaysdesiredthattheimpedanceofarecordingmicroelectroderemainsthesameallthroughthelifetimeoftheelectrode.Unfortunately,severalfactorsaltertheimpedanceofthemicrowireelectrode,whichinturnhasanegativeinuenceonthesignalrecordingreliabilityoftheelectrodes.Thisphenomenonislargelyobservedinchronicapplications[ 121 ].Atthispointoftime,thereisnoclearunderstandingofwhatfactorsinuencethevariationinimpedanceandtowhatextent.Thesecondpartofthisresearchisaimedataddressingthatquestion,withaspecialfocusonabioticfactors.Thischapterdiscussestheimpactofsomeabioticfactorsincludinginsulationdelamination,cracking,andvariationsinrecordingsitesurfacemorphologyonelectrodeimpedancebasedontheliteraturereviewandoursystematicstudies. Thischapterisorganizedasfollows.Section 4.2 givesanoverviewofthephysicsoftheelectrode-electrolyteimpedanceandadescriptionofthechargetransportmechanismsattheelectrode-electrolyteinterface.Section 4.3 reviewsthechronicperformanceofcommonlyusedinsulationmaterialsandprovidesaqualitativediscussionoftherelationbetweenelectrodesurfacevariationsandimpedancebasedonpriorScanningElectronMicroscope(SEM)imagingstudiesandin-vivoimpedancemeasurements.AlsothemethodofelectrodesurfaceroughnessmeasurementusingLaserScanningMicroscopeisdescribedinsubsection 4.3.4 .Section 4.4 discussesthe 69

PAGE 70

impactofrecordingsitevariationssuchascorrosionofmetalandcoatingofconductivematerialsonthequalityofrecordedsignal. 4.2PhysicsofElectrodeImpedance Thissectionprovidesthephysicalbackgroundinformationonelectrodeimpedancewhichisnecessarytounderstandthesubsequentsectionsandchapters.Forabetterunderstandingoftherelationbetweentheelectrodeinsulationanditsimpedance,thephysicsbehindtheelectrode-electrolyteinterfaceisexplainedrst,followedbyadetaileddescriptionofthechargetransportmechanisminelectrolytesandanoverviewoftheelectrode-electrolyteinterfaceinabiologicalenvironment. 4.2.1Electrode-ElectrolyteInterface Theelectrode-electrolyteinterfacecanbeclassiedasfaradaicandnonfaradaicinterfaces.Inafaradaicinterface,theelectrodeisreactiveandelectrochemicalreactionsoccurattheelectrodesurface.Inanonfaradaicinterface,theelectrode,oftenreferredasblockingelectrode,isnonreactiveandthereisnoelectrochemicalreactionatthesurface.Thephysicsgoverningtheelectrode-electrolyteinterfaceispracticallythesameforbothfaradaicandnonfaradaicinterfaces;theonlyadditioninfaradaicinterfaceisthemodelrepresentingthefaradaicchargetransfermechanism.Thenextfewpassageswillprovideabriefdescriptionofthephysicsbehindthenon-faradaicandfaradaicinterfaces. Onceametalelectrodeisintroducedintoanelectrolytesolution,aspacechargeregionisformedattheelectrode-electrolyteinterface.Thespacechargeregion,commonlyidentiedasadoublelayerwasrstdescribedbyHelmholtzin1853andlaterstudiedbyGouy,ChapmanandSterninearly1900s[ 122 ].Figure 4-1 givesanillustrationofthedoublelayerelectrode-electrolyteinterfaceatequilibrium.ItcanbeseenthattheinterfacehasaninnerHelmholtzplane(IHP)andanouterHelmholtzplane(OHP),whichtogetherformstheHelmholtzdoublelayer.TheIHPcontainstheionsandmoleculesthatareadsorbedatthesurface,suchaspolarizedwatermolecules,and 70

PAGE 71

Figure4-1. Electrode-electrolyteinterfaceatequilibrium[ 5 ]. theOHPcontainsthesolvatedadsorbedions[ 123 ].TheregionbetweenIHPandOHPconsistsofaxedlayerofchargethatisusuallymodeledusingaxedcapacitanceCHdl.OutsidetheOHPisthediffuselayerwhichisaregionwithanetelectricalchargethatdecaysexponentiallytozerowithdistance,thusmergingwiththebulkelectrolyte.ThisagainismodeledbythediffuselayercapacitanceCdi. Intheory,thereciprocaloftheequivalentdoublelayercapacitance,Cdl,isthesumofthereciprocalsoftheHelmholtzdoublelayercapacitanceandthediffuselayercapacitance,asgiveninequation 4 1 Cdl=1 CHdl+1 Cdi,(4) However,inpractice,whentheconcentrationoftheelectrolyteissufcientlyhigh,Cdibecomeslargeandcouldbeneglected[ 124 ].Foranonfaradaicinterface,theequivalentelectricalcircuitisgivenbyaseriescombinationofthedoublelayercapacitance,Cdl,andthebulkelectrolyteresistance,Re,asshowninFigure 4-2 71

PAGE 72

Figure4-2. Electricalequivalentcircuitfornonfaradaicelectrode-electrolyteinterface. Figure4-3. Randleselectricalequivalentcircuitforelectrode-electrolytefaradaicinterface. Incaseoffaradaicsystems,theelectrode-electrolyteinterfacecanberepresentedbytheequivalentelectricalcircuit(Randlescircuit)[ 5 ]asshowninFigure 4-3 ,wheretheelectrolyteresistanceisrepresentedbyRe,andtheelectrode-electrolyteinterfaceimpedanceisrepresentedbyaparallelcombinationofthedoublelayercapacitanceCdlandthechargetransferresistance,Rct.Athigherfrequencies,theresistanceandthecapacitancemodelingthedoublelayercanbeneglected,andtheequivalentcircuitcanberepresentedonlybythetermRe. ConstantPhaseElement(CPE):Innon-idealFaradaicsystems,thedoublelayercapacitanceisusuallyrepresentedbytheconstantphaseelement(CPE).CPEisa 72

PAGE 73

Figure4-4. Electricalequivalentcircuitfornon-idealfaradaicinterfacewithCPE. highlydynamicelement,anditsbehaviorislargelydependentonthenon-homogeneityofthesurface[ 125 ].TheimpedanceoftheCPEisgivenby, ZCPE=1 (j!)Q(4) where!isthefrequency,isanumberfrom0to1,andQistheCPEcoefcientwiththeunitss=cm2[ 125 ].When=1,CPEbecomesthedoublelayercapacitance.Figure 4-4 showsthenon-idealfaradaiccircuitwithCPE. 4.2.2ChargeTransportinElectrode-ElectrolyteInterface Thetransportofionspeciesinanelectrolyteisusuallyattributedtothreemechanisms:migration,diffusion,andconvection[ 123 ].Migrationisthemechanismwheretheionsinanelectrolytesolutionaredrivenbyanappliedelectriceld.Diffusionisthemechanismwheretheconcentrationgradientoftheelectrolytedrivesthespecies.Convectionisthebulktransportoftheuid.Theuxdensityoftheionsisacombinationofallthreemechanisms,anditisgivenby[ 123 ], Ni=)]TJ /F7 11.955 Tf 9.3 0 Td[(ZiuiFCir)]TJ /F7 11.955 Tf 11.95 0 Td[(DirCi+CiV(4) whereNiistheuxdensityoftheionicspeciesiwithunitsmol/s.cm2,Ziisthenumberofprotonchargescarriedbyanion,uiisthemobilityofionswithunitscm2.mol/J.s,FistheFaraday'sconstant=96485.3415Coulomb,Ciistheconcentrationofthe 73

PAGE 74

ions[]/cm3],ristheelectriceld[V/cm],Diisthediffusioncoefcientofspeciesi[cm2/s],rCiistheconcentrationgradient[1 cm.] cm3],andVisthevelocityofbulkuidmotion[cm/s]. Thecurrentintheelectrolytecanbeexpressedas, i=FXiZiNi=)]TJ /F7 11.955 Tf 9.3 0 Td[(F2rXiZi2uiCi)]TJ /F7 11.955 Tf 11.96 0 Td[(FXiZiDirCi+FvXiZiCi(4) assumingtheintracellularandextracellularuidicspaceishomogeneouswithuniformdensity[ 126 ],andthetransportofionsduetoconvectioncanbeneglectedandthetermcanbetreatedaszero.Sincetheintracellularandextracellularionicconcentrationsareapproximatelyconstant,itcanbeassumedthattheconcentrationgradientiszero[ 126 ],andthediffusiontermcanalsobeneglected.Thissimpliestheaboveexpressioninto i=)]TJ /F8 11.955 Tf 9.29 0 Td[(r,(4) where, =F2Xiz2iuiCi.(4) istheconductivityofthesolution.Theowofionsaffectsthepotentialbychangingthelocalchargedensity,.ThisisgivenbyGauss'slawas, "r2=)]TJ /F8 11.955 Tf 9.3 0 Td[(.(4) Theintracellularandextracellularuidsareelectricallyneutral[ 126 ].Hencethechargedensity,iszero,andGausslawisessentiallyrepresentedbythefollowingLaplace'sequation, r2=0.(4) 74

PAGE 75

Figure4-5. Electricalequivalentcircuitforelectrode-electrolyteinterfaceinabiologicalenvironment. ItisunderstandablethatsolvingtheLaplace'sequationwillprovidethepotentialdistributionacrosstheelectrode.Foradiskelectrodewiththereferenceelectrodepositionedatasemi-innitedistance,theresistanceofthebulkelectrolyteReisgivenby[ 123 ], Re=1 4r0,(4) where,r0istheradiusoftheelectrode. 4.2.3Electrode-ElectrolyteInterfaceinBiologicalEnvironment Whentheelectrodeisimplantedintoananimal,anadditionalimpedancelayerisformedattheelectrode-electrolyteinterfaceduetotissueencapsulation.Severalequivalentcircuitshavebeenproposedtomodeltheelectrode-tissue-electrolyteinterfaceimpedance[ 127 129 ].Inessence,allthemodelsincludeaparallelcombinationofacapacitorandresistorrepresentingthetissueimpedanceplacedinserieswiththenon-idealRandlescircuit.Somecircuitsincludesubcomponentsmodelingtheproteinencapsulationandextracellularmatrixseparately[ 127 129 ].Figure 4-5 showstheelectrode-tissue-electrolyteinterfaceequivalentcircuit. 75

PAGE 76

4.3ReviewofResultsfromPriorIn-vivoandIn-vitroStudies ThissectionprovidesaliteraturereviewofthelongtermperformanceofcommonlyusedelectrodeinsulationmaterialssuchasSiO2,Si3N4,polyimideandparylene-C.TheDARPAfundedHistologyforInterfaceStabilityoverTime(DARPA-HIST)project,whichisacollaborativeeffortofUniversityofFloridaandUniversityofMiami,isbrieyintroduced.AlsoadiscussionoftheobservationsobtainedfrompreviousSEMcharacterizationstudies,in-vivoimpedancestudies,andsurfaceroughnessanalysisprocedureisprovided. 4.3.1ChronicPerformanceofCommonlyUsedInsulationMaterials SomeofthecommonlyusedelectrodeinsulationmaterialsareinorganicdielectricssuchasSiO2andSi3N4andorganicpolymerssuchaspolyimideandparylene-C.Najaet.al.usedSiO2andSi3N4asprobeinsulationmaterialfortheirpolysiliconortantalumconductingelectrodes[ 82 ].ThechoiceofSiO2andSi3N4fortheprobeinsulationwasdrivenbytheideathattheconductorandtheinsulatorwillhaveasmoothcontactandtheremightnotbeanyinterfaceissuesarisingduetodistinctadhesionpropertiesoftennotedwhileusingdissimilarmaterials.However,underchronicconditions,theinsulationbecameunstable.After250hoursofsoakinginsaline,theprobeimpedancestarteddroppingfromitsstablevalue[ 82 ].TheauthorsinferthatthechosenSiO2andSi3N4arenotsufcientforencapsulatingelectrodesforchronicapplications,andthereisaneedforadditionalcoatingoforganicpolymerssuchaspolyimideorparylene-Cforimprovedstability. Severalelectrodedesignsusingpolyimideorparylene-Casprobeencapsulationmaterialshavebeenpublishedintheliteratureandareavailablecommercially[ 2 3 79 80 ].Thoughthesematerialshavemultipleadvantagessuchasexibility,biocompatibility,andsimplefabricationprocess,theytendtofailinchronicconditionsowingtotheirchangesincrystallineandmechanicalpropertiesunderprolongedexposuretoanaqueousenvironment.Anin-vitrostudyconductedbyagroupatthe 76

PAGE 77

UniversityofFreiburg,Germanyevaluatedthefeasibilityofpolyimideasamaterialforchronicneuralimplantbycharacterizingitstensilestrength,crystallinity,andchemicalcompositionatdifferenttemperatures[ 130 ].SamplesofBPDA-PPD(biphenyldianhydride/1,4phenylenediamine)typepolyimidewerestoredinPBSsolutionat37Ctosimulatein-vivoconditionsandat60Cand85Ctosimulateacceleratedagingconditions.ItwasobservedthatsamplesstoredinPBSsolutionat85Cfor20months,weresensitivetosodiumions,constantlyabsorbedwater,anddisplayedchangesinitscrystallinestructureandadecreaseinitsmechanicalproperties.Morerecently,thesamegrouphasreportedtheirobservationsonchronicadhesionpropertiesofpolyimidetodifferentimplantmaterialsincludingplatinum,gold,andtungsten-titanium[ 131 ].Sampleswerestoredinsalineforoveroneyearat60Ctosimulateacceleratedagingconditions.Allsamplesattachedtoplatinum,gold,andtungsten-titaniumsubstratesshoweddelaminationafterweeks.Whilepolyimideattachedtogoldfailedimmediately,thoseattachedtoplatinumandW-Ti,showedbubbleformationafterfewweeks,becamebrittle,anddelaminated.Thehighwaterabsorption(4wt%)ofpolyimide[ 132 133 ]isoftenattributedasaprimarylimitationforeffectiveinsulationunderchronicconditions. Thoughparylene-Chasasmallermoistureabsorptionrate(0.06%-0.1%)[ 134 135 ]thanpolyimide,itisstillfoundtobeunstableinsalineenvironmentduringchronicexposure.ThestudiesconductedbyHassleret.al.[ 134 136 ]haveshownthatparylene-Cincontactwithsalinesolutionfor38daysshowedanincreaseinmodulusofelasticityandadecreaseintensilestrength.AlsoPtandSi3N4samplescoatedwithparylene-Cwhenimmersedinsalinesolutionshowedcondensationofwateranddelaminationofparylene-Clmafter38days[ 136 ]. 4.3.2ReviewofSEMImagesofimplantedTucker-DavisTechnologies(TDT)andUFTungstenMicroelectrodes TheDARPA-HISTreliabilityproject,whichinvolvesresearchgroupsfrommultipledisciplinesattheUniversityofFloridaandtheUniversityofMiami,isaimedatunderstanding 77

PAGE 78

thefailuremechanismsforin-vivoneuralimplants,bydevelopingaresearchplanthatcouplesthestudyofbioticandabioticmetricsinasystematicway.Asaneffortfortheabioticstudy,25polyimideinsulated16-channelstungstenmicrowireelectrodearraysacquiredfromTucker-DavisTechnologiesInc.(TDT)werecharacterizedusingscanningelectronmicroscope(SEM)(CarryScopeSEM,JEOL,Inc.,andFEIXL-40eldemissiongunSEM)bothpriortoandafterimplantationinratmodels.Thesewireshaveatungstendiameterof50m,coatedwithagoldlmofapproximatethickness2-3m,andencapsulatedbyathinlm(5mthickness)ofpolyimide.Theimplantdurationoftheseelectrodearraysrangedfromafewhoursto9monthstostudytheacute,recovery,andchronicbehaviorrespectively. Comparisonbetweenthepre-implantandthepost-implantSEMimagesofthetungstenmicroelectrodearrayspresentedsignicantinsulationdamageinsomeofthewires.Thedamagesobservedinthewireshadwidevariationsincludingnarrowslits,mediumsizedcracks,andlargedelamination.Interestingly,thewiresthathaddamagedinsulationalsodisplayedlowneuronalyield.Thearraysthatwereimplantedforacuteandrecoverytimeperiods,didnotshowsignicantinsulationdamage.However,therecordingsiteoftheseelectrodearraysshowedsomesurfacedeteriorationincludingchangesinsurfaceroughnessandappearanceofshallowpitstodeepcavitiessuggestingcorrosionoftungsten.Adetailedreportoftheobservationfromalltheanimalscanbefoundin[ 137 ].Figures 4-6 Aand 4-6 BshowthepreandpostimplantSEMimagesofarepresentativetungstenmicroelectrodearrayusedinthisstudy,and 4-6 Cshowstheneuronalyieldplottedagainstthewiresforthatarray.Thiselectrodearraywasimplantedintheanimalformorethan130days.ItcouldbeseenfromFigure 4-6 Bthatwirenumber7(leftontheimage)hasundergonesomeinsulationdamageneartherecordingtipduringimplantation.AlsoitcanbenotedfromFigure 4-6 Cthatwire7isoneoftheprobeswithlowneuronalyieldinthearray. 78

PAGE 79

A B C Figure4-6. SEMimagesandneuronalyieldplotoftungstenmicroelectrodeswithpolyimideinsulation.A)Pre-implantationimageofelectrodes1,3,5and7forpolyimidecoatedtungstenelectrodearrayR9usedintheDARPA-HISTreliabilitystudy,B)Post-implantationimageofelectrodes7and2showingalargecrackinpolyimideinsulationnearthetipofwire7,C)Plotshowingtheneuronalyieldforallthewiresinthearrayduringtheimplantedperiod. Similarly,anumberofparylene-CinsulatedplatinummicrowireelectrodearraysobtainedfromMicroProbesInc.wereimplantedintheratmodelsfordifferenttimeintervals.TheelectrodearrayswerecharacterizedusingSEMbeforeandafterimplantation.Similartothepolyimideinsulatedtungstenarrays,theseelectrodearraysalsodisplayedinsulationdamage.Figures 4-7 Aand 4-7 BshowthepreandpostimplantSEMimagesofarepresentativeplatinummicroelectrodearrayusedinthisstudy.Itcan 79

PAGE 80

benotedfromFigure 4-7 bthatwire1hasalargepieceofparylene-Cmissingatthetipwhichwasquiteintactbeforeimplantation. A B Figure4-7. SEMimagesofplatinum-iridiumelectrodeswithparylene-Cinsulation.A)Pre-implantationimageofelectrodes1and2forparylene-CcoatedplatinumelectrodearrayF01-MA2usedintheDARPA-HISTreliabilitystudy,B)Post-implantationimageofelectrodes1and2showingmissingpieceofparylene-Cinsulationnearthetipofwire1. ThisphenomenonofinsulationdamagewasalsoobservedinelectrodearraysfabricatedandtestedatUF.Tungstenmicrowiresinsulatedwithathinlmofpolyimide(similartoTDTarrays)wereimplantedin-vivofor87days.SEMimagesweretakenbeforeimplantationandafterexplantation.Postimplantimagesshowedswellingofpolyimideinsulationandsomemetalexposedunderneaththeinsulation[ 5 ].Thesequalitativeobservationssupportthewidelyacceptedinferencethatbothpolyimideandparylene-Carepronetostructuralchangesunderchronicexposuretosalineenvironment. 4.3.3ReviewofIn-vivoImpedanceMeasurements TheDARPA-HISTreliabilityprojectalsoincludedquantitativeandfunctionalstudyoftungstenandplatinummicrowireelectrodearraysinanin-vivosetup.Asapartofthateffort,commerciallyacquiredpolyimideinsulatedtungstenmicroelectrodearrays(TDTInc.)wereimplantedin22ratsfordifferenttimeperiodsupto21weeks,and 80

PAGE 81

dailyimpedancemeasurementsandelectrophysiologicalrecordingswereobtainedfromalltheanimals[ 121 ].Theresultsfromthisstudyconrmedastrongfunctionalcorrelationbetweentheelectrodeimpedanceandtheoverallneuronalyieldduringtheimplantedduration.Itwasalsoobservedthatlowarrayyieldswereassociatedwithverylowimpedancevaluesorveryhighimpedancevalues,andthebestarrayyieldwasobservedforanimpedancerangeof40-150Kfortheimplantedtungstenwires.However,thisoptimalrangeofimpedancedisplayedinconsistenttemporalevolutionacrossalltheimplantedanimals.Thoughtheseobservationssuggestthattheelectrodeimpedanceisaffectedbysomeshort-termandlong-termfactors,itisnotclearwhattheunderlyingdrivingmechanismis.Thismotivatestheneedforasystematicevaluationoftheeffectoftheprobestructuralvariationonimpedance. 4.3.4AnalysisofElectrodeSurfaceRoughnessusingKeyencerLaserScanningMicroscope ThepreandpostimplantSEMimagesoftheelectrodeprovidedonlyaqualitativeideaofthevariationinthesurfacemorphology.Inordertoquantitativelyassessthesurfaceroughnessandcorrosiondepthoftheimplantedelectrodes,someofthearrayswereanalyzedforsurfaceroughnessusingalaserscanningmicroscope(KeyencerVK-9700Color3DLaserScanningMicroscope)attheMaterialsEvaluationandTestingLaboratory,SouthDakotaStateUniversity.Fourimplantedtungstenelectrodearrays,tworepresentativearrays(R9andR13)implantedforchronicduration,onerepresentativearray(R19)thatwasimplantedforacuteduration,andonerepresentativearray(R24)thatwasimplantedforrecoveryduration,wereanalyzedforsurfaceroughness.Onlythearraysthathadgoodelectrophysiologicalrecordingswereselectedfortheanalysis.Thedepthproleofalmostallthewiresineacharrayweremeasured,andtheaverageheightwithrespecttothereferencelinewascalculated.Figure 4-8 showsthelasermicroscopeimageandthe3Ddepthproleofthecorrodedrecordingsiteofwire1ofchronicelectrodeR9. 81

PAGE 82

A B Figure4-8. ImagesofelectrodeR9obtainedfromKeyencerlaserscanningmicroscope.A)Laserimageoftherecordingsiteofwire1ofR9,B)3Ddepthproleoftherecordingsiteofwire1ofR9. 4.4ImpactofElectrodeSurfaceModicationonSignalQuality Asseeninsection 4.2 ,theimpedanceoftheelectrodeislargelydecidedbytheelectrolyteimpedancewhichinturnisafunctionofitssurfacearea.Anymodicationsorvariationsinthesurfaceoftherecordingsitewillbetranslatedtotheelectrodeimpedanceandadverselyaffectthequalityoftherecordedsignal.Thissectiondiscussestheeffectofcorrosionoftherecordingelectrodeonthemeasuredsignalquality,andreviewsdifferentschemesadoptedtoimprovetherecordingsitesurfacearea. 4.4.1EffectsofRecordingSiteCorrosiononSignalQuality Corrosionofthemetalrecordingsiteisconsideredasoneoftheabioticsourcesoffailuremodesoflongtermneuralrecording[ 138 ].Fewstudiesinthepasthaveobservedacorrelationbetweenchangesinthesurfacemorphologyoftherecordingsiteduetocorrosionanddegradationoftherecordedelectrophysiologicalsignals.Thisbehaviorismorepronouncedintungstenmicroelectrodes.StudiesconductedbySanchezet.al.[ 139 ]noticedvisiblestructuralchangesonthesurfaceofthetungstenmicroelectrodeafterfourweeksofin-vivoimplantation,suggestingpossiblecorrosionofthemetal.In 82

PAGE 83

addition,theresultsshowedareductioninthepeak-to-peakamplitude(from12%-68%)oftherecordedsignalin10outof14neuronsoverthedurationoftheimplant. Theinuenceoftungstencorrosionontherecordingpropertiesisnotstraightforward.StudiesconductedbyPatricket.al.[ 4 ],foundcorrosionoftungstenmetalafter42daysofin-vivoimplantation.However,thesignalqualitydidnotshowmuchdegradationduringtheimplantedperiod.Theauthorsinferthatthegoldcoatingonthetungstenmicrowireremainedintactandmaintainedagoodelectricalcontactbetweentheneuronsandtheexternalsystem.However,therecordingelectrodefailedafter42days,andtheroleofcorrosionoftungstencannotbeoverlooked.Afollow-upstudyconductedbyPatricket.al.[ 58 ]showedthattungstenissensitivetoelectrochemicalreactionsandhasacorrosionrateof300-700m/yearinin-vitroconditionsandacorrosionrateof100m/yearinin-vivoconditions.Resultsfromthesamestudyhaverecommendedplatinumasabetterchoicefortherecordingsitematerialbecauseofitsbettercorrosionresistanceinbiologicalenvironments. 4.4.2ImprovedRecordingUsingConductivePolymerandCarbonNanotubeCoatedNeuralElectrodes Forthepastfewyears,therehasbeenanincreasedinterestoncarbonnanotube(CNT)andconductivepolymercoatedneuralelectrodes.Theimprovedelectrochemicalproperties,lesscomplicatedfabricationprocesses,tissue-friendlynature,andtheirdurabilitymakethempromisingcandidatesforcoatingonneuralelectrodes.AnumberofstudieshavebeingundertakentoreassurethesuperiorrecordingqualitiesofCNTcoatedelectrodes.Websteret.al.investigatedthepossibilityofcarbonnanobercoatedpolyurethaneasapotentialneuralororthopedicimplant[ 140 ].Theyobservedincreasedneuralattachmentsanddecreasedimmunecellsproliferationinthecarbonnanobercoatedimplants.StudiesconductedbyLovatet.al.[ 141 ],Gabayet.al.[ 142 ],andKeeferet.al.[ 143 ]alsoconrmedimprovedneuraltissueadhesiontotheCNTcoatings,aswellasdemonstratedbetterelectrochemicalproperties[ 142 ].Keeferet.al 83

PAGE 84

showedthatCNTcoatingontungstenandstainlesssteelwireelectrodesimprovedtheelectrophysiologicalrecordingsbydecreasingtheimpedanceandincreasingthechargetransfer[ 143 ]. SimilartoCNT,conductivepolymerssuchasPEDOT(poly(3,4-ethylenedioxythiophene))andPolypyrrole(Ppy)arealsogarneringmuchinterestaspotentialcoatingmaterialsforneuralrecordingelectrodes.Studiesshowthattheincreaseinsurfaceareacausedduetothecoatingofthesepolymerswillresultinadecreaseinimpedancebytwoordersofmagnitude[ 144 145 ]andanimprovementinsignal-to-noiseratioandchargeinjection[ 146 147 ].Itwasalsoobservedthatthesepolymercoatedelectrodeshavebetterelectrochemicalstability[ 144 ].Thesepropertiesarehighlydesirableforareliablechronicneuralrecordingelectrode. 4.5Summary Variationinthesurfacemorphologyoftheelectroderecordingsiteandinsulationduetocorrosionanddelaminationisoneofthemajorcomponentspotentiallyresponsibleforthemechanismofchronicelectrodefailure.Inconsistenciesinelectrodeimpedancecausedbyunstablerecordingsiterequirein-depthunderstandingandattentionfordesigningelectrodeswithimprovedrecordingefcacy.Thischapterprovidesthebackgroundinformationneededtounderstandthesecondpartofthisresearchworkwhichisfocusedonstudyingtheeffectofrecordingsiteandinsulationvariationonimpedanceandinvestigatingnewdielectricmaterialsforstabilizingtheelectrode.Thebackgroundonelectrodeimpedance,performanceofcommonlyemployedinsulationmaterials,andobservationsfromexistingSEMimagesandin-vivoimpedanceresultsandsurfaceroughnessanalysisarediscussed.Also,theeffectofrecordingsitecorrosiononimpedanceandthemeanstoreducetheimpedanceusingconductivepolymerandCNTcoatingareintroduced. 84

PAGE 85

CHAPTER5COMSOLFINITEELEMENTMODELINGSTUDYOFTHEEFFECTOFELECTRODESURFACEVARIATIONSONIMPEDANCE 5.1Introduction AsseeninChapter 4 ,theimpedanceofachronicallyimplantedrecordingmicroelectrodeisinuencedbothbybioticandabioticfactors.Asystematicstudyidentifyingtheextentoftheimpactoftheindividualcomponentsonimpedanceisrequiredforabetterunderstandingofthetemporalvariationoftheimpedance.Furthermore,theanalyticalequationforbulkelectrolytegiveninEquation 4 isonlyforaperfectdiscelectrode.Theinuenceofelectrodegeometricalchangesonelectrolyteresistanceneedstobequantied.ThischapterdescribesaCOMSOLniteelementmodelingstudyoftheeffectofelectrodesurfacevariationsonbulkelectrolyteimpedanceanddiscussestheresultsinrelationtotheobservationsfrompriorin-vivoimpedancemeasurements.Ultimately,amodelexplainingtheimpactofelectrodesurfacemorphologicalvariationsontheelectrolyteimpedanceisestablished. Theorganizationofthischapterasfollows.Themodelforelectrodesurfacemorphologybasedonlong-termin-vivocharacterizationandin-vitrocorrosionstudiesisexplainedinSection 5.2 .TheCOMSOLniteelementmodelingparametersfortwocommoncasesofsurfacevariations,corrosionwithinsulationdelaminationandcorrosionwithinsulationcracks,aredescribedinSection 5.3 .TheresultsfromCOMSOLsimulationarepresentedinSection 5.4 .Finally,thechapterissummarizedinSection 5.5 5.2ModelforElectrodeSurfaceMorphologybasedonLong-TermIn-vivoCharacterizationandIn-vitroCorrosionStudies Ourmodelforelectrodesurfacemorphologywasdevelopedbasedonourknowledgeobtainedfrompriorin-vitro[ 58 ]andin-vivo[ 121 ]impedancestudies.Itcomprisesmorphologicalchangestothemetal,tungsten,viacorrosionandtothepolymerinsulationviacrackformationanddelamination. 85

PAGE 86

Theelectrodeundergoescontinuouscorrosionduringtheimplantedperiod.However,thecorrosionratemaybehighlydynamic.Therateofcorrosionisinuencedbythepresenceofchemicalspeciessuchashydrogenperoxide(H2O2).Anin-vitrocorrosionstudy[ 58 ]showsthatAuplatedtungstenwirescorrodeatarateof10,000-20,000m/yrinphosphatebufferedsaline(PBS)inthepresencemillimolarconcentrationsofH2O2,whiletheratedropsto300-700m/yrintheabsenceofH2O2. Inthein-vivosetting,theimmuneresponsegivesanindicationforhowcorrosionmaytemporallyevolve.Afewhoursafteranelectrodeisimplanted,theinammatoryresponseactivatestheimmune-responsemicroglialcells.ThesereactivemicroglialcellsproducesuperoxidesandaddsunknownconcentrationsofH2O2totheCSF[ 148 ].Duringtherstfewweeksofimplantation,weassumethatthereisahigherconcentrationofH2O2,whichleadstoahigherrateoftungstencorrosion.Whentheinammationresponseprogressestothegliosisphase,thenumberofsurroundingglialcellsdecreasesandtheglialsheathbecomesmorecompact[ 49 ].Hence,weassumetherateofcorrosiondecreasesbecause1)thereisasmalleramountofH2O2beingproducedbytheremainingglialcells,and2)thedenseglialsheathmayimpedediffusionofreactingspecies,therebymakingthecorrosionratemass-transferlimited.Thusfar,ourmodelfortemporaltungstenmorphologyisthattheexposedtungstenwilluniformlycorrode(ascomparedtopittingcorrosion)assoonasitisimplantedwitharatethatwillincreasewithtimeuptoformationofthecompactglialsheathandthendecreasetoaconstantvalue. Toincorporateamodelforelectrodeinsulationdegradation,weuseresultsfrom[ 121 ].TheSEMimagesofexplantedelectrodesshowthatinsulationdamageisobservedonlyinelectrodesthathavebeenimplantedforachronicperiod(42daysormore),andnonoticeabledamageormodicationisobservedinacuteorrecoveryanimals[ 137 ].Thissuggeststhattheinsulationremainsintactduringtheearlyweeksofimplantationandundergoessomemodicationafteracertainperiodontheorderof 86

PAGE 87

40days.However,therateofmodicationisunknownatthispoint.TheSEMimagesofpost-implantelectrodesshowsignsofcracksintheinsulationaswellasdelaminationatthesurfaceoftheelectrode,tovariousdegrees. Thus,thegeometricalmodelweuseintheCOMSOLanalysisiscomprisedofacylindricalgold-cladtungstenelectrodeinsulatedinapolymerwhosetungstensurfacewillrecessintothepolymerinsulationaccordingtoourmodelfordynamiccorrosionratewhiletheinsulationwillformcracksanddelaminateovertime.Howthesemorphologicalchangesaffectthemeasuredin-vivoelectrodeimpedanceindependentofbioticcontributionsistheultimatequestion.Thischapterapproximatestheanswertothisquestionbymakingsomeassumptionsregardingthefrequencydependentnatureoftheelectrode/tissueinterface. Basedontheabovemodelofmorphologicalchanges,weproposeourhypothesisforimpedancetrendasfollows;Initialbioticeffectsdominatetherstphaseofhighcorrosionrateoftherecordingsite,thenabioticeffectsdominatethesecondphaseofsimultaneouscorrosionandinsulationdamage(atconstantrate)withasaturationinbioticimpedancewhichresultsintheriseandsubsequentfalloftheimpedancevaluesofimplantedtungstenelectrodes.Figure 5-1 showstheschematicillustrationoftheproposedhypothesis. 5.3COSMOL3DModel Finiteelementanalysisoftheelectrode-electrolyteinterfacewasconductedusingtheCOMSOLMultiphysicssoftwarepackage.Thecylindricalelectrodeandthesurroundingelectrolyteregionweremodeledin3D.AC/DCmodule(orelectriccurrentphysics)wasselectedasthephysics.Themodelincludeda50mdiametertungstenmicrowiresurroundedbya1mthickgoldlayerandinsulatedwith5mthickpolyimideinsulationsurroundedbyaregionofsalineelectrolyte.Theelectrolyteboundarieswereextendedupto1cmx1cmtosatisfythesemiinniteboundarycondition.Apointcurrentsourceof1Awasplacedatthecenterofthetop(cortexboundary)surfaceof 87

PAGE 88

Figure5-1. Schematicillustrationoftheproposedhypothesisexplainingthein-vivoimpedancetrend.Insetshowsthebioticandabioticimpedancecomponentsintheequivalentcircuit. theelectrode.Figure 5-2 showsanillustrationoftheCOMSOLniteelement3Dmodeloftheelectrode-electrolyteinterface. 5.3.1ModelParameters Theelectrodeisassumedtobeauniformconductorandhencethepotentialdistributionacrosstheelectrodewillbeconstant.ADirichletboundaryconditionofV=0isassumedattheelectrolyteboundariessincetheboundariesaresupposedtobesemi-innite,whileaVonNeumannboundaryconditionofn.J=0isassumedatthecortexboundary,whichensureselectricalinsulationattheboundaries.The 88

PAGE 89

Figure5-2. IllustrationoftheCOMSOLniteelement3Dmodeloftheelectrode-electrolyteinterface.Right:Zoomedinimageofthegoldplatedtungstenmicrowireandpolyimideinsulationmodel. conductivitiesofdifferentmaterialsusedforthemodelaregiveninTable 5-1 .Tetrahedralelementswithsizerangingfrom1mto400mwereusedformeshingthedomains.Figure 5-3 showsanillustrationoftheboundaryconditionsusedintheCOMSOLniteelementmodel. Table5-1. ConductivityvaluesusedforCOMSOLniteelementmodeling. MaterialConductivity(S/m)) Tungsten18.94106Gold40.98106Polyimide110)]TJ /F3 7.97 Tf 6.59 0 Td[(6Saline1.8 89

PAGE 90

Figure5-3. IllustrationoftheboundaryconditionsusedintheCOMSOLniteelementmodel. Twocasesofelectrodesurfacevariationswerestudiedindetail:(i)theeffectofrecordingsitecorrosionandinsulationdelamination/peelingonimpedance,and(ii)theeffectofrecordingsitecorrosionandinsulationcrackingonimpedance. 5.3.2Case1:EffectofRecordingSiteCorrosionandInsulationDelaminationonImpedance Fortherstcase,theinsulationdelaminationwasmodeledasagapbetweentheelectrodeandtheinsulation.Impedancewascalculatedatsevendifferentinstances,eachrepresentingadifferentconditionofsurfacevariationthatmaybeobservedinanimplantedelectrode.Therstinstancewasapristinewirewithnocorrosionorinsulationdelamination.Impedancecalculatedatthisinstanceservedasthebaselinereferenceforthesuccessiveinstances.Thenexttwoinstancesmodelthehighcorrosionphaseoftheelectrode.Nomodicationininsulationwasincludedinthesetwoinstances.Thenalfourinstancesmodelthesimultaneouscorrosionandinsulationdelaminationphase.Arateof15mperinstancewasassumedformodelingthehighcorrosionrate,whilea 90

PAGE 91

Figure5-4. IllustrationofCase1electrodesurfacemodicationoverimplantduration.A)pristineelectrodewithintactmetalandinsulation,B)electrodewithcorrodedmetalandnoinsulationdelamination,andC)electrodewithcorrodedmetalandnoticeableinsulationdelamination.Inseton(B)and(C)showsacloserviewofthegoldlayeraroundthetungsten. rateof0.5mperinstancewasassumedforthelowcorrosionrate.Thisratioof15:0.5waschoseninaccordancewiththeexperimentallyestimatedcorrosionrateratioof10,000-20,000:300-700torepresentcorrosioninthepresenceandintheabsenceofhydrogenperoxide.Aconstantrateof2mperinstancewasassumedfortheincreaseininsulationgapwidthandlengthinthenalfourinstances.Figure 5-4 showstheillustrationsofthepristineelectrode,corrodedelectrode,andcorrodedelectrodewithinsulationdelaminationorpeeling. 5.3.3Case2:EffectofRecordingSiteCorrosionandInsulationCrackingonImpedance Similartotherstcase,theimpedancewascalculatedatsevendifferentinstancesfortheinsulationcrackedcase.Asbefore,therstinstancewasapristinewireservingasthebaselinereferenceforthesuccessiveinstances,andthenexttwoinstancesmodelthehighcorrosionphaseoftheelectrode,andthenalfourinstancesmodelthe 91

PAGE 92

Figure5-5. IllustrationofCase2electrodesurfacemodicationoverimplantduration.A)pristineelectrodewithintactmetalandinsulation,B)electrodewithcorrodedmetalandnoinsulationcrack,andC)electrodewithcorrodedmetalandnoticeableinsulationcrack.Insetontheimagesshowsacloserviewofthegoldlayeraroundthetungsten. simultaneouscorrosionandcrackingininsulationphase.Thesamerateof15mperinstancewasassumedformodelingthehighcorrosionrate,whilearateof0.5mperinstancewasassumedforthelowcorrosionrate.Aconstantrateof2.5mperinstancewasassumedfortheincreaseininsulationcracklengthandthewidthofthecrackwassetat15,30,60and90forthenalfourinstances.Figure 5-5 showstheillustrationsofthepristineelectrode,corrodedelectrode,andcorrodedelectrodewithinsulationcracks. 5.4ResultsandDiscussion ThissectionwilldescribetheCOMSOLniteelementmodelingresultsforbothcases.Themodelingresultsarecomparedwithin-vivomeasurementsforfurthervalidation.Dailyin-vivoimpedancedatafromonelong-termimplantedrepresentativeelectrodewasusedforthecasestudypurpose.AdetaileddescriptionofthecasestudyisgiveninSection 6.5.2 92

PAGE 93

COMSOLModelingResults:COMSOLsimulatedpercentagechangeinimpedanceforcorrosionwithinsulationcracksandcorrosionwithinsulationdelaminationcasesisgiveninFigure 5-6 .AlsothedimensionsofmetalrecessionandinsulationdelaminationandcracksaregiveninthemiddleandthebottomplotsofFigure 5-6 .ItcanbeobservedfromthetopplotinFigure 5-6 thatthepercentchangeintheelectrolyteimpedance(Re)increasessteadilyduringtheinitialhighcorrosionratephaseanddecreasesduringthelowcorrosionandinsulationdamagephase,givingabellshapeforboththeinsulationdelaminationandcrackcases.However,itcanbeobservedthattheincreaseinimpedanceisonlyupto14%andthedecreaseinimpedanceisupto7%.Thisisnearlyoneorderofmagnitudelessthantheobservedin-vivoimpedancechangesasshownintheFigure 6-13 .ThisimpliesthatthecontributionofRetotheoverallimpedanceinanin-vivoconditionissmall.However,thissimulationstudywasdoneforonlyfewtypesofsimplesurfacevariations,andinrealitymorecomplexvariationssuchasrougheningofthesurfacecanoccur.Undersuchcomplexsurfacevariations,thecontributionofRecanbedifferentthanourestimate.ItshouldalsobenotedfromthetopplotofFigure 5-6 thatthesolutioncouldnotbeobtainedforinstances4and5forinsulationdelaminationcasesincetheFEAsolvercouldnotgenerateameshforverysmallvariationsingeometry. 5.5Summary TheeffectofelectrodesurfacevariationsincludingrecordingsitecorrosionandinsulationdamagesonelectrolyteimpedancewasextensivelystudiedbydevelopingnumericalmodelsinCOMSOLMultiphysics.Themodelingresultsshowedthattheelectrolyteresistance(Re)isinuencedbythechangesinthegeometryoftheelectrodesurface.Thesimulationresultsshowedanincreaseintheelectrolyteresistanceastherecordingmetalrecessesduetocorrosionandadecreaseinresistanceasmoremetalisexposedduetoinsulationdelaminationandcracking.TheincreaseinRecouldbeduetotheincreaseindriftdistancecausedbytherecessionofthemetalandthedecrease 93

PAGE 94

Figure5-6. COMSOL3Dsimulatedresults.A)Percentagechangeinelectrodeimpedanceforcorrosionwithinsulationcrackcaseandcorrosionwithinsulationdelaminationcaseatdifferentinstances,B)Variationinthemetalrecessionlengthatdifferentinstances,andC)Variationofinsulationcrackwidthandlength,anddelaminationdimensionsatdifferentinstances. 94

PAGE 95

inRecouldbeduetotheincreaseinsurfaceareacausedbytheincreasedexposureofmetal.ThecalculatedpercentagechangeinRe,althoughsignicant,issmallerthanthatseeninin-vitroandin-vivomeasurements,suggestingthatthecontributionfromtheelectrolyteimpedancetotheoverallin-vivoimpedancemaybesmall.However,undercomplexsurfacevariationssuchaschangesinthemicroscaleroughnessontherecordingsitetheresultsmayvary.Furthermore,basedonourpreviousin-vivoimpedancestudy,amodeltodescribethein-vivoimpedancebehaviorwasdeveloped,whichcanserveasascienticbasisfordescribingtheinuenceofelectrodesurfacevariationonimpedance. 95

PAGE 96

CHAPTER6IN-VITROEXPERIMENTALEVALUATIONOFTHEEFFECTOFELECTRODESURFACEVARIATIONSONIMPEDANCE 6.1Introduction SincetheCOMSOLFEAfocussedonlyontheelectrolyteresistance,afollowupexperimentalinvestigationwasneededtounderstandtheeffectofthedoublelayerimpedanceontheelectrode-electrolyteimpedance.Inthischapter,themodeldescribedinChapter 5 wasvalidatedthroughexperimentsconductedinanin-vitrosetup.Tungstenmicrowiresampleswereengineeredforcommonlyobservedsurfacevariationsusingcontrolledcorrosionandionmillingmethodsandmeasuredforimpedancechangesin-vitro.Theresultswerecomparedwithin-vivoimpedancedata. Thischapterisorganizedasfollows.Thein-vitroexperimentalplanandthestepsinvolvedaredescribedinSection 6.2 .ThemethodsfordefectengineeringthetungstenmicrowiresusingcontrolledcorrosionandFIBionmillingandthecharacterizationoftheengineeredwiresusingSEMimagingprocedureareexplainedinSection 6.3 .DetailsoftheimpedancemeasurementmethodsincludingthesamplepreparationandcalibrationoftheNanoZrinstrumentaregiveninSection 6.4 .Theresultsalongwithathoroughcomparisonwithin-vivoobservationsareprovidedinSection 6.5 .ThechapterissummarizedinSection 6.6 6.2ExperimentalDesign Thissectiondescribestheexperimentalplanthatwasfollowedtovalidateourhypothesisforabioticfactorsaffectingelectrodeimpedanceusingin-vitromethods.Adualpathexperimentalschemewasperformedinparalleltostudytheimpactofcorrosionandinsulationcracksontungstenelectrodeimpedance.Thedualpathwasadaptedtostudytheeffectofeachkindofsurfacevariationindependently.Theexperimentalstepsincludeddicingthewiresusingadicingsaw,SEMimaging,impedancemeasurementusingNanoZrinstrument,andFIBsurfacemodicationsorcontrolledcorrosionoftungstenmicrowiresamplesfollowedbyfurtherSEMimagingand 96

PAGE 97

Figure6-1. FlowchartofthestepsinvolvedinIn-vitroimpedancemeasurementstudyofcorrodedwires. impedancemeasurements.Twosamplesweretestedineachstudy.Thecorrosionstudysamplesweregiventhenamessample2andsample4,whilethenomenclatureusedfortheFIBinsulationcrackstudysamplesweresample3andsample5respectively.Figures 6-1 and 6-2 listthestepsinvolvedinthecorrosionandinsulationdamageexperimentaldesign. 6.3MicrowireSurfaceModicationsusingFocusedIonBeamandCorrosion Thissectionprovidesadescriptionofthemethodsusedforcreatingthesurfacemodicationsintungstenmicrowiresamples.Recessionoftheexposedmetalwasachievedbyadaptingacontrolledcorrosionprocess,whichisexplainedinsubsection 6.3.1 .ThecracksininsulationwereengineeredviaionmillingofthepolyimideusingaFocussedIonBeam(FIB)tool,whichisdescribedinsubsection 6.3.2 .Also,theprocedurefollowedforimagingthemodiedwiresisdescribed. 97

PAGE 98

Figure6-2. FlowchartofthestepsinvolvedinIn-vitroimpedancemeasurementstudyofinsulationcrackedwires. 6.3.1EngineeringMetalRecessionusingControlledCorrosion Controlledcorrosionmethodwasemployedtocreatearecessionintherecordingmetalsiteofthetungstenmicrowiresamples2and4.Thestartingpristine50mdiameter,3cmlongtungstenmicrowireswerefabricatedusingdicingsawtocuttherecordingend.Thetipsofthepristinetungstenmicrowireswereimmersedina0.9%phosphatebufferedsaline(PBS)solutionforatotaltimeperiodofaround194hoursor8days.Sufcientcarewastakennottosoakthewholewireinthesalinesolution.ThemicrowiresampleswereremovedfromthePBSsalineafter64hours,106hoursand194hoursofimmersionandimagedusingaSEMtoassessthechangesinthesurfacemorphologyatdifferenttimepoints.Fromliterature,itisestimatedthatthewireswillberecessedabout3.5mafter64hoursofimmersioninPBS,about6mafter106hours, 98

PAGE 99

andabout11mafter194hoursofimmersioninPBSsolution[ 58 ].Inaddition,theimpedancewasmeasuredateachtimepointusingaNanoZrimpedancemeter. 6.3.2EngineeringInsulationDamagesusingFocussedIonBeam(FIB) Agalliumionbasedfocussedionbeam(FIB)(FEIDual-BeamStrataDB235)wasusedtoetchpatternsonthepolyimideinsulationandtocreatewelldenedinsulationcracksontwodicingsawcuttungstenmicrowires. Rectangularpatternsofdimension15mx10mweremilledonsamples3and5.Theetchratedependeduponthebeamcurrentintensity.Qualitatively,forahigherbeamcurrent,theetchratewasfast,buttheresolutionwaspoorleadingtodiffusedpatternboundaries.Theetchratewasslowforlowcurrentdensities,buttheresolutionwasgoodandthepatternhadsharpboundaries.Inordertostrikeabalancebetweentheetchtimeandresolution,thepatternwasetchedoffwithahighbeamcurrent(20,000pA)atthebeginningandslowlythecurrentwasreducedtoalowervalue(40pA)forimprovingthebeamfocusandnetuningthepattern.Withthistechnique,theaveragetimetakentomill5mdeeppatternsontheinsulationsurfacewasaround15minutes. 6.3.3ElectronMicroscopyImagingofSurfaceModiedMicrowires MicrowiresampleswithengineeredsurfacedefectswerecharacterizedusingSEMimagesaftereachsurfacemodicationstep.Thepristinetungstenmicrowiresafterbeingcutbythedicingsawwereimagedusingthevariablepressurescanningelectronmicroscope(VP-SEM).Similarly,thewiresrecessedviacontrolledcorrosionmethodswerealsoimagedusingtheVP-SEMaftereachcorrosionperiod.Imagesweretakenatlowvacuumenvironmentalsecondaryelectrondetector(ESED)modewithanaccelerationpotentialof12kVandanaverageworkingdistanceof27mm.Figure 6-3 showstheSEMimagesoftungstenmicrowiresamplenumber2atpristinecondition,after64hoursofPBSimmersion,after106hoursofPBSimmersion,andafter194hoursofPBSimmersion.Figure 6-4 showstheSEMimagesoftungstenmicrowiresamplenumber4atpristinecondition,after64hoursofPBSimmersion,after106hours 99

PAGE 100

A B C D Figure6-3. SEMimagesoftungstenmicrowiresamplenumber2.A)pristinecondition,B)after64hoursofPBSimmersion,C)after106hoursofPBSimmersion,andD)after194hoursofPBSimmersion. ofPBSimmersion,andafter194hoursofPBSimmersion.Itcanbenotedfromtheguresthatthesurfaceroughnessoftherecededmetalincreasedwiththeexposuretosalineandmoremicroscalecracksandpitsformedanddevelopedonthesurfaceofthemetal. ThewiresmodiedforinsulationcracksusingFIBwereimagedwiththeelectronbeamoftheFIBtoolwithanacceleratingvoltageof5kVandaspotsizeof3.Figure 6-5 showstheSEMimagesoftungstenmicrowiresamplenumber3atpristinecondition,aftercreatingtherstrectangularnotchusingFIB,andaftercreatingthesecond 100

PAGE 101

A B C D Figure6-4. SEMimagesoftungstenmicrowiresamplenumber4.A)pristinecondition,B)after64hoursofPBSimmersion,C)after106hoursofPBSimmersion,andD)after194hoursofPBSimmersion. rectangularnotchusingFIB.Figure 6-6 showstheSEMimagesoftungstenmicrowiresamplenumber5atpristinecondition,aftercreatingtherstrectangularnotchusingFIB,andaftercreatingthesecondrectangularnotchusingFIB. 6.4ImpedanceMeasurementusingNanoZrImpedanceMeter Thissectionprovidesadescriptionofthein-vitroexperimentalimpedancemeasurementprocedureincludingthecalibrationoftheNanoZrimpedancemeterandthepreparationofthemicrowiresampleswhoseimpedancewasmeasured. 101

PAGE 102

A B C Figure6-5. SEMimagesoftungstenmicrowiresamplenumber3.A)pristinecondition,B)aftercreatingtherstrectangularnotchusingFIB,andD)aftercreatingthesecondrectangularnotchusingFIB. 102

PAGE 103

A B C Figure6-6. SEMimagesoftungstenmicrowiresamplenumber5.A)pristinecondition,B)aftercreatingtherstrectangularnotchusingFIB,andD)aftercreatingthesecondrectangularnotchusingFIB. 103

PAGE 104

Figure6-7. PhotographoftheNanoZrNZ-CALcalibrationadaptor.FigureadaptedfromNanoZrusermanual[ 6 ]. 6.4.1CalibrationofNanoZrInstrument ThecalibrationprocedureoftheNanoZrInstrumentissimpleandstraightforward.ItinvolvesconnectingtheNZ-CALadaptertotheNanoZrandrunningthedevicecalibrationprocedurefromtheNanoZrsoftware.ThecalibrationadaptorNZ-CALincludesanarrayofresistorsandcapacitorsofdifferentimpedancevaluesasshowninFigure 6-7 .Dependingontheuserrequirement,theusercanselectthecalibrationfortheimpedancetestingmodeortheelectroplatingmode.Sinceinthisstudy,theNanoZrisprimarilyusedforimpedancemeasurement,theimpedancecalibrationmodewasselectedandperformed.Thecalibrationprocessrunsforafewminutes.Attheendofthecalibrationprocedure,itisrequiredtoruntheimpedancetestingoftheNZ-CALadaptor.Theimpedancetestinggeneratesareportwiththemeasuredimpedanceoftheresistorsintheadaptor.WhenthevaluesinthereportmatcheswiththevaluesprovidedintheNanoZrusermanual[ 6 ],withintheacceptabletolerancelimit,thenitcanbeassumedthatthedevicehasbeencalibratedandisfunctioningwithgoodaccuracy.Ifnot,thecalibrationproceduremustberepeateduntilthemeasuredimpedancevaluesmatchwiththestandardvalues.Attheendofthedevicecalibration,itwasfoundthatthemeasuredvaluesmatchedthestandardvalues. 104

PAGE 105

6.4.2MicrowireSamplePreparation Tungstenmicrowiresamplesforin-vitroimpedancemeasurementusingNanoZrinstrumentwerepreparedinthefollowingway.Pristine50mdiametermicrowireswerecutintoseveral3cmlongpieces.Theinsulationfromoneendofthedicedwiresweremanuallyremovedusingabladeexposingthemetal.Carewastakennottodamagetheotherendofthewire.Athickergaugewireofapproximately5-7cmlengthwasbondedtotheexposedendofthemicrowireusingsilverepoxy(AbleBondr84-1LMI),thusservingasaninterfacebetweenthemicrowiresampleandanexternalcable.AnotherlongwireofthickergaugewassolderedtotheinterfaceandservedasacablefortheNanoZradapter.Thesolderedcablewasremovedaftereachimpedancemeasurementexperimentandreattachedbeforethenextroundofimpedancemeasurement.Theotherendofthewireissubjectedtosurfacemodications. 6.4.3ImpedanceMeasurementProcedure Impedancemeasurementsaremadeonpolyimideinsulated50mdiametertungstenmicrowiresamplesthatarepristinewithnodamagetotherecordingmetalortheinsulationmaterialaswellasonsamplesthatareengineeredwithdefectssuchasdamagedinsulationandmodiedrecordingsurface.TheNanoZrimpedancemeter(WhiteMatter,LLC.)measurestheimpedanceusingthetwoelectrodesmethodwithaworkingelectrodeandareferenceorgroundelectrode.Theworkingprincipleinvolvestheoperationofasimplevoltagedividercircuit.TheunknownresistorisconnectedinserieswithaninternalreferenceelectrodeZrefandagenerated4mVppsinusoidalvoltageisappliedastheinputvoltage.Thevoltagebetweentheresistorsismeasured,and,usingthevoltagedivisionequation,thevalueoftheunknownresistoriscalculated.TheimpedancewasmeasuredforallmeasurablefrequenciesintheNanoZr,i.e.,1Hz,2Hz,5Hz,10Hz,20Hz,50Hz,100Hz,200Hz,500Hz,1kHz,and2kHz.Impedancemeasurementsweremadeforvetrialsoneachmicrowiresamplewitheachtrialinvolving40cyclesofmeasurements.Theexperimentaluncertaintywasanalyzedby 105

PAGE 106

calculatingthecondenceinterval.Sincethenumberoftrials(measurements)waslessthan30,thet-distributionwasusedtocalculatethecondenceinterval.Thecondenceintervalforat-distributionisgivenby, C.I.= xt=2,n)]TJ /F3 7.97 Tf 6.58 0 Td[(1s p n,(6) where x=mean,=signicancelevel(0.05for95%C.I.),n-1=numberofdegreesoffreedom,n=samplesize,ands=samplestandarddeviation,whichisgivenby s=vuut 1 n)]TJ /F12 11.955 Tf 11.95 0 Td[(1nXi=1(xi)]TJ ET q .478 w 279.79 -204.88 m 286.44 -204.88 l S Q BT /F7 11.955 Tf 279.79 -212.2 Td[(x)2.(6) ThesampletungstenmicrowireisattachedtoathickerwireusingconductivesilverepoxyandinterfacedwiththeNanoZrwithanadapter(NeuralynxInc.).Thisservesastheworkingelectrode.Anotherwirewithasurfacearealargerthanthetungstenmicrowireisconnectedtothegroundpinoftheadapterandservesasthereferenceelectrodeandclosestheloop.Bothwiresareimmersedina90%phosphatebufferedsaline(PBS)solution.Sincethetestsignalisalowamplitudesinewaveandnoiseduetoelectromagneticinterference(EMI)couldpotentiallyaffecttheinputsignalleadingtoerroneousmeasurements,themeasurementsweremadeusingabatterypoweredlaptopinsteadofanACsupplyconnectedcomputer.Figure 6-8 showsthephotographoftheexperimentalsetup. 6.5ResultsandDiscussion Thissectiondescribesthein-vitroexperimentalresults.NanoZrmeasuredimpedancevaluesinsalineenvironmentarecomparedwithin-vivomeasurements.Similartosection 5.4 ,dailyin-vivoimpedancedatafromonelong-termimplantedrepresentativeelectrodewasusedforcasestudypurpose. 106

PAGE 107

Figure6-8. In-vitroimpedancemeasurementexperimentsetup.A)Photographshowingthein-vitroimpedancemeasurementexperimentsetup,B)EnlargedimageoftheNanoZrimpedancemeterandtheadapter. 6.5.1In-vitroImpedanceMeasurementResults 6.5.1.1ImpedanceMeasurementonPhosphateBufferedSaline(PBS)CorrodedMicrowires Thefrequencyspectrumplotsoftheaverageimpedancemeasuredfromthetwocorrosionstudytungstenmicrowiresamples2and4areshowninFigures 6-9 and 6-10 .Itcanbeobservedfromtheplotsthatathigherfrequenciestheimpedanceofthe3.6mcorrodedwiresishigherthanthepristinewire.At1kHz,itwasmeasuredthattheaverageimpedanceincreasedbyalmost45%inbothsamples2and4after64hoursofPBSexposure.Thiscouldbeduetotheinitialincreaseinthedriftdistancecausedbytherecessionofthemetal.However,theaverageimpedancedecreasedinthesubsequentmeasurements.Itwasfoundthattheaverageimpedanceat1kHzdecreasedby32%insample2and63%insample4afterthewireswereimmersedinPBSsolutionforadditional42hours(106hoursintotal).Similarly,theaverageimpedanceat1kHzdecreasedby35%insample2and63%insample4afterthewireswereimmersedinPBSsolutionforanadditional88hours(194hoursintotal).Thisdecreaseinimpedanceisclearlysupportedbytherapidincreaseinthemicroscale 107

PAGE 108

Figure6-9. Averageimpedancemeasuredfromsamplenumber2beforeandafterimmersioninPBS. surfaceroughnessontherecededmetalasseenintheSEMimages.TheincreasedsurfaceroughnessincreasestheeffectiveareawhichdecreasesRe.ItalsomayaltertheCPEbehaviorwhichisalsoreectedasadropinimpedance.ItshouldbenotedthatthepercentagedecreaseinimpedanceismuchhigherthanthatofReasseenintheCOMSOLmodelingresults.Thisimpliesthatthecontributionofthedoublelayerimpedanceontheoverallelectrodeimpedancemaybesignicantatthemeasurementfrequencies. 6.5.1.2ImpedanceMeasurementonFIBInsulationModiedMicrowires Thefrequencyspectrumplotsoftheaverageimpedancemeasuredfromthetwoinsulationcrackstudytungstenmicrowiresamples3and5areshowninFigures 6-11 and 6-12 .Itisobservedfromtheplotsthatthepercentagechangeinimpedance 108

PAGE 109

Figure6-10. Averageimpedancemeasuredfromsamplenumber4beforeandafterimmersioninPBS. betweeneachsurfacemodicationstepissmallforthisparticularmodication.Forinstance,insample3,theaverageimpedanceat1kHzisreducedby2.5%aftertherstengineered15m10mcrackandbyanother6.5%afterthesecondcrack.However,insample4,theaverageimpedanceat1kHzisdecreasedbynearly17.5%aftertherstcrackandbyanother14%afterthesecondcrack.Thoughthisishigherthanwhatwasobservedinsample3,thesenumbersarestilllessthanthecorrosioneffects. 6.5.2ComparisonwithIn-vivoResults Inordertocheckourhypothesis,thenumericalandexperimentalresultswerecomparedwiththeimpedanceresultsobtainedfromapreviouslongtermin-vivostudy.Theexperimentalmethodsandresultsofthestudyarereportedin[ 121 ].Inthatstudy, 109

PAGE 110

Figure6-11. Averageimpedancemeasuredfromsamplenumber3beforeandafterFIBinsulationmodication.Insetshowsthemeasuredimpedanceat1kHzfrequency. dailyimpedancemeasurementsandneuralsignalsignalrecordingsweremadeon22rats,fourofwhichwereimplantedforacutephase(fewhoursaftersurgery),vewereimplantedforrecoveryperiod(uptotwoweekspost-implant),and13wereimplantedforchronicphase(upto6monthspost-implant).Dailyimpedancemeasurementsmadeupto21weeksshowedatimevaryingtrendintheimpedancewhichwasrelatedtotheelectrodeperformance.Thetimevaryingimpedancetrendhadabellshapedprolewithalongtail.Thestudyalsorevealedthattheelectrodeperformancewasatitsbestwhentheimpedancewasbetween40and150k.Amongthe13chronicelectrodearrays,electrodeR9waschosenfortheanalysisasitwasimplantedfor217daysandprovidedstablerecordingduringtheentireimplantedduration. 110

PAGE 111

Figure6-12. Averageimpedancemeasuredfromsamplenumber5beforeandafterFIBinsulationmodication.Insetshowsthemeasuredimpedanceat1kHzfrequency. Comparisonbetweenthepre-implantandthepost-implantSEMimagesofR9revealedcorrosioninall16wiresandvisibleinsulationdamageincludingdelaminationandcracksin50%oftheelectrodes.Surfaceroughnessanalysiswasperformedonthewiresusinga3Dlaserscanningmicroscope(KeyencerVK-9700,METLABSDSU)tomeasurethedepthofcorrosionineachwire.Themeasuredaveragecorrosiondepthforall16microwiresinR9was38.8m,respectively. Fromthemeasuredin-vivoimpedancedata,thepercentagechangeinimpedancewascalculatedforeachindividualwireandaveragedforallwiresovertheimplantedduration.Theimpedancevaluemeasuredonthesurgerydaywasconsideredasthereferencevalue,andthepercentagechangewascalculatedcomparedtothisreference. 111

PAGE 112

Figure6-13. Percentagechangeintheaveragein-vivoimpedanceforallwiresofelectrodeR9plottedagainsttheimplanteddurationandttedwithaGaussiancurve. Sincetheshapeofthein-vivoimpedancetrendappearstohaveaGaussianprole,thecalculatedpercentagechangeinimpedancevalueswerettedwithaGaussiancurveusingthebuilt-inGaussiancurvettingtoolinOriginPror.Valueshigherthan600%weretreatedasoutliersandwerenotincludedintheplot.TheequationfortheGaussiancurveisgivenby, y=y0+A wp 2e)]TJ /F3 7.97 Tf 6.58 0 Td[(2(x)]TJ /F14 5.978 Tf 5.76 0 Td[(xc)2 w2,(6) wherewisthestandarddeviation,xistheindividualcalculatedvalue,xcistheaverage,andAandy0areconstants.Figure 6-13 showsthepercentchangeinaveragein-vivoimpedanceforallwiresofelectrodeR9plottedagainsttheimplanteddurationandttedwithaGaussiancurve. 112

PAGE 113

ItcanbenotedthattheGaussiancurvetsthedatawellwithanR-squarevalueof0.639.Figures 6-14 6-15 6-16 and 6-17 showsthepercentchangeinthein-vivoimpedanceforeachindividualwireofelectrodeR9plottedagainsttheimplanteddurationandttedwithaGaussiancurve.AlsoshownarethepreandpostimplantSEMimagesandthecorrosiondepthprolemeasuredusingthelaserscanningmicroscopeforeachwire.Ofthe16wires,6wireshadagoodt(i.e.,R-square>0.5),while5wireshadfairlyacceptablet(i.e.,R-squarebetween0.25to0.5)and4hadpoort(i.e.,R-square'0).TheR-squarevalueforeachcurveisgiveninthegure.ThepercentagechangevalueisaboutoneorderofmagnitudehigherthantheCOMSOLcalculatednumericalvaluesandnearlytwicethemeasuredin-vitroimpedancechanges.However,thepeakpercentchangevalueisnotthesameinallwires,anditrangesbetween100-600%. WhenthebeforeandafterSEMimagesofthewireswithpoorGaussiant(wirenumber4,6,9and16)wereexamined,itwasfoundthatallthesefourwireshadrecordingsitecorrosion.Also,thelaserscanningdepthproleshowedthatthecorrosiondepthofthesewiresaresimilartothoseobservedintherestofthewiresinthearray.Sincethereisnoabnormalityincorrosion,whichistheprimaryabioticfactorinuencingtheimpedancechanges,itisdeemedthattheanomalousimpedancebehaviorofthesewiresarepotentiallycausedbysomecomplextissuechangesduringtheimplantedperiod. 113

PAGE 114

A B C D Figure6-14. Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforindividualwiresofelectrodeR9.A)wire1,B)wire2,C)wire3,D)wire4.Thepre-implantandpost-implantSEMimagesalongwiththelaserscanningdepthproleareshownintheinset. 114

PAGE 115

A B C D Figure6-15. Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforindividualwiresofelectrodeR9.A)wire5,B)wire6,C)wire7,D)wire8.Thepre-implantandpost-implantSEMimagesalongwiththelaserscanningdepthproleareshownintheinset. 115

PAGE 116

A B C D Figure6-16. Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforindividualwiresofelectrodeR9.A)wire9,B)wire10,C)wire11,D)wire12.Thepre-implantandpost-implantSEMimagesalongwiththelaserscanningdepthproleareshownintheinset. 116

PAGE 117

A B C D Figure6-17. Percentagechangeinin-vivoimpedanceplottedagainsttheimplanteddurationandttedwithaGaussiancurveforindividualwiresofelectrodeR9.A)wire13,B)wire14,C)wire15,D)wire16.Thepre-implantandpost-implantSEMimagesalongwiththelaserscanningdepthproleareshownintheinset. 117

PAGE 118

Inordertounderstandtheindividualcontributionsofthebiotic(ortissue)andabiotic(orsurfacevariation)componentstotheoverallin-vivoimpedance,thereal(ortheresistive)andtheimaginary(orthereactive)partsofthecompleximpedancewereseparatedandplottedasafunctionoffrequencyforthreewires(wire1,3and7)intheR9electrodearray.SixtimestampswerechosentorepresentpointsfromthethreesegmentsoftheGaussianprole,namely,therisingphase,thefallingphase,andthelongtail.Thetimestampschosenwerepre-implantday,day1ofimplant,day7postimplant,days25and31ofpostimplant,day85postimplant,anddays175,177,and205postimplant.Similarly,therealandimaginarypartsofthecompleximpedancewereplottedagainstfrequencyforin-vitrotestsamplesnumber2andnumber3.Itshouldberecalledthatthein-vitroexperimentsdidnothaveanybioticcomponentassociatedwithit,andthereactanceorimaginarypartoftheimpedanceissolelyduetothedoublelayerasseeninFigure 4-4 .Ontheotherhand,thein-vivomeasurementsincludetissueimpedance,whichalsocontributestothereactanceortheimaginarypartoftheimpedanceasseeninFigure 4-5 .Inboththecases,thebulkelectrolyteimpedanceispurelyresistive.Figures 6-18 6-19 ,and 6-20 showthefrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire1,wire3,andwire7respectively.Figures 6-21 and 6-22 showthefrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofcorrodedtungstenwiresamplenumber2andtheFIBmodiedtungstenwiresamplenumber3.Figures 6-23 6-24 ,and 6-25 showtherealandimaginarypartsoftheimpedanceofR9wire1,wire3,andwire7at1kHzfrequencyasafunctionofimplantduration.Figures 6-26 and 6-27 showthetherealandimaginarypartsoftheimpedanceofcorrodedtungstenwiresamplenumber2andtheFIBmodiedtungstenwiresamplenumber3at1kHzfrequencyasafunctionofthedifferentsurfaceconditions. Itcanbeobservedfromthein-vivorealandimaginaryimpedanceplotsofallthreewiresthat,athigherfrequencies(>500Hz),theimpedancecurvetendstocurve 118

PAGE 119

A B Figure6-18. FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire1.A)ResistivecomponentandB)Reactivecomponent. 119

PAGE 120

A B Figure6-19. FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire3.A)ResistivecomponentandB)Reactivecomponent. 120

PAGE 121

A B Figure6-20. FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofR9wire7.A)ResistivecomponentandB)Reactivecomponent. 121

PAGE 122

A B Figure6-21. Frequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofcorrodedtungstenwiresample2.A)ResistivecomponentandB)Reactivecomponent. 122

PAGE 123

A B Figure6-22. FrequencyspectrumoftheresistiveandreactivecomponentsoftheaverageimpedanceofFIBmodiedtungstenwiresample3.A)ResistivecomponentandB)Reactivecomponent. 123

PAGE 124

A B Figure6-23. RealandimaginarypartsoftheimpedanceofR9wire1at1kHzwithrespecttoimplantduration.A)ResistivecomponentandB)Reactivecomponent. 124

PAGE 125

A B Figure6-24. RealandimaginarypartsoftheimpedanceofR9wire3at1kHzwithrespecttoimplantduration.A)ResistivecomponentandB)Reactivecomponent. 125

PAGE 126

A B Figure6-25. RealandimaginarypartsoftheimpedanceofR9wire7at1kHzwithrespecttoimplantduration.A)ResistivecomponentandB)Reactivecomponent. 126

PAGE 127

A B Figure6-26. Realandimaginarypartsoftheimpedanceofcorrosionstudysample2at1kHzwithrespecttodifferentsurfaceconditions.A)ResistivecomponentandB)Reactivecomponent. 127

PAGE 128

A B Figure6-27. Realandimaginarypartsoftheimpedanceofinsulationcrackingstudysample3at1kHzwithrespecttodifferentsurfaceconditions.A)ResistivecomponentandB)Reactivecomponent. 128

PAGE 129

upwardsinsteadofcontinuingtodecrease.Alsothisnon-linearityismorepronouncedondays7,25,and31afterimplantation,whichisusuallytheperiodduringwhichtheinammationresponseisactive,andtheglialsheathisformed.Itcanalsobenoticedthat,inwires1and7,thenon-linearityoftheimaginaryimpedanceremainedthesameondays85,177,and205.Inwire3,thecurvetendstobecomemorelinearondays85and175.However,suchatrendisnotseenintheimaginaryimpedanceplotsofin-vitrosamples.Thesesuggestthatthetissueorthebioticimpedanceisrapidlyincreasingduringtheearlyfewweeksofimplantationafterwhichittendstosaturateandshowaverygradualdrop.Theseplotsalsosuggestthattherateofincreaseinbioticimpedanceduringtheearlyimplantationperiodappearstobehigherthantherateofdecreaseinabioticimpedance.Thesendingssupportourhypothesisthattheoverallimpedanceisprimarilydominatedbytheincreaseinbioticimpedanceduringthepeakinammationphase,andaftergliosis,theabioticfactordominatesandbringsdowntheoverallimpedancegivingaGaussianshape.However,thereducedcorrosionrateduringthelatterweeksofimplantationslowsdownthereductioninoverallimpedanceandthenearlyconstanthighbioticimpedanceaddsalongtailtotheGaussianprole. 6.6Summary Anin-vitroexperimentalevaluationoftheeffectofabioticelectrodevariationswasconducted.Twomethodswereadaptedforcreatingthesurfaceroughness.Exposedmetalwasrecessedthroughcontrolledcorrosionprocess,andcracksininsulationwerecreatedusingFIBionmillingprocedure.ThesurfacemorphologyofthedefectengineeredwireswerecharacterizedusingSEMimages,andtheimpedancewasmeasuredusingtheNanoZrimpedancemeterinasalineenvironment.Theexperimentalresultswerecomparedwithin-vivoimpedancemeasurements.Onerepresentativechronicallyimplantedelectrodearraywasselected,andthedailyimpedancedataforeachwireinthatarraywasanalyzedindepth.AGaussiancurvewasttedtotherawdata.Theresistiveandthereactivecomponentsofthemeasured 129

PAGE 130

impedanceforthreewireswereplottedandstudiedindetail.Itwasobservedthatthein-vivorealimpedanceat1kHzincreasesandthenfalls,whiletheimaginaryimpedanceat1kHzincreaseswithtimeandstabilizesaftercertainperiod.Therealimpedanceofthein-vitroPBSimmersedtungstenwirealsoincreasesandfalls,buttheimaginaryimpedanceshowsaslightinitialincreaseandthenaconstantdecrease.Thissuggestthatthetissueorbioticimpedanceplaysanimportantroleintherecoveryphaseincreaseinelectrodeimpedancewhiletheabioticsurfacemodicationmediatesthebioticledincreaseandthenleadstothedecreaseinimpedance,andthecombinedeffectyieldalongtailbehavior. 130

PAGE 131

CHAPTER7INVESTIGATIONOFMATERIALSANDENGINEERINGMETHODSTOBUILDMICROELECTRODESWITHIMPROVEDSTABILITYANDREDUCEDIMPEDANCE 7.1Introduction ItwasseeninChapters 4 5 and 6 thatinsulationdamageandrecordingsitecorrosionaretwoofthekeyabioticfactorsaffectingthechronicimpedancebehaviorofimplantedneuralelectrodes.Existingmicrowireelectrodeinsulationmaterialssuchaspolyimideandparylene-Careknowntodegradeatafasterrateinaqueoussolutions.Alsothecommonlyusedelectrodematerial,goldplatedtungstenhasbeenshowntocorrode[ 58 ].Thischapterinvestigatessomenewmaterialsforrobustelectrodeinsulationandmaterialsforimprovedrecording.AlsoFIBionmillingoftheelectroderecordingsurfaceisinvestigatedforincreasingthesurfaceroughnessandachievingareducedimpedance.TheHfO2orBCBinsulatedPt-Irmicroelectrodeswithmodiedrecordingsitepresentedinthischaptercanpotentiallyserveasamorestableandhighperformancerecordingelectrodeforchronicapplications. Thischapterisorganizedasfollows.Section 7.2 givesanoverviewoftheproposedHfO2orBCBinsulatedelectrodeswithSU-8coatedrecordingsiteforimprovedstabilityandreducedimpedance.Propertiesofthenewmaterialsarereviewedinthatsection.ThedesignconsiderationsandfabricationprocessaredescribedinSections 7.3 and 7.4 respectively.ThemethodsfollowedtocharacterizetheBCBandHfO2insulatedPt-IrmicrowiresareexplainedinSection 7.5 andthecharacterizationmethodsadaptedforSU-8coatedwiresareexplainedinSection 7.6 .FocusedIonBeam(FIB)surfacemodicationoftungstenandplatinum-iridiumrecordingsitearediscussedinSection 7.7 .TheresultsarediscussedinSection 7.8 andthechapterissummarizedinSection 7.9 7.2MaterialChoicesforProbeInsulationandImpedanceReduction Thechoiceofinsulationandstabilizationmaterialsisguidedbytheirbiocompatibility,moistureuptakepercentage,stabilityanddielectricbehavior.Successfuldemonstration 131

PAGE 132

ofBCBasaneuralelectrodesubstratematerialbyLeeet.al.[ 132 ]andthesimplefabricationprocesshavemotivatedtouseitasaninsulationmaterial.TheincreasinguseofHfO2asaninsulationmaterialinthesemiconductorindustry[ 149 150 ],alongwithitsstabilityinaqueousmedium[ 151 ]haveignitedourinterestinconsideringitasanadditionalstabilizationlayer.Thissectiondescribesapotentialdesignapproachforbuildingrobustmicroelectrodeswithlowimpedancerecordingsite.Therobustmicroelectrodedesignincludesplatinum-iridium(80%-20%)microwiresencapsulatedwithaninsulationlayercomposedofathinlmofbenzocyclobutene(BCB)andanoptionaladditionalstabilizationlayerusingahigh-kdielectricmaterialsuchashafniumoxide(HfO2).Platinum-iridiumisselectedastherecordingmaterialbecauseofthehighcorrosionresistancecharacteristicofplatinumandthehightensilestrengthofthealloy(whichis3-5timeshigherthanpureplatinum).Moreover,successfulneuralrecordingusingPt-Irmicrowireelectrodeshasbeenreportedindifferentstudies[ 152 154 ].ThedesignalsoincludescarbonizedSU-8nanobersasalowimpedancepromotinglayer.Theincreasedsurfacearea,goodmechanicalproperties,andexcellenttissuecompatibility[ 155 157 ]ofSU-8hasattractedtheattentionofresearchersfordrugdeliveryapplications[ 158 ]andalsoforrecordingneuralsignalsfromPNS[ 159 ].AconceptualdrawingofthemicroelectrodeisgiveninFigure 7-1 BackgroundoftheMaterialProperties:Thekeymaterialpropertiesconsideredwerewateradsorptionrate,absorptionrate,stabilityinaqueousmedium,biocompatibility,anddielectricconstant.ItwasfoundthatbothBCBandHfO2havegoodbiocompatibilityandlowwaterabsorptionrate.BCBishydrophobic,whilethesurfacewettabilityofHfO2canbecontrolledbythefabricationprocess.ThedielectricconstantofBCBisintherangeofpolyimideorparylene,andtheaveragevalueofthedielectricconstantofHfO2ishigh(20)[ 150 ].AdetailedcomparisonofthematerialpropertiesofcandidateelectrodeinsulationmaterialsandtheexistingmaterialsisgiveninTable 7-1 132

PAGE 133

Table7-1. Tablecomparingthekeymaterialpropertiesoftheproposedandtheexistingelectrodeinsulationmaterials. ProposedmaterialsforprobeinsulationandstabilizationCommonlyusedmaterialsforprobeinsulation Insulationlayer(BCB)Stabilizationlayer(HfO2)SiO2forinsulationSi3N4forinsulationPolyimideforinsulationParyleneCforinsulation Wateradsorption/SurfacewettabilityHyrdophobic[ 160 ]HydrophilicatlowRFpowerandhydrophobicathighRFpower[ 161 ]HydrophilicHydrophobicHydrophilicHydrophilic Waterabsorptionrate0.12wt%[ 132 ]Small[ 162 ]SmallSmallLarge[ 133 ]Small(<0.1%)[ 135 ] Delamination/stabilityinsalineorin-vivoenvironmentUnknownStableinaqueouselectrolytemedium[ 151 ]Unstableunderchronicconditions[ 82 ]Unstableunderchronicconditions[ 82 ]UnstableinPBSUnstableinPBS BiocompatibilityGood[ 132 ]Good[ 151 163 164 ] Good[ 82 ]Good[ 82 ]GoodGood DielectricConstant2.64[ 132 ]Averagevalue=20[ 150 ]3.9[ 165 ]7[ 166 ]2.9@1KHz[ 117 ]3.10@1KHz[ 167 ] 133

PAGE 134

Figure7-1. IllustrationshowingBCB-HfO2insulatedPt-Irmicrowiredesign.ThetextureonthetopsurfacerepresentsthethinlmofcarbonizedSU-8nanobercoatedonthePt-Irtip. 7.3DesignConsiderations Thethicknessoftheinsulationlayersiscriticalinthedesignofmicroelectrodes.Caremustbetakentoensurethattheinsulationlayerisoptimallythickfornegligiblecapacitivecouplingbetweenthemetalmicroelectrodeandtheelectrolyte.IntheBCB-HfO2design,BCBisconsideredastheprimaryinsulationlayer,whileHfO2isthestabilizationlayersurroundingthePt-Iralloy.SincetheHfO2layerismuchthinner(10-20nm)thantheBCBlayer,itseffectivecapacitancecanbeneglected. 7.3.1AnalyticalCalculationofOptimalInsulationThickness Basedonearlierimpedancemeasurementexperiments,theaverageimpedancevalueofa50mdiameterplatinumwireattherecordingsiteisZRecSite=166.29Kat1KHz.Tominimizecapacitivecouplingalongtheshank,wesettheimpedanceoftheprobeinsulationistwoordersofmagnitudehigherthanthatoftherecordingsite.Then,theimpedanceattheshankis,ZShank=166.29x102K. 134

PAGE 135

Figure7-2. Schematicshowingthechargedistributionandelectriceldpatternofthecylindricalcapacitancemodel. Thecapacitivereactanceattheinsulationisgivenby, jZj=1 j!C=1 2fC(7) Thisimplies, C=1 21103166.29105=9.5710)]TJ /F3 7.97 Tf 6.59 0 Td[(12(F)(7) Sincetheplatinum-iridiummicroelectrodeiscylindrical,itisrationaltoassumethattheelectrode-electrolyteinterfacewillalsotakeacylindricalshapearoundtheinsulation.Therefore,themetalelectrode,insulation,andtheelectrolyteinterfacewillformacylindricalcapacitanceasshowninFigure 7-2 Letusassumethatcharge+Qisdistributedallalongtheprobeshankandcharge-Qisdistributedalongtheelectrolyteinterface.Thedirectionoftheelectriceldisrepresentedby!inthegure.Thecapacitanceofacylindricalcapacitorisgivenby[ 168 ], 135

PAGE 136

C=2"0"rL ln(b=a)(7) whereL=lengthoftheplatinummicroelectrode=5mm(assumed),a=radiusoftheplatinum-iridiummicroelectrode=25m,b=radiusoftheelectrolyteinterface=tobecalculated,"r=relativepermittivityoftheinsulationmaterial(forbenzocyclobuteneitisassumedtobe2.64[ 132 ]),and"0=permittivityoffreespace=8.85410)]TJ /F3 7.97 Tf 6.58 0 Td[(12(F/m).SubstitutingthevaluesinEquation 7 andsolvingforb,b=26.99310)]TJ /F3 7.97 Tf 6.59 0 Td[(6mforBCB.Hence,thethicknessoftheBCBinsulationisb-a=26.99310)]TJ /F3 7.97 Tf 6.59 0 Td[(6-2510)]TJ /F3 7.97 Tf 6.59 0 Td[(6(m)=1.99m'2m. 7.3.2NumericalAnalysisofOptimalSU-8Thickness AnumericalanalysiswasperformedusingCOMSOLniteelementmodelingsoftwaretodeterminetheoptimalthicknessofSU-8nanobersforachievingasignicantreductionintheelectrodeimpedance.Themodelincludeda50mdiameterPt-Irelectrodeinsulatedwitha2mthickBCBlayer.TheexposedrecordingsiteoftheelectrodeiscoatedwithalayerofSU-8.ThethicknessoftheSU-8lmwasvariedfrom20-80minstepsof20m.Theinitialassumptionsandtheboundaryconditionswereadaptedfromthemodelforelectrodesurfacevariationandareexplainedindetailinsection 5.3.1 .TheconductivityvaluesofthematerialsusedforthesimulationaregiveninTable 7-2 .1104(S/m)istheconductivityvalueforbulkSU-8thatispyrolyzedat900-1000C[ 169 ]. Table7-2. ConductivityvaluesusedforCOMSOLniteelementmodeling. MaterialConductivity(S/m)) Platinum-Iridium4106BCB110)]TJ /F3 7.97 Tf 6.59 0 Td[(17SU-81104Saline1.8 136

PAGE 137

Figure7-3. PercentagereductioninimpedancefordifferentthicknessesofanuniformlmofSU-8coatingonPt-IrwiresascomparedtononSU-8coatedtungstenwires. Figure 7-3 showsthepercentagereductioninimpedancefordifferentthicknessesofSU-8coatingonPt-IrwiresascomparedtononSU-8coatedtungstenwires.Theresultsshowthatmorethan50%reductioninimpedancewithrespecttoanoncoatedgoldplatedtungstenmicroelectrodecanbeachievedwitha40mthicklmofcarbonizedSU-8onaplatinum-iridiumelectrode. 7.4FabricationProcess ThefabricationprocessowoftheBCB-HfO2insulatedmicroelectrodealongwithcarbonizedSU-8nanobercoatingontherecordingsiteisasequentialprocedureofatomiclayerdepositionofhafniumoxideondicedbarePt-Irmicrowires,followedbyelectrospinningandcarbonizationofSU-8,andnallydipcoatingofBCB.DepositingHfO2overbarePt-IrpriortoBCBdipcoatingwillpreventthedegradationofBCBduringelectrospinningofSU-8andsubsequentpyrolysisprocesseswhichisperformedat1000C.ItisnotedthatHfO2isarefractorymaterialwithameltingpointof2758Candhencecaneasilywithstandthepyrolysisprocess.Aschematicillustrationof 137

PAGE 138

Figure7-4. SchematicshowingthefabricationprocessowforBCB-HfO2insulatedPt-IrwirescoatedwithcarbonizedSU-8nanobers. thefabricationprocessowwiththekeystepsisgiveninFigure 7-4 .However,forinvestigatingandcharacterizingeachmaterialandcompatibilitybetweenmaterials,thissequencewasnotfollowed.EachmaterialwasdepositeddirectlyonabarePt-Irmicrowireandcharacterized.Thedepositionprocedureemployedforeachmaterialisdescribedasfollows. BCBdipcoatingonPt-Irmicrowires:BarePt-Irmicrowireof50mdiameter(A-MSystems)weredicedintosmallpiecesofapproximatelength5cmusingadicingsaw(AdvancedDicingTechnologiesInc.)attheNanoscaleResearchFacility(NRF)atUF.Benzocyclobutene(BCB)wascoatedconformallyonthebarePt-Irwiresusingadip-coatingmethod.BCBisaliquidpolymerwithaviscosityrangingfrom14cStto1950cStfordifferentcommercialformulations[ 170 171 ].TwotypesofBCBformulationwere 138

PAGE 139

tested.Initially,Cyclotene4026,whichisaphotopatternableBCBfromDowChemicalswasinvestigated.Cyclotene4026isahighviscosityBCBwithaviscosityof1100cStat25C,whichcanyieldlmsof7-14mthicknessuponcuring.ThefabricationstepsinvolvedbringingtherefrigeratorstoredBCBtoroomtemperatureanddipcoatingonthecutwires.DicedwireswereintroducedintoasmallpuddleofBCBandremovedrapidly.Coatedsampleswerethensubjectedtonalcuringundervacuum.SinceBCBcontainssiliconinitscomposition,vacuumcuringisrequiredtopreventtheoxidationofsiliconathighertemperature.Thesampleswererampedto150C,andaftercuringatthattemperaturefor15minutes,thetemperaturewasincreasedto250C,andthesampleswereheldatthattemperaturefor60minutes.After1hour,thetemperaturewasbroughtdowntoroomtemperature.Theentirepostbakingwasdoneunderavacuumpressureof-29inofHg(IsotempvacuumovenModel281A).However,aftercuring,itwasfoundthatthecoatingwasnotuniformonthewires.ThehighviscosityoftheBCBresultedinbubblesandbersonthecoatedsurface. Toimprovetheuniformityofthecoatedlm,alowviscosityBCBwastested.Cyclotene3022(DowChemicals)isadryetchlowviscosityBCBofviscosity14cStat25C.Almofthickness1-2.4mcanbeobtainedbyusingthisformulation.DipcoatingandnalcuringstepswerethesameasthoseforhighviscosityBCB.Aftercuring,itwasfoundthatthislowviscosityBCBresultedinamuchimproveduniformityinthecoatedlm.Afewwiresstillshowedsomenon-uniformityonthesurface,butthenon-uniformityisnotaslargeasseeninthehighviscosityBCBlm.ThickeningofBCBonexposuretoatmosphereduringthedipcoatingprocesscanaccountforsomenon-uniformitiesinthosewiresthatwereimmersedlast.Figure 7-5 showstheSEMimagesofhighviscosityBCBandlowviscosityBCBcoatedPt-Irmicrowires.ThelowviscosityBCBcoatedwiresweresubjectedtosoaktestwhichisreportedinSection 7.5.1 139

PAGE 140

A B Figure7-5. SEMimagesofhighviscosityBCBandlowviscosityBCBcoatedPt-Irmicrowires.A)HighviscosityBCBcoatedwiresandB)LowviscosityBCBcoatedwires. HfO2depositiononPt-Irmicrowires:DicedPt-IrwireswerecoatedwithathinlmofHfO2.Atomiclayerdeposition(ALD)(CambridgeNanoFiji200)wasemployedfordepositingHfO2withathicknessof10-20nm.AstandardHfO2plasmarecipewasfollowed.3DMAH((HfN(Me)2)4)wasusedasaprecursorandwasheatedto50C.Thisprecursorwasexposedtothesubstratemaintainedat200C,andthenoxygenplasmawasintroducedtoformHfO2.Theprocesswasrepeatedtogetadesiredthickness.However,subsequentTransmissionElectronMicroscope(TEM)imagesofthesamplesshowedthatthedepositionwasnotgood,andtherecipeneedstobereevaluated.SincethesamplesdidnothavegoodHfO2coating,furthercharacterizationstepswerenotcarriedout. SU-8coatingontherecordingtips:DicedPt-IrmicrowiresamplesweresubjectedtotheelectrospinningprocesswhereSU-8nanoberswereelectrospunonthePt-Iralloy.Twokindsofelectrospinningtechniquesareavailable,namely,singleneedleelectrospinningandtubeneedleelectrospinning[ 172 ].SingleneedleelectrospinningwasusedforcoatingtheSU-8bersonthePt-Irsamples.Themicrowiresweremountedonacopperplateandgroundedwiththeplate.SU-8polymerisloadedintoa 140

PAGE 141

syringeconnectedtoaneedleofdiameter0.5mm.TheSU-8polymerwaspumpedataconstantrateof0.5ml/hrandtheberswerecollectedatthesamples.Needlevoltagewasmaintainedbetween10kVand15kV.Figure 7-6 showsthephotographsoftheSU-8electrospinningprocess.TheelectrospunSU-8nanoberswerethenpostbaked,developed,andcarbonized.Thecarbonizationorpyrolysiswasdoneinatubefurnaceinaninertatmosphereat1000C. 7.5CharacterizationofBCBInsulation CharacterizationofBCBinsulationisathreestepprocess.TherststepistoevaluatethesurfacemorphologyandtheuniformityoftheBCBcoatingonthePt-Irwiresthroughscanningelectronmicroscope(SEM)images.ThesecondstepistosubjecttheBCBinsulatedmicrowiresamplestoasoaktestfordurabilityevaluation.Finally,afterimmersioninPBS,thesampleswereimagedagainusingSEMtoqualitativelyassessanyvariationinthesurfacemorphology. 7.5.1DurabilityEvaluationusingSoakTest ThedurabilityoftheBCBwasassessedusingasoaktest.TheBCBencapsulatedPt-Irmicrowiresampleswereimmersedina0.9%PBS(SigmaAldrich)solutioncontaining30mMof30%H2O2(FisherScientic)for57hourstomimictheneuralenvironmentwithinammatoryresponse.Thesoaktestwereconductedatroomtemperature.AlongwithBCBinsulatedplatinum-iridiumwires,polyimideinsulatedtungstenmicrowireswerealsosubjectedtosoaktestsunderthesameconditions.Allthesampleswereimagedbothbeforeandafterthesoaktests. 7.5.2SurfaceMorphologyEvaluationusingSEMImages BCBdipcoatedplatinum-iridiummicrowiresampleswerecharacterizedbycomparingthevariablepressurescanningelectronmicroscope(VP-SEM)imagesbeforeandaftera57hoursoaktest.SEMimagesofdicingsawcutpolyimideinsulatedtungstenmicrowireswerealsotakenpriortoandafter57hoursofsoaktestsandcomparedwithBCBinsulatedPt-Irwires.ThebeforeandaftersoaktestSEMimages 141

PAGE 142

A B Figure7-6. PhotographsofSU-8electrospinningprocess.A)Photographofelectrospinningexperimentsetup,andB)CloseupofthePt-Irmicrowiresamplesmountedonametalsubstrate. 142

PAGE 143

ofpolyimideinsulatedtungstenmicrowiresaregiveninFigures 7-7 and 7-8 .ThebeforeandaftersoaktestSEMimagesofBCBcoatedPt-IrwiresaregiveninFigures 7-9 and 7-10 .Itcanbereadilyobservedfromtheimagesthatthepolyimideinsulatedtungstenwiresunderwentsignicantmetalcorrosionandnoticeableinsulationdamage,whiletheBCBcoatedPt-Irwiresdidnotshowanyremarkablesurfacevariationinboththealloyandthepolymer. 7.6SurfaceMorphologyEvaluationofSU-8nanobers TheuniformityandthethicknessoftheelectrospunSU-8wereassessedbyobtaininghighmagnicationSEMimages.CarbonizedSU-8coatedplatinum-iridiummicrowiresampleswerecharacterizedusingthescanningelectronmicroscope(CarryScopeSEM,JEOL,Inc.)images.Theimagesgaveaqualitativeideaofhowwellthebersareattachedtothewires.Figure 7-11 showstheSEMimagesofPt-IrmicrowirescoatedwithcarbonizedSU-8nanobers.ItcanbeseenfromtheimagesthattheattachmentofSU-8bersisnotgood.ThebersappeartopeelawayfromthePt-Irsurface.FurtherworkisnecessarytobetterunderstandtheadhesionofSU-8tothePt-Irsurfaceandtorenethecoverage. 7.7FIBSurfaceModicationforReducedImpedance TrialexperimentswereconductedtocreateametalrecessioninthetungstenmicrowiresampleusingFocussedIonBeam(FIB)milling.TheexposedmetalsurfaceinonedicingsawcuttungstenmicrowiresamplewasionmilledusingFIBtoadepthof15m.Fromthetrialexperiments,itisobservedthatthesurfaceroughnessoftheelectroderecordingsitecanbesignicantlyimprovedbydiscretelybombardingthemetalsurfacewithFIBgalliumions.SEMimagesofthepreliminarytrialsshowedaremarkableincreaseinthemicroscaleroughnessonthesurfaceofthetungstenrecordingsite.Figure 7-12 showstheSEMimagesofatungstenmicrowirebeforeandafterFIBsurfaceroughnessmodication.Impedancemeasurementsontheroughnessengineeredwireshowedabouta56%reductionintheaverageimpedance 143

PAGE 144

A B C D E F Figure7-7. SEMimagesofpolyimideinsulatedtungstenmicrowirestakenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2.A)sample1beforesoak,B)sample1aftersoak,C)sample2beforesoak,D)sample2aftersoak,E)sample3beforesoak,andF)sample3aftersoak. 144

PAGE 145

A B C D E F Figure7-8. SEMimagesofpolyimideinsulatedtungstenmicrowirestakenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2.A)sample4beforesoak,B)sample4aftersoak,C)sample5beforesoak,D)sample5aftersoak,E)sample6beforesoak,andF)sample6aftersoak. 145

PAGE 146

A B C D E F Figure7-9. SEMimagesofBCBinsulatedplatinum-iridiummicrowirestakenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2.A)sample1beforesoak,B)sample1aftersoak,C)sample2beforesoak,D)sample2aftersoak,E)sample3beforesoak,andF)sample3aftersoak. 146

PAGE 147

A B C D E F Figure7-10. SEMimagesofBCBinsulatedplatinum-iridiummicrowirestakenbeforeandafter57hoursofsoaktestin0.1MPBSwith30mMH2O2.A)sample4beforesoak,B)sample4aftersoak,C)sample5beforesoak,D)sample5aftersoak,E)sample6beforesoak,andF)sample6aftersoak. 147

PAGE 148

A B C D Figure7-11. SEMimagesofPt-IrmicrowirescoatedwithcarbonizedSU-8nanobers.A)sample1,B)sample1zoomedintoshowthenanobertexture,C)sample2,andD)sample2zoomedintoshowthenanobertexture. at1kHzcomparedwiththepristinewire.Thisisaninterestingandpromisingoptionforincreasingthesurfaceroughnessoftherecordingsiteandtherebyachievingareductioninimpedance.AsimilarstudywascarriedoutonaBCBinsulatedPt-Irmicrowiresampleandthesurfaceroughnessmodiedsamplewasmeasuredforimpedanceinanin-vitrosetup.TheexposedmetalsurfaceinonedicingsawcuttungstenmicrowiresamplewasionmilledusingFIBforadepthof10m.Figure 7-13 showstheSEM 148

PAGE 149

imagesofBCBinsulatedPt-IrmicrowirebeforeandafterFIBsurfaceroughnessmodication. A B Figure7-12. SEMimagesoftungstenmicrowirebeforeandafterFIBsurfaceroughnessmodication.A)BeforeandB)After.Insetshowsthecloserviewofthemicroscaleroughnessontungstensurface. A B Figure7-13. SEMimagesofBCBinsulatedPt-IrmicrowirebeforeandafterFIBsurfaceroughnessmodication.A)BeforeandB)After. ImpedanceMeasurementusingNanoZrImpedanceMeter:TheNanoZrimpedancemeter(WhiteMatter,LLC.)isusedformeasuringtheimpedanceofFIB 149

PAGE 150

surfacemodiedPt-Irmicrowires.TheimpedancemeasurementsmadeontheFIBsurfacemodiedtungstenwireswasmuchlessthantheimpedancemeasuredfromtheunmodiedpristinewireatallmeasuredfrequencies.Thevariabilitybetweeneachmeasurementwasless.At1kHz,theimpedanceoftheFIBsurfacemodiedtungstenmicrowirewas27.80.5k,whichisabout56%lessthantheimpedancemeasuredfromapristinewire,whichwas641.5k.Ontheotherhand,theimpedancemeasurementsmadeonPt-IrmicrowiresbeforeandafterFIBsurfacemodicationshowedlargevariationsbetweeneachmeasurements.Duetothissignicantmeasurementerror,thedifferenceinimpedancebetweenpristineandFIBmodiedwirecouldnotbeascertainedprecisely.At1kHz,theimpedanceoftheFIBsurfacemodiedPt-Irmicrowirewas21.011.9k,andthatofapristinePt-Irwirewas19.68.7k.Atarstlook,itmayappearthatthereisanincreaseintheimpedanceoftheFIBsurfacemodiedwire.However,anychangeinimpedancecannotbeascertainedwithintheexperimentalvariability.Figure 7-14 showsthefrequencyspectrumoftheaverageimpedancemeasuredfromthetungstenmicrowiresamplebeforeandafterFIBsurfacemodication.Figure 7-15 showsthefrequencyspectrumoftheaverageimpedancemeasuredfromtheplatinum-iridiummicrowiresamplebeforeandafterFIBsurfacemodication. 7.8ResultsandDiscussion SEMimagesofthehighviscosityBCB(Cyclotene4026)andlowviscosityBCB(Cyclotene3022)coatedPt-IrwiresclearlyshowthatalowviscosityBCBpolymerisdenitelybetterforobtainingauniformcoating.However,thelowviscosityBCB(Cyclotene3022)stillhadsomebubblesinafewwires.ItwasfoundthattheCyclotene3022solutionthickenswhenexposedtoatmosphereforafewminutesandresultsinbubblesandbersonthecoating.Thiscouldpossiblybemitigatedintwoways.TheBCBsolutioncanbefurtherdilutedusingaMesitylene(T1100-DowChemicals)solventorthedipcoatingprocesscanbedoneinaninertenvironmentsuchasinaglovebox. 150

PAGE 151

Figure7-14. FrequencyspectrumoftheaverageimpedancemeasuredfromtungstenmicrowiresamplebeforeandafterFIBsurfacemodication. ItcanbeobservedfromtheSEMimagestakenpriortoandafterthesoakteststhattheBCB(Cyclotene3022)insulatedwiresdidnotshowsignicantsurfacemorphologicalchangesandalmostremainedthesame.However,tungstenmicrowiresinsulatedwithpolyimidedisplayedlargerecessionofexposedmetalandsignicantsurfaceroughnessasaresultofcorrosion,andtheinsulationwasalsodelaminatedowingtothehighermoistureabsorptionrateofpolyimide.ThesoaktestresultsconrmthatPt-Irdoesnotcorrode,andBCBismorestableinanaqueousmediumthanpolyimideorparylene-C. TheSEMimagesofSU-8coatedmicrowiresamplesshowthattheSU-8nanobersarepoorlyattachedtothePt-Irsurface.Similarly,theinitialsurfacecharacterizationofHfO2coatedPt-IrwiresrevealedpoordepositionofHfO2onthePt-Irsurface.ThisimpliesthatfurtherdevelopmentisneededtodeterminetheusabilityofHfO2andSU-8 151

PAGE 152

Figure7-15. FrequencyspectrumoftheaverageimpedancemeasuredfromPt-IrmicrowiresamplebeforeandafterFIBsurfacemodication. onPt-Irwires.Morein-depthstudyintermsofthesurfacechemistryofthesematerialsandintermsofprocessingconditionsisneededtoimprovethedepositionofthesematerials. ThemethodofincreasingthesurfaceroughnessofPt-IrwireusingFIBionmillingseemstobepromising.Thoughthepreliminaryin-vitromeasurementsdidnotshowsignicantdecreaseintheimpedanceintheFIBmodiedPt-Irwiresduetolargeexperimentalvariability,moresamplesmilledtodifferentdepthsareneededforabetterevaluationoftheeffects. 7.9Summary Newmaterialsformicroelectrodeinsulationandstabilizationwereinvestigated.Alsoengineeringmethodstoimprovethesurfaceroughnessoftherecordingsite 152

PAGE 153

andtoreducetheimpedancewereinvestigated.Sampleswerefabricatedandtestedasaproofofconceptdemonstration.Benzocyclobutene(BCB),whichisalowdielectricpolymer,wasdipcoatedontoPt-Irmicrowiresamplesandtestedfordurability.ThepreliminarysoaktestresultsshowedthatBCBcanbeagoodcandidateformicroelectrodeinsulationandcanpotentiallyfunctionreliablyinchronicapplications.Moreover,carbonizedSU-8,whichisaconductivepolymer,wascoatedontherecordingsiteofthewiresusingelectrospinningandpyrolysismethods.InitialresultsshowedpooradhesionofSU-8nanoberstothePt-Irwiresandsuggestedanin-depthstudyofthesurfacepropertiesofPt-IrandSU-8arenecessary.FIBsurfaceroughnessmodiedPt-Irsamplewasmeasuredforimpedanceunderin-vitroconditions.PreliminaryresultssuggestthatFIBsurfaceroughnessmodiedandBCBinsulatedPt-Irmicrowiresmayhaveimprovedperformanceandhighreliabilityaslong-termneuralimplantscomparedwithpolyimideinsulatedtungstenmicrowires. 153

PAGE 154

CHAPTER8SUMMARYANDFUTUREWORK 8.1ResearchSummary Thisdissertationinvestigatedsomeofthekeyabioticfactorsidentiedasmajorcontributorsforthefailureofchronicneuralimplantsandproposedengineeringsolutionsformitigatingthesefactors.Micromotioninducedtissuestrainisoneofthesignicantfailuremodesofmicroelectrodesinchronicapplications.Thisresearchstudiedthephysicsbehindthemicromotionstrainanddemonstratedahighlycompliantserpentinecabledesignforimprovedfront-endstrainrelief.Thecomplianceofstraightandserpentineelectrodecableswerestudiedusinganalyticalandnumericalmodels.PrototypesoftheserpentinecableswerefabricatedusingMEMSmicrofabricationtechniquesandwereexperimentallymeasuredforcompliance.Itwasshownthattheserpentinecablesareupto10timesmorecompliantthanthetraditionalstraightcablesofthesameoveralldimensions.Alsoaplanforbuildingtheinterfacecablesandsubstrateforintegratedelectronicsasseparateexibleandrigidmodulesispresented.Prototypesofboththemoduleswerefabricated.Thesimpledesign,straightforwardprocessing,andpackagingstepsandtheimprovedexibilityoftheserpentineshapedcablescanbeusedforbuildingfuturelowstrainneuralimplantswithhighchronicreliability. Impedanceisoneoftheimportantmetricsfordeningtheefcacyofarecordingmicroelectrode.However,recentstudies[ 121 ]haveobservedalargetemporalvariationoftheelectrodeimpedanceinchronicapplications.Thereiscurrentlynoscienticbasistoexplainthisbehavior,andthisisperceivedasoneoftheimportantunsolvedquestionsfortheBMIresearchcommunityatpresent.Chapters 5 and 6 approachthisquestionfromanabioticperspective.Theinuenceofcommonlyobservedelectrodesurfacevariationssuchasrecordingsitecorrosionandinsulationdelaminationandcrackingonimpedancewereanalyzednumericallyandexperimentally.Thenumerical 154

PAGE 155

andin-vitroexperimentalresultswerecomparedtotheimpedancedataobtainedfrompriorin-vivostudies[ 121 ].Thein-vivoimpedancebehaviorofwiresfromachronicallyimplantedelectrodearraywascomparedwiththein-vitroimpedanceresults.Basedonthecomparison,ascienticmodelwasdevelopedtoexplainthetimevaryingin-vivoimpedanceofchronicneuralelectrodes.Themodelexplainsthebellshaped,longtailed,temporalvariationoftheoverallin-vivoimpedanceisaresultofthecombinationoftheinitialincreaseinbioticimpedanceandadecreaseinabioticimpedanceduetoahighcorrosionrateduringtherstphase,followedbyasaturationinbioticimpedanceandasimultaneouslowratecorrosionandinsulationdamageleddecreaseinabioticimpedanceduringthepost-gliosisphase. Moreover,newmaterialsforrobustmicroelectrodeinsulationandmaterialsandengineeringmethodsforachievingreducedimpedancewereinvestigated.Chapter 7 presentsthendingsofthestudy.Highandlowdielectricconstantmaterials,hafniumoxideandbenzocyclobutene(BCB),wereconsideredforprobeinsulationandstabilization.CarbonizedSU-8nanoberwasalsoconsideredasaconductivepolymercoatingtoreducetheelectrodeimpedanceattherecordingsite.Designparameterswereevaluatedandprototypeswerefabricatedandcharacterized.ThequalitativecharacterizationofsurfacemorphologythroughSEMimagesshowedarelativelyuniformdepositionofthelowviscosityBCBonPt-Irmicrowiresamples,andtheimagestakenafterthesoaktestsshowedthattheBCBremainedintactwhilethepolyimideinsulationwasdamaged.FocussedIonBeam(FIB)surfacemodicationforincreasedmicroscalesurfaceroughnessandreducedimpedancewasalsoinvestigated.In-vitroimpedancemeasurementsmadeusingtheNanoZrimpedancemetersuggestedthatFIBsurfacemodicationmayleadtoareductionintheelectrodeimpedance.ThesendingssuggestthatFIBsurfacemodiedmicroelectrodesinsulatedwithBCBwillpotentiallyhavebetterlong-termstabilityandimprovedperformance. 155

PAGE 156

8.2RecommendationsforFutureWork Theserpentineshapedcableswereshowntobeupto10timesmorecompliantthanstraightcablesunderbenchtopexperimentalconditions.However,theperformanceofthenewcablesunderin-vivoconditionsisyettobeevaluated.Thecorrelationbetweenthetenfoldincreaseincomplianceandthereductionintissuestrainandpossiblereductioninbiologicalresponseduetomicromotionneedstobestudiedtoassesstheimprovementinthereliabilityofthenewdesign.Itissuggestedthatanin-vivostudybemadeononesetofanimalsimplantedwiththenewdesigncablesandthesecondsetofanimalsimplantedwiththeoldstraightcablesandelectrophysiologicalrecordingsandimpedancemeasurementsaremadeforachronicduration.Sinceimpedanceandneuronalrecordingsaredirectlyinuencedbytissueinammationresponse,acomparisonofthesemetricsbetweenthetwoanimalsetswillbeusefulinestimatingtheimpactofincreasedcompliancerealizedbytheserpentinecables.Asimilarstudyissuggestedformodularandmonolithicelectrodes. Throughthemodeldevelopedinthiswork,itwasestablishedthatduringtheearlyweeksofimplantation(i.e.,duringtheformationofglialsheath),tissueresponseplaysamajorroleinincreasingtheelectrodeimpedance.Duringthelaterweeks(afterthegliosisphase),abioticfactorsincludingrecordingsitecorrosion,insulationdelamination,andcrackingdominateovertissueeffectsandlowertheimpedanceuntilitreachesapointwheretheeffectofconstanttissueimpedanceislargelyseen.ThismodelclearlyexplainstheGaussianshapedin-vivoimpedanceprolewithalongtailobservedat1kHzfrequency.Inordertoobtainabetterunderstandingofthebehaviorofeachcomponentoftheelectrode-electrolyteimpedance,andalsototestthevalidityofthemodelathigherfrequencies,itisrecommendedthatanin-vivostudybeconductedwheretheimpedanceismeasuredforabroadspectrumoffrequencies,from1Hzto100kHz.Theresultsfromsuchastudymayprovidemoreinsightsontheimpactofthedoublelayercapacitanceandtheconstantphaseelementonimpedancevariation. 156

PAGE 157

ThepreliminaryresultsobtainedforinsulatingthePt-IrmicroelectrodewithBCBandreducingtheimpedanceusingFIBsurfacemodicationmethodareencouraging.ItisrecommendedthattheBCBprocessingstepbestudiedinmoredepthsoauniformandsmoothinsulationlayercanbedepositedaroundthePt-Iralloymicrowire.Similarly,itissuggestedthattheFIBsurfacemodicationbestudiedindetailonmoresamplesforstatisticalvalidation.Uponsuccessfulvalidationunderin-vitroconditions,theexperimentcanberepeatedin-vivoforfurtherevaluationoftheimpedancebehavior. PreliminaryresultsonHfO2andSU-8werenotasexpected.However,thisdoesnotimplythatthesematerialsarenotsuitableforprobestabilizationandrecordingsurfacecoating.MorestudiesareneededtounderstandthesurfacechemistryofthesematerialsandtoimprovetheiradhesiontoPt-Ir.InadditiontoHfO2andBCB,otherdielectricmaterialssuchasnanoporousaluminaoranodizedaluminumoxide(AAO)maybecandidatematerialsforinsulatingtherecordingmicrowires.AAOisapromisingelectrodeinsulationmaterialwithgoodbiocompatibility[ 173 175 ]andhydrophobicproperties[ 176 ].Itisrecommendedthateffortsbesteeredtowardsthosematerialsalso. 157

PAGE 158

REFERENCES [1] A.Prasad,V.Sankar,A.Dyer,E.Knott,Q.Xue,T.Nishida,J.Reynolds,G.Shaw,W.Streit,andJ.Sanchez,Couplingbioticandabioticmetricstocreateatestbedforpredictingneuralelectrodeperformance.inConfProcIEEEEngMedBiolSoc.2011,2011,pp.3020. [2] E.Patrick,M.Ordonez,N.Alba,J.C.Sanchez,andT.Nishida,Designandfabricationofaexiblesubstratemicroelectrodearrayforbrainmachineinterfaces,inEngineeringinMedicineandBiologySociety,2006.EMBS'06.28thAnnualInternationalConferenceoftheIEEE,2006,pp.2966. [3] E.Patrick,V.Sankar,W.Rowe,S.-F.Yen,J.C.Sanchez,andT.Nishida,Flexiblepolymersubstrateandtungstenmicroelectrodearrayforanimplantableneuralrecordingsystem,inEngineeringinMedicineandBiologySociety,2008.EMBS'08.30thAnnualInternationalConferenceoftheIEEE,2008,pp.3158. [4] E.Patrick,V.Sankar,W.Rowe,J.C.Sanchez,andT.Nishida,Animplantableintegratedlow-poweramplier-microelectrodearrayforbrain-machineinterfaces,inEngineeringinMedicineandBiologySociety,2010.EMBS'10.32ndAnnualInternationalConferenceoftheIEEE,2010,pp.1816. [5] E.Patrick,Design,fabrication,andcharacterizationofmicroelectrodesforbrain-machineinterfaces,Ph.D.dissertation,UniversityofFlorida,2010. [6] NanoZUser'sManual,WorldWideWeb, http://neuralynx.com/manuals/nanoZ manual.pdf ,2013. [7] NationalInstituteofNeurologicalDisordersandStroke(NINDS)DisorderIndex,WorldWideWeb, http://www.ninds.nih.gov/disorders/disorder index.htm ,2012. [8] ParalysisFacts&Figures,WorldWideWeb, http://www.christopherreeve.org/site/c.mtKZKgMWKwG/b.5184189/k.5587/Paralysis Facts Figures.htm ,2012. [9] StrokeStatistics,WorldWideWeb, http://www.strokecenter.org/patients/about-stroke/stroke-statistics/ ,2012. [10] SpinalCordInjuryFactsandFiguresataGlance,WorldWideWeb, https://www.nscisc.uab.edu/PublicDocuments/fact gures docs/Facts%202012%20Feb%20Final.pdf ,2012. [11] EpilepsyStatsandFacts,WorldWideWeb, http://www.epilepsy.com/node/986825 ,2012. [12] D.Hirtz,D.Thurman,K.Gwinn-Hardy,M.Mohamed,A.Chaudhuri,andR.Zalutsky,Howcommonarethecommonneurologicdisorders?Neurol-ogy,vol.68,no.5,pp.326,2007. 158

PAGE 159

[13] J.Bach,U.Ziegler,G.Deuschl,R.Dodel,andG.Doblhammer-Reiter,Projectednumbersofpeoplewithmovementdisordersintheyears2030and2050,MovDisord,vol.26,no.12,pp.2286,2011. [14] E.Leuthardt,G.Schalk,D.Moran,andJ.Ojemann,Theemergingworldofmotorneuroprosthetics:aneurosurgicalperspective,Neurosurgery,vol.59,no.1,pp.1,2006. [15] P.Leake,G.Hradek,andR.Snyder,Chronicelectricalstimulationbyacochlearimplantpromotessurvivalofspiralganglionneuronsafterneonataldeafness,JCompNeurol,vol.412,no.4,pp.543,1999. [16] M.Humayun,E.deJuanJr,J.Weiland,G.Dagnelie,S.Katona,R.Greenberg,andS.Suzuki,Patternelectricalstimulationofthehumanretina,VisionRes,vol.39,no.15,pp.2567,1999. [17] P.Limousin,P.Krack,P.Pollak,A.Benazzouz,C.Ardouin,D.Hoffmann,andA.Benabid,Electricalstimulationofthesubthalamicnucleusinadvancedparkinson'sdisease,NEnglJMed,vol.3396,no.16,pp.1105,1998. [18] M.Hodaie,R.Wennberg,J.Dostrovsky,andA.Lozano,Chronicanteriorthalamusstimulationforintractableepilepsy,Epilepsia,vol.43,no.6,pp.603,2002. [19] S.Lim,S.Lee,Y.Tsai,I.Chen,P.Tu,J.Chen,H.Chang,Y.Su,andT.Wu,Long-termanteriorthalamusstimulationforintractableepilepsy,ChangGungMedJ,vol.31,no.3,pp.287,2007. [20] J.Bach,U.Ziegler,G.Deuschl,R.Dodel,andG.Doblhammer-Reiter,Happinessisawarmelectrode,PopularScience,pp.57,2007. [21] D.Tweney,Deankamen'srobotarmgrabsmorepublicity,WorldWideWeb, http://blog.wired.com/gadgets/2008/05/dean-kamens-rob.html ,2008. [22] M.Carrozza,P.Dario,F.Vecchi,S.Roccella,M.Zecca,andF.Sebastiani,Thecyberhand:onthedesignofacyberneticprosthetichandintendedtobeinterfacedtotheperipheralnervoussystem,inProc.IEEEConf.IntelligentRobotsandSystems,vol.3,LasVegas,Nevada,2003,pp.2642. [23] M.A.L.Nicolelis,Actionsfromthoughts,Nature,vol.409,pp.403,2001. [24] M.A.L.NicolelisandM.A.Lebedev,Principlesofneuralensemblephysiologyunderlyingtheoperationofbrain-machineinterfaces.NatureReviewsNeuro-science,vol.10,pp.530,2009. [25] M.A.Lebedev,A.J.Tate,T.L.Hanson,Z.Li,J.E.O'Doherty,J.A.Winans,P.J.Ifft,K.Z.Zhuang,N.A.Fitzsimmons,D.A.Schwarz,A.M.Fuller,J.H.An,andM.A.L.Nicolelis,Futuredevelopmentsinbrain-machineinterfaceresearch.Clinics,vol.66,no.S1,pp.25,2011. 159

PAGE 160

[26] H.Berger,berdaselektrenkephalogrammdesmenschen,ArchivfrPsychiatrie,vol.87,pp.527,1929. [27] J.Wolpaw,N.Birbaumer,W.Heetderks,D.McFarland,P.Peckham,G.Schalk,E.Donchin,L.Quatrano,C.Robinson,andT.Vaughan,Brain-computerinterfacetechnology:areviewoftherstinternationalmeeting.IEEETransRehabilEng,vol.8,no.2,pp.164,2000. [28] J.Wolpaw,N.Birbaumer,D.McFarland,G.Pfurtscheller,andT.Vaughan,Brain-computerinterfacesforcommunicationandcontrol.ClinNeurophysiol,vol.113,no.6,pp.767,2002. [29] E.Schmidt,Singleneuronrecordingfrommotorcortexasapossiblesourceofsignalsforcontrolofexternaldevices.AnnBiomedEng,vol.8,no.4-6,pp.339,1980. [30] J.Chapin,K.Moxon,R.Markowitz,andM.Nicolelis,Real-timecontrolofarobotarmusingsimultaneouslyrecordedneuronsinthemotorcortex.NatNeurosci,vol.2,no.7,pp.664,1999. [31] S.Talwar,S.Xu,E.Hawley,S.Weiss,K.Moxon,andJ.Chapin,Ratnavigationguidedbyremotecontrol.Nature,vol.417,no.6884,pp.37,2002. [32] J.Wessberg,C.Stambaugh,J.Kralik,P.Beck,M.Laubach,J.Chapin,J.Kim,S.Biggs,M.Srinivasan,andM.Nicolelis,Real-timepredictionofhandtrajectorybyensemblesofcorticalneuronsinprimates.Nature,vol.408,no.6810,pp.361,2000. [33] D.Taylor,S.Tillery,andA.Schwartz,Directcorticalcontrolof3dneuroprostheticdevices.Science,vol.296,no.5574,pp.1829,2002. [34] J.Carmena,M.Lebedev,R.Crist,J.O'Doherty,D.Santucci,D.Dimitrov,P.Patil,C.Henriquez,andM.Nicolelis,Learningtocontrolabrain-machineinterfaceforreachingandgraspingbyprimates.PLoSBiol,vol.1,no.2,p.E42,2003. [35] M.Velliste,S.Perel,M.Spalding,A.Whitford,andA.Schwartz,Corticalcontrolofaprostheticarmforself-feeding.Nature,vol.453,no.7198,pp.1098,2008. [36] M.Serruya,N.Hatsopoulos,L.Paninski,M.Fellows,andJ.Donoghue,Instantneuralcontrolofamovementsignal.Nature,vol.416,no.6877,pp.141,2002. [37] L.Hochberg,M.Serruya,G.Friehs,J.Mukand,M.Saleh,A.Caplan,A.Branner,D.Chen,R.Penn,andJ.Donoghue,Neuronalensemblecontrolofprostheticdevicesbyahumanwithtetraplegia.Nature,vol.442,no.7099,pp.164,2006. 160

PAGE 161

[38] J.Simeral,S.Kim,M.Black,J.Donoghue,andL.Hochberg,Neuralcontrolofcursortrajectoryandclickbyahumanwithtetraplegia1000daysafterimplantofanintracorticalmicroelectrodearray.JNeuralEng,vol.8,no.2,p.025027,2011. [39] L.Hochberg,D.Bacher,B.Jarosiewicz,N.Masse,J.Simeral,J.Vogel,S.Haddadin,J.Liu,S.Cash,P.vanderSmagt,andJ.Donoghue,Reachandgraspbypeoplewithtetraplegiausinganeurallycontrolledroboticarm.Nature,vol.485,no.7398,pp.372,2012. [40] M.LebedevandM.Nicolelis,Brain-machineinterfaces:past,presentandfuture.TrendsNeurosci,vol.29,no.9,pp.536,2006. [41] A.Sodagar,G.Perlin,Y.Yao,K.Wise,andK.Naja,Animplantablemicrosystemforwirelessmulti-channelcorticalrecording.inSolid-StateSensors,ActuatorsandMicrosystemsConference,2007.TRANSDUCERS2007.International,June2007,pp.69. [42] H.Miranda,V.Gilja,C.Chestek,K.Shenoy,andT.Meng,Hermesd:Ahigh-ratelong-rangewirelesstransmissionsystemforsimultaneousmultichannelneuralrecordingapplications.BiomedicalCircuitsandSystems,IEEETransactionson,vol.4,no.3,pp.181,2010. [43] Y.Song,D.Borton,S.Park,W.Patterson,C.Bull,F.Laiwalla,J.Mislow,J.Simeral,J.Donoghue,andA.Nurmikko,Activemicroelectronicneurosensorarraysforimplantablebraincommunicationinterfaces.IEEETransNeuralSystRehabilEng,vol.17,no.4,pp.339,2009. [44] M.Rizk,C.Bossetti,T.Jochum,S.Callender,M.Nicolelis,D.Turner,andP.Wolf,Afullyimplantable96-channelneuraldataacquisitionsystem.JNeuralEng,vol.6,no.2,p.26002,2009. [45] C.Nordhausen,E.Maynard,andR.Normann,Singleunitrecordingcapabilitiesofa100microelectrodearray.BrainRes,vol.726,no.1-2,pp.129,1996. [46] D.Purves,G.J.Augustine,D.Fitzpatrick,L.C.Katz,A.-S.LaMantia,J.O.McNamara,andS.M.Williams,Eds.,Neuroscience.Sunderland,MA:SinauerAssociates,2004. [47] S.Murphy,Astrocytes:pharmacologyandfunction.SanDiego,CA:AcademicPress,1993. [48] H.KettenmannandB.R.Ransom,Eds.,Neuroglia.NewYork,NY:OxfordUniversityPress,1995. [49] V.Polikov,P.Tresco,andW.Reichert,Responseofbraintissuetochronicallyimplantedneuralelectrodes.JNeurosciMethods,vol.148,no.1,pp.1,2005. [50] T.Trappenberg,FundamentalsofComputationalNeuroscience.NewYork,NY:OxfordUniversityPress,2002. 161

PAGE 162

[51] W.M.Reichert,IndwellingNeuralImplants:StrategiesforContendingwiththeInVivoEnvironment.BocaRaton,FL:CRCPress,2008. [52] R.Clark,J.Schiding,S.Kaczorowski,D.Marion,andP.Kochanek,Neutrophilaccumulationaftertraumaticbraininjuryinrats:comparisonofweightdropandcontrolledcorticalimpactmodels.JNeurotrauma,vol.11,no.5,pp.499,1994. [53] C.Heldin,B.Westermark,andA.Wasteson,Platelet-derivedgrowthfactor.isolationbyalarge-scaleprocedureandanalysisofsubunitcomposition.BiochemJ,vol.193,no.3,pp.907,1981. [54] R.Assoian,A.Komoriya,C.Meyers,D.Miller,andM.Sporn,Transforminggrowthfactor-betainhumanplatelets.identicationofamajorstoragesite,purication,andcharacterization.JBiolChem,vol.258,no.11,pp.7155,1983. [55] R.Mohle,D.Green,M.Moore,R.Nachman,andS.Rai,Constitutiveproductionandthrombin-inducedreleaseofvascularendothelialgrowthfactorbyhumanmegakaryocytesandplatelets.ProcNatlAcadSciUSA,vol.94,no.2,pp.663,1997. [56] J.SilverandJ.Miller,Regenerationbeyondtheglialscar.NatRevNeurosci,vol.5,no.2,pp.146,2004. [57] S.Davies,P.Field,andG.Raisman,Regenerationofcutadultaxonsfailseveninthepresenceofcontinuousalignedglialpathways.ExpNeurol,vol.142,no.2,pp.203,1996. [58] E.Patrick,M.Orazem,J.Sanchez,andT.Nishida,Corrosionoftungstenmicroelectrodesusedinneuralrecordingapplications.JNeurosciMethods,vol.198,no.2,pp.158,2011. [59] A.Schwartz,X.Cui,D.Weber,andD.Moran,Brain-controlledinterfaces:movementrestorationwithneuralprosthetics.Neuron,vol.52,no.1,pp.205,2006. [60] E.Leuthardt,G.Schalk,J.Wolpaw,J.Ojemann,andD.Moran,Abrain-computerinterfaceusingelectrocorticographicsignalsinhumans.JNeuralEng,vol.1,no.2,pp.63,2004. [61] A.Gunduz,M.Ozturk,J.Sanchez,andJ.Principe,Echostatenetworksformotorcontrolofhumanecogneuroprosthetics,inNeuralEngineering,2007.CNE'07.3rdInternationalIEEE/EMBSConferenceon,2007,pp.514. [62] A.Sodagar,G.Perlin,Y.Yao,K.Naja,andK.Wise,Animplantable64-channelwirelessmicrosystemforsingle-unitneuralrecording,Solid-StateCircuits,IEEEJournalof,vol.44,no.9,pp.2591,2009. 162

PAGE 163

[63] G.Santhanam,M.Linderman,V.Gilja,A.Afshar,S.Ryu,T.Meng,andK.Shenoy,Hermesb:acontinuousneuralrecordingsystemforfreelybehavingprimates.IEEETransBiomedEng,vol.54,no.11,pp.2037,2007. [64] C.Chestek,V.Gilja,P.Nuyujukian,R.Kier,F.Solzbacher,R.Ryu,S.andHarrison,andK.Shenoy,Hermesc:Low-powerwirelessneuralrecordingsystemforfreelymovingprimates.IEEETransNeuralSystRehabilEng,vol.17,no.4,pp.330,2009. [65] H.Gao,R.Walker,P.Nuyujukian,K.Makinwa,K.Shenoy,B.Murmann,andT.Meng,Hermese:A96-channelfulldataratedirectneuralinterfacein0.13mcmos,Solid-StateCircuits,IEEEJournalof,vol.47,no.4,pp.1043,april2012. [66] Y.Song,W.Patterson,C.Bull,D.Borton,Y.Li,A.Nurmikko,J.Simeral,andJ.Donoghue,Abrainimplantablemicrosystemwithhybridrf/irtelemetryforadvancedneuroengineeringapplications.inConfProcIEEEEngMedBiolSoc.2007,2007,pp.445. [67] M.Yin,D.Borton,J.Aceros,W.Patterson,andA.Nurmikko,A100-channelhermeticallysealedimplantabledeviceforwirelessneurosensingapplications,inCircuitsandSystems(ISCAS),2012IEEEInternationalSymposiumon,may2012,pp.2629. [68] P.Irazoqui-Pastor,I.Mody,andJ.Judy,Transcutaneousrf-poweredneuralrecordingdevice,inEngineeringinMedicineandBiology,2002.24thAnnualConferenceandtheAnnualFallMeetingoftheBiomedicalEngineeringSocietyEMBS/BMESConference,2002.ProceedingsoftheSecondJoint,vol.3,oct.2002,pp.21052106vol.3. [69] M.YinandM.Ghovanloo,Aexibleclockless32-chsimultaneouswirelessneuralrecordingsystemwithadjustableresolution,inSolid-StateCircuitsConference-DigestofTechnicalPapers,2009.ISSCC2009.IEEEInternational,feb.2009,pp.432,433a. [70] S.Lee,H.Lee,M.Kiani,U.Jow,andM.Ghovanloo,Aninductivelypoweredscalable32-channelwirelessneuralrecordingsystem-on-a-chipforneuroscienceapplications.DigTechPapIEEEIntSolidStateCircuitsConf,vol.4,no.6,pp.360,2010. [71] S.RoyandX.Wang,Wirelessmulti-channelsingleunitrecordinginfreelymovingandvocalizingprimates.JNeurosciMethods,vol.203,no.1,pp.28,2012. [72] H.Grundfest,R.W.Sengstaken,W.H.Oettinger,andR.W.Gurry,Stainlesssteelmicro-needleelectrodesmadebyelectrolyticpointing,ReviewofScienticInstruments,vol.21,no.4,pp.360,1950. 163

PAGE 164

[73] D.Hubel,Tungstenmicroelectrodeforrecordingfromsingleunits.Science,vol.125,no.3247,pp.549,1957. [74] M.Wolbarsht,E.MacnicholJr,andH.Wagner,Glassinsulatedplatinummicroelectrode.Science,vol.132,no.3436,pp.1309,1960. [75] C.Guld,Aglass-coveredplatinummicroelectrode.MedElectronBiolEng,vol.2,pp.317,1964. [76] M.SalcmanandM.Bak,Design,fabrication,andinvivobehaviorofchronicrecordingintracorticalmicroelectrodes.IEEETransBiomedEng,vol.20,no.4,pp.253,1973. [77] ,Anewchronicrecordingintracorticalmicroelectrode.MedBiolEng,vol.14,no.1,pp.42,1976. [78] J.Williams,R.Rennaker,andD.Kipke,Long-termneuralrecordingcharacteristicsofwiremicroelectrodearraysimplantedincerebralcortex.BrainResBrainResProtoc,vol.4,no.3,pp.303,1999. [79] TDTMicrowireArray,WorldWideWeb, http://www.tdt.com/products/MW16.htm ,2012. [80] MicroprobesInc.Multi-ChannelArrays,WorldWideWeb, http://www.microprobes.com/index.php?option=com content&view=article&id=135&Itemid=82 ,2012. [81] K.D.Wise,J.B.Angell,andA.Starr,Anintegrated-circuitapproachtoextracellularmicroelectrodes,BiomedicalEngineering,IEEETransactionson,vol.BME-17,no.3,pp.238,july1970. [82] K.Naja,K.Wise,andT.Mochizuki,Ahigh-yieldic-compatiblemultichannelrecordingarray.ElectronDevices,IEEETransactionson,vol.32,no.7,pp.1206,1985. [83] K.NajaandK.Wise,Animplantablemultielectrodearraywithon-chipsignalprocessing,Solid-StateCircuits,IEEEJournalof,vol.21,no.6,pp.10351044,dec1986. [84] J.JiandK.Wise,Animplantablecmoscircuitinterfaceformultiplexedmicroelectroderecordingarrays,Solid-StateCircuits,IEEEJournalof,vol.27,no.3,pp.433,mar1992. [85] J.Hetke,J.Lund,K.Naja,K.Wise,andD.Anderson,Siliconribboncablesforchronicallyimplantablemicroelectrodearrays.IEEETransBiomedEng,vol.41,no.4,pp.314,1994. [86] A.HoogerwerfandK.Wise,Athree-dimensionalmicroelectrodearrayforchronicneuralrecording.IEEETransBiomedEng,vol.41,no.12,pp.1136,1994. 164

PAGE 165

[87] K.Wise,D.Anderson,J.Hetke,D.Kipke,andK.Naja,Wirelessimplantablemicrosystems:high-densityelectronicinterfacestothenervoussystem,Proceed-ingsoftheIEEE,vol.92,no.1,pp.7697,jan2004. [88] P.Campbell,K.Jones,R.Huber,K.Horch,andR.Normann,Asilicon-based,three-dimensionalneuralinterface:manufacturingprocessesforanintracorticalelectrodearray.IEEETransBiomedEng,vol.38,no.8,pp.758,1991. [89] P.RouscheandR.Normann,Chronicrecordingcapabilityoftheutahintracorticalelectrodearrayincatsensorycortex.JNeurosciMethods,vol.82,no.1,pp.1,1998. [90] A.Branner,R.Stein,andR.Normann,Selectivestimulationandrecordingusingaslantedmultielectrodearray,in[EngineeringinMedicineandBiology,1999.21stAnnualConf.andthe1999AnnualFallMeetingoftheBiomedicalEngineeringSoc.]BMES/EMBSConference,1999.ProceedingsoftheFirstJoint,vol.1,1999,p.377vol.1. [91] P.Norlin,M.Kindlundh,A.Mouroux,K.Yoshida,andU.G.Hofmann,A32-siteneuralrecordingprobefabricatedbydrieofsoisubstrates,JournalofMicrome-chanicsandMicroengineering,vol.12,no.4,p.414,2002. [92] M.Kindlundh,P.Norlin,andU.G.Hofmann,Aneuralprobeprocessenablingvariableelectrodecongurations,SensorsandActuatorsB:Chemical,vol.102,no.1,pp.5158,2004. [93] J.Muthuswamy,M.Okandan,andN.Jackson,Singleneuronalrecordingsusingsurfacemicromachinedpolysiliconmicroelectrodes.JNeurosciMethods,vol.142,no.1,pp.45,2005. [94] P.Rousche,D.Pellinen,D.PivinJr,J.Williams,R.Vetter,andD.Kipke,Flexiblepolyimide-basedintracorticalelectrodearrayswithbioactivecapability.IEEETransBiomedEng,vol.48,no.3,pp.361,2001. [95] D.Pellinen,T.Moon,R.Vetter,R.Miriani,andD.Kipke,Multifunctionalexibleparylene-basedintracorticalmicroelectrodes.inConfProcIEEEEngMedBiolSoc.,vol.5,2005,pp.5272. [96] J.SeymourandD.Kipke,NeuralprobedesignforreducedtissueencapsulationinCNS.Biomaterials,vol.28,pp.3594,2007. [97] T.StieglitzandM.Gross,FlexibleBIOMEMSwithelectrodearrangementsonfrontandbacksideaskeycomponentinneuralprosthesesandbiohybridsystems.SensorsandActuatorsB-Chemical,vol.83,no.1-3,pp.8,2002. [98] T.Stieglitz,M.Schuetter,andK.Koch,Implantablebiomedicalmicrosystemsforneuralprostheses,EngineeringinMedicineandBiologyMagazine,IEEE,vol.24,no.5,pp.5865,sept.-oct.2005. 165

PAGE 166

[99] S.Takeuchi,T.Suzuki,K.Mabuchi,andH.Fujita,dexiblemultichannelneuralprobearray,JournalofMicromechanicsandMicroengineering,vol.14,no.1,p.104,2004. [100] Y.Kato,I.Saito,T.Hoshino,T.Suzuki,andK.Mabuchi,Preliminarystudyofmultichannelexibleneuralprobescoatedwithhybridbiodegradablepolymer,inEngineeringinMedicineandBiologySociety,2006.EMBS'06.28thAnnualInternationalConferenceoftheIEEE,302006-sept.32006,pp.660. [101] R.Biran,D.Martin,andP.Tresco,Thebraintissueresponsetoimplantedsiliconmicroelectrodearraysisincreasedwhenthedeviceistetheredtotheskull,JBiomedMaterResA,vol.82,no.1,pp.169,2007. [102] Y.Kim,R.Hitchcock,M.Bridge,andP.Tresco,Chronicresponseofadultratbraintissuetoimplantsanchoredtotheskull.Biomaterials,vol.25,no.12,pp.2229,2004. [103] R.Biran,D.Martin,andP.Tresco,Neuronalcelllossaccompaniesthebraintissueresponsetochronicallyimplantedsiliconmicroelectrodearrays.ExpNeurol,vol.195,no.1,pp.115,2005. [104] A.GillettiandJ.Muthuswamy,Brainmicromotionaroundimplantsintherodentsomatosensorycortex.JNeuralEng,vol.3,no.3,pp.189,2006. [105] H.Lee,R.Bellamkonda,W.Sun,andM.Levenston,Biomechanicalanalysisofsiliconmicroelectrode-inducedstraininthebrain.JNeuralEng,vol.2,no.4,pp.81,2005. [106] J.Subbaroyan,D.Martin,andD.Kipke,Anite-elementmodelofthemechanicaleffectsofimplantablemicroelectrodesinthecerebralcortex,JBiomedMaterResA,vol.2,no.4,pp.103,2005. [107] J.SubbaroyanandD.Kipke,Theroleofexiblepolymerinterconnectsinchronictissueresponseinducedbyintracorticalmicroelectrodes-amodelingandaninvivostudy,inEngineeringinMedicineandBiologySociety,2006.EMBS'06.28thAnnualInternationalConferenceoftheIEEE,302006-sept.32006,pp.3588. [108] Y.Yao,M.Gulari,B.Casey,J.Wiler,andK.Wise,Siliconmicroelectrodeswithexibleintegratedcablesforneuralimplantapplications,in3rdInternationalIEEE/EMBSConferenceonNeuralEngineering,2007,pp.398. [109] P.J.Gilgunn,R.Khilwani,T.D.Y.Kozai,D.J.Weber,X.T.Cui,G.Erdos,O.B.Ozdoganlar,andG.K.Fedder,Anultra-compliantscalableneuralprobewithmoldedbiodissolvabledeliveryvehicle.in25thAnnualInternationalConferenceoftheIEEEMEMS,2012,pp.56. 166

PAGE 167

[110] E.Kim,H.Tu,C.Lv,H.Jiang,H.Yu,andY.Xu,Arobustpolymermicrocablestructureforexibledevices.App.Phy.Letters,vol.102,p.033506,2013. [111] J.Thelin,H.Jrntell,E.Psouni,M.Garwicz,J.Schouenborg,N.Danielsen,andC.Linsmeier,Implantsizeandxationmodestronglyinuencetissuereactionsinthecns.PLoSOne,vol.6,no.1,p.e16267,2011. [112] M.Ward,P.Rajdev,C.Ellison,andP.Irazoqui,Towardacomparisonofmicroelectrodesforacuteandchronicrecordings.BrainRes,vol.1282,pp.183,2009. [113] G.Barillaro,A.Molfese,A.Nannini,andF.Pieri,Analysis,simulationandrelativeperformancesoftwokindsofserpentinesprings,J.Micromech.Microeng.,vol.15,no.4,pp.736,2005. [114] G.K.Fedder,Simulationsofmicroelectromechanicalsystems,Ph.D.dissertation,UniversityofCaliforniaatBerkeley,1994. [115] DupontKaptonBTechnicalDataSheet,WorldWideWeb, http://www2.dupont.com/Kapton/en US/assets/downloads/pdf/Kapton B.pdf ,2012. [116] J.DolbowandM.Gosz,Effectofout-of-planepropertiesofapolyimidelmonthestresseldsinmicroelectronicstructures,MechanicsofMaterials,vol.23,pp.311,1996. [117] HDMicrosystemsPI2611ProcessGuide,WorldWideWeb, http://hdmicrosystems.com/HDMicroSystems/en US/pdf/PI-2600 ProcessGuide.pdf ,2012. [118] ABAQUSAnalysisUser'sManual(6.12),WorldWideWeb, http://abaqus.me.chalmers.se/v6.12/books/usb/default.htm?startat=pt03ch07s02aus51.html ,2013. [119] Series:DualColumnTabletop(5kn-50kn)Manual,InstronInc.WorldWideWeb, http://www.instron.us/wa/product/5900-Series-Mechanical-Testing-Systems.aspx ,2012. [120] Ti900TriboindenterManual,HysitronInc.WorldWideWeb, http://www.hysitron.com ,2012. [121] A.PrasadandJ.Sanchez,Quantifyinglong-termmicroelectrodearrayfunctionalityusingchronicinvivoimpedancetesting.JNeuralEng,vol.9,no.2,p.026028,2012. [122] J.BockrisandA.Reddy,ModernElectrochemistry.NewYork,NY:PlenumPress,1970. [123] J.Newman,ElectrochemicalSystems.Hoboken,NJ:JohnWileyandSons,Inc.,2004. 167

PAGE 168

[124] A.Hassibi,R.Navid,R.Dutton,andT.Lee,Comprehensivestudyofnoiseprocessesinelectrodeelectrolyteinterfaces,JournalofAppliedPhysics,vol.96,pp.1074,2004. [125] G.Brug,A.V.denEeden,M.Sluyters-Rehbach,andJ.Sluyters,Theanalysisofelectrodeimpedancescomplicatedbythepresenceofaconstantphaseelement.Electroanal.Chem.,vol.176,pp.275,1984. [126] G.R.Holt,Acriticalreexaminationofsomeassumptionsandimplicationsofcabletheoryinneurobiology,Ph.D.dissertation,CaliforniaInstituteofTechnology,1998. [127] J.Williams,J.Hippensteel,J.Dilgen,W.Shain,andD.Kipke,Compleximpedancespectroscopyformonitoringtissueresponsestoinsertedneuralimplants.JNeuralEng,vol.4,no.4,pp.410,2007. [128] J.Buitenweg,W.Rutten,W.Willems,andJ.vanNieuwkasteele,Measurementofsealingresistanceofcell-electrodeinterfacesinneuronalculturesusingimpedancespectroscopy.MedBiolEngComput,vol.36,no.5,pp.630,1998. [129] A.Kyle,C.Chan,andA.Minchinton,Characterizationofthree-dimensionaltissueculturesusingelectricalimpedancespectroscopy.BiophysJ,vol.76,no.5,pp.2640,1999. [130] B.RubehnandT.Stieglitz,Invitroevaluationofthelong-termstabilityofpolyimideasamaterialforneuralimplants.Biomaterials,vol.31,no.13,pp.3449,2010. [131] S.Ordonez,C.Boehler,M.Schuettler,andS.T,Long-termadhesionstudiesofpolyimidetoinorganicandmetalliclayers,inMRSProceedings,vol.1466,2012. [132] K.Lee,J.He,R.Clement,S.Massia,andB.Kim,Biocompatiblebenzocyclobutene(BCB)-basedneuralimplantswithmicro-uidicchannel.BiosensBioelectron,vol.20,no.2,pp.404,2004. [133] Y.NishiandR.Doering,Eds.,HandbookofSemiconductorManufacturingTechnology.CRCPress,2007. [134] C.Hassler,T.Boretius,andT.Stieglitz,Polymersforneuralimplants,JournalofPolymerSciencePartB:PolymerPhysics,vol.49,no.1,pp.18,2011. [135] M.Dhindsa,S.Kuiper,andJ.Heikenfeld,Reliableandlow-voltageelectrowettingonthinparylenelms,ThinSolidFilms,vol.519,no.10,pp.33463351,2011. [136] C.Hassler,R.vonMetzen,P.Ruther,andT.Stieglitz,Characterizationofparylenecasanencapsulationmaterialforimplantedneuralprostheses.JBiomedMaterResBApplBiomater,vol.93B,no.1,pp.266274,2010. 168

PAGE 169

[137] A.Prasad,Q.Xue,V.Sankar,T.Nishida,G.Shaw,W.Streit,andJ.Sanchez,Comprehensivecharacterizationandfailuremodesoftungstenmicrowirearraysinchronicneuralimplants.JNeuralEng,vol.9,no.5,p.056015,2012. [138] W.Streit,Q.Xue,A.Prasad,V.Sankar,E.Knott,A.Dyer,J.Reynolds,T.Nishida,G.Shaw,andJ.Sanchez,Electrodefailure:tissue,electrical,andmaterialresponses.IEEEPulse,vol.3,no.1,pp.30,2012. [139] J.Sanchez,N.Alba,T.Nishida,C.Batich,andP.Carney,Structuralmodicationsinchronicmicrowireelectrodesforcorticalneuroprosthetics:acasestudy.IEEETransNeuralSystRehabilEng,vol.14,no.2,pp.217,2006. [140] T.J.Webster,M.C.Waid,J.L.McKenzie,R.L.Price,andJ.U.Ejiofor,Nano-biotechnology:carbonnanobresasimprovedneuralandorthopaedicimplants,Nanotechnology,vol.15,no.1,pp.48,2004. [141] V.Lovat,D.Pantarotto,L.Lagostena,B.Cacciari,M.Grandolfo,M.Righi,G.Spalluto,M.Prato,andL.Ballerini,Carbonnanotubesubstratesboostneuronalelectricalsignaling.NanoLett,vol.5,no.6,pp.1107,2005. [142] T.Gabay,M.Ben-David,I.Kalifa,R.Sorkin,Z.Abrams,E.Ben-Jacob,andY.Hanein,Electro-chemicalandbiologicalpropertiesofcarbonnanotubebasedmulti-electrodearrays.Nanotechnology,vol.18,no.3,p.35201,2007. [143] E.Keefer,B.Botterman,M.Romero,A.Rossi,andG.Gross,Carbonnanotubecoatingimprovesneuronalrecordings.NatNanotechnol,vol.3,no.7,pp.434,2008. [144] J.YangandD.Martin,Microporousconductingpolymersonneuralmicroelectrodearrays:Ii.physicalcharacterization.SensorsandActuatorsA:Physical,vol.113,pp.204,2004. [145] M.Abidian,J.Corey,D.Kipke,andD.Martin,Conducting-polymernanotubesimproveelectricalproperties,mechanicaladhesion,neuralattachment,andneuriteoutgrowthofneuralelectrodes.Small,vol.6,no.3,pp.421,2010. [146] M.Abidian,L.Salas,A.Shahmorad,T.Marzullo,D.Martin,andD.Kipke,In-vivoevaluationofchronicallyimplantedneuralmicroelectrodearraysmodiedwithpoly(3,4-ethylenedioxythiophene)nanotubes,inProceedingsofthe3rdInternationalIEEEEMBSConferenceonNeuralEngineering,2007,2007,pp.61. [147] S.Venkatraman,J.Hendricks,Z.King,A.Sereno,S.Richardson-Burns,D.Martin,andJ.Carmena,Invitroandinvivoevaluationofpedotmicroelectrodesforneuralstimulationandrecording.IEEETransNeuralSystRehabilEng,vol.19,no.3,pp.307,2011. [148] B.BurkeandC.E.Lewis,TheMacrophage.NewYork,NY:OxfordUniversityPress,2002. 169

PAGE 170

[149] J.ShiandL.Eastman,CorrelationbetweenAlGaN/GaNMISHFETperformanceandHfO2insulationlayerquality,IEEEElectrodeDeviceLetters,vol.32,no.3,pp.312,2011. [150] M.Jones,Y.Kwon,andD.Norton,DielectricconstantandcurrenttransportforHfo2thinlmsonITO,AppliedPhysicsA:MaterialsScience&Processing,vol.81,pp.285,jul2005. [151] Y.Chen,M.Liu,T.Kanekoa,andP.McIntyre,Atomiclayerdepositedhafniumoxidegatedielectricsforcharge-basedbiosensors.Electrochem.Solid-StateLett.,vol.13,pp.G29G32,2010. [152] C.Palmer,Amicrowiretechniqueforrecordingsingleneuronsinunrestrainedanimals.BrainResBull,vol.3,no.3,pp.285,1978. [153] G.WestbyandH.Wang,Aoatingmicrowiretechniqueformultichannelchronicneuralrecordingandstimulationintheawakefreelymovingrat.JNeurosciMethods,vol.76,no.2,pp.123,1997. [154] X.Zhang,R.Zhang,Z.Zhang,andM.Chopp,Measurementofneuronalactivityofindividualneuronsafterstrokeintheratusingamicrowireelectrodearray.JNeurosciMethods,vol.162,no.1-2,pp.91,2007. [155] G.Kotzar,M.Freas,P.Abel,A.Fleischman,S.Roy,C.Zorman,J.Moran,andJ.Melzak,Evaluationofmemsmaterialsofconstructionforimplantablemedicaldevices.Biomaterials,vol.23,no.13,pp.2737,2002. [156] G.Voskerician,M.Shive,R.Shawgo,H.vonRecum,J.Anderson,M.Cima,andR.Langer,Biocompatibilityandbiofoulingofmemsdrugdeliverydevices.Biomaterials,vol.24,no.11,pp.1959,2003. [157] M.Hennemeyer,F.Walther,S.Kerstan,K.Schurzinger,A.M.Gigler,andR.W.Stark,Cellproliferationassaysonplasmaactivatedsu-8,Microelectron.Eng.,vol.85,no.5-6,pp.1298,2008. [158] S.TaoandT.Desai,Microfabricationofmultilayer,asymmetric,polymericdevicesfordrugdelivery,AdvancedMaterials,vol.17,no.13,pp.1625,2005. [159] S.-H.Cho,H.M.Lu,L.Cauller,M.Romero-Ortega,J.-B.Lee,andG.Hughes,Biocompatiblesu-8-basedmicroprobesforrecordingneuralspikesignalsfromregeneratedperipheralnervebers,SensorsJournal,IEEE,vol.8,no.11,pp.1830,nov.2008. [160] A.Modafe,Benzocyclobutene-basedElectricMicromachinesSupportedonMicroballBearings:Design,Fabrication,andCharacterization,Ph.D.dissertation,UniversityofMaryland,2007. 170

PAGE 171

[161] S.-S.LinandC.-S.Liao,Thehydrophobicityandopticalpropertiesofthehfo2-depositedglass,CeramicsInternational,vol.39,no.1,pp.353358,2013. [162] Y.Zhao,Designofhigher-kandmorestablerareearthoxidesasgatedielectricsforadvancedCMOSdevices,Materials,vol.5,no.8,pp.1413,2012. [163] J.Field,A.Luna-Velasco,S.Boitano,F.Shadman,B.Ratner,C.Barnes,andR.Sierra-Alvarez,Cytotoxicityandphysicochemicalpropertiesofhafniumoxidenanoparticles.Chemosphere,vol.84,no.10,pp.1401,2011. [164] M.Lee,N.Zine,A.Baraket,M.Zabala,F.Campabadal,R.Caruso,M.G.Trivella,N.Jaffrezic-Renault,andA.Errachid,Anovelbiosensorbasedonhafniumoxide:Applicationforearlystagedetectionofhumaninterleukin-10,SensorsandActuatorsB:Chemical,no.0,pp.,2012. [165] P.R.Gray,P.J.Hurst,S.H.Lewis,andR.G.Meyer,AnalysisandDesignofAnalogIntegratedCircuits.NewYork:Wiley,2009. [166] MITMaterialPropertyDatabase,WorldWideWeb, http://www.mit.edu/6.777/matprops/pecvd sin.htm ,2012. [167] PayleneEngineering-TypicalSpecicationofParylene-C,WorldWideWeb, http://www.paryleneengineering.com/specications parylene c astm.html ,2012. [168] H.YoungandR.Freedman,UniversityPhysics.Addison-Wesley,2011. [169] G.J.Kim,G.H.Kim,andY.K.Yoon,Fabricationofhighenergydensitycpaacitorswithmicromachinedcarbonnanoberelectrodes,inPowerMEMSConference,2011,pp.1. [170] ProcessingProceduresforCYCLOTENE4000SeriesPhotoBCBResins:DS3000ImmersionDevelopProcess,WorldWideWeb, http://www.dow.com/cyclotene/docs/cyclotene 4000 immersion dev.pdf ,2012. [171] ProcessingProceduresforCYCLOTENE3000SeriesDryEtchResins,WorldWideWeb, http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh 08b4/0901b803808b4ed7.pdf?lepath=/888-00006.pdf&fromPage=GetDoc ,2012. [172] P.F.Jao,S.-P.Fang,E.David,K.T.Kim,andY.-K.Yoon,Nanomanufacturingoflargeareacarbonnanobersusingtubenozzleelectrospinning(TNE),lithographyandcarbonizationprocesses,inElectronicComponentsandTechnologyConfer-ence(ECTC),2012IEEE62nd,292012-june12012,pp.2075. [173] K.LaFlamme,K.Popat,L.Leoni,E.Markiewicz,T.LaTempa,B.Roman,C.Grimes,andT.Desai,Biocompatibilityofnanoporousaluminamembranesforimmunoisolation.Biomaterials,vol.28,no.16,pp.2638,2007. 171

PAGE 172

[174] S.D.Alvarez,Stabilityandbiocompatabilityofporoussiliconandporousaluminaforcellandbiomolecularsensing,Ph.D.dissertation,UniversityofCaliforniaatSanDiego,2008. [175] W.AhmedEl-Said,C.Yea,M.Jung,H.Kim,andJ.Choi,Analysisofeffectofnanoporousaluminasubstratecoatedwithpolypyrrolenanowireoncellmorphologybasedonafmtopography.Ultramicroscopy,vol.110,pp.676,2010. [176] C.Ran,G.Ding,W.Liu,Y.Deng,andW.Hou,Wettingonnanoporousaluminasurface:Transitionbetweenwenzelandcassiestatescontrolledbysurfacestructure.Langmuir,vol.24,pp.9952,2008. 172

PAGE 173

BIOGRAPHICALSKETCH ViswanathSankarwasborninMadras(Chennai),Indiain1982.HereceivedhisBachelorofEngineeringdegreeinelectricalandelectronicsengineeringfromtheUniversityofMadrasinMay2003andhisMasterofSciencedegreeinelectricalengineeringfromtheUniversityofIllinoisatChicagoinMay2006.SinceFall2007,ViswanathhasbeenaresearchassistantintheInterdisciplinaryMicrosystemsGroup(IMG)attheUniversityofFloridaworkingwithDr.ToshikazuNishidaonMicroelectroMechanicalSystems(MEMS)basedneuralelectrodes.Hisresearchinterestincludeinvestigatingandimprovingthereliabilityofmicroelectrodesusedinchronicneuralrecordingapplications.HeworkedasaprocessengineerinternatIntelCorporationfromFebruary2012toAugust2012.HereceivedhisPh.DdegreeinelectricalengineeringfromtheUniversityofFloridainAugust2013.HeiscurrentlywithIntelCorporationasaprocessengineer. 173