Optimizing electrode thickness and material for laser-induced electrothermal flow

Transcription

1 Purdue University Purdue e-pubs Open Access Theses Theses and Dissertations Spring 2015 Optimizing electrode thickness and material for laser-induced electrothermal flow Jian Wei Khor Purdue University Follow this and additional works at: Part of the Mechanical Engineering Commons Recommended Citation Khor, Jian Wei, "Optimizing electrode thickness and material for laser-induced electrothermal flow" (2015). Open Access Theses This document has been made available through Purdue e-pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu for additional information.

2 Graduate School Form 30 Updated 1/15/2015 PURDUE UNIVERSITY GRADUATE SCHOOL Thesis/Dissertation Acceptance This is to certify that the thesis/dissertation prepared By Jian-Wei Khor Entitled OPTIMIZING ELECTRODE THICKNESS AND MATERIAL FOR LASER-INDUCED ELECTROTHERMAL FLOW For the degree of Master of Science in Mechanical Engineering Is approved by the final examining committee: Steven T. Wereley Chair Jun Chen Liang Pan To the best of my knowledge and as understood by the student in the Thesis/Dissertation Agreement, Publication Delay, and Certification Disclaimer (Graduate School Form 32), this thesis/dissertation adheres to the provisions of Purdue University s Policy of Integrity in Research and the use of copyright material. Approved by Major Professor(s): Steven T. Wereley Approved by: Ganesh Subbarayan 4/15/2015 Head of the Departmental Graduate Program Date

3

4 OPTIMIZING)ELECTRODE)THICKNESS)AND)MATERIAL)FOR)LASER4INDUCED) ELECTROTHERMAL)FLOW) ) A)Thesis) )Submitted)to)the)Faculty)) of) Purdue)University)) by) Jian4Wei)Khor) ) In)Partial)Fulfillment)of)the)) Requirements)for)the)Degree)) of) Master)of)Science)in)Mechanical)Engineering) ) May)2015) )Purdue)University) )West)Lafayette,)Indiana) )

5 ii ACKNOWLEDGMENTS Firstandforemost,Iwouldliketothankmyparentsforsupportingmefinancially andemotionallythroughoutmygraduatestudiesatpurdueuniversity.theyhave providedmewitheaseofmind,whichcontributedtomytotalfocusinmystudies andresearch.iwouldalsoliketothankmyfiancé,lingsimhong,forencouraging metopursuemygoalsandhadfaithinmyendeavors. IwouldalsoliketothankProfessorStevenWereleywhoismyacademicadvisorfor guidingmethroughmygraduateresearch.hehasinspiredmetodogoodresearch andmademeunderstandtheworkingsofacademia.iamalsogratefulforbeingable tousetheoverpricedequipmentsinhisresearchlab. Iwouldliketothankmylabmates,AvanishMishra,KatieClayton,RavirajThakur, andnathanmiller,labalumni,dr.stuartwilliams,dr.craigsnoeyink,anddr. JaesungKwon,andpreviousvisitingscholars,Dr.XudongPanandComlanMagloire, forgivingmeadviceandguidanceontheproperprocedurestoconductresearch.i wouldliketospecificallyacknowledgementavanishmishraashehasbeenmy mentorforallofmyundergraduateandgraduateresearch.

6 iii Avanishhasfollowedmyresearchverycloselyandhastrainedmetobecomea betterresearcher.

7 iv TABLEOFCONTENTS Page LISTOFTABLES..v LISTOFFIGURES...vi ABSTRACT..ix CHAPTER1:BRIEFINTRODUCTIONOFMICROFLUIDICS CHAPTER2:RAPIDELECTROKINETICPATTERNING Introduction Physics ElectrothermalFlow Particle[ElectrodeInteraction Dielectrophoresis ParticleDipole[DipoleAttraction ParametersforParticleClustering Methodology CurrentApplications DifferentiationofBiologicalOrganisms PlasmonicNanostructuretoOptimizeREP REPinCoplanarElectrodes ThinFilmHeatersasHeatSourceforREP ConclusionsandFutureWork CHAPTER3:OPTIMIZINGELECTRODETHICKNESSANDMATERIALFOR LASER[INDUCEDELECTROTHERMALFLOW Introduction Theory ReflectionoftheLaserattheElectrodeSurface ProductionofTemperatureGradientinElectrodes AnalyticalSolutionofLaserHeatingontoaThinSubstrate ACElectricFieldinSolution FluidMechanicsandElectrothermalFlow MethodandMaterials ChipDesignandFabrication ExperimentalSetup ParticlePreparation SoftwareandImageProcessing NumericalSimulations ProductionofTemperatureGradientinElectrodes ACElectricFieldinSolution..47

8 v Page 3.4.3ElectrothermalFlow Simulation ComparisonBetweenExperimentalResultsandComputational Results ResultsandDiscussions AnalyticalResultsofTemperatureDifferenceperLaserPower AnalyticalSolutionandComputationalResultsofthe TemperatureGradient ComparisonofFlowVelocityofElectrothermalFlowforMaterials ofsimilarthickness OptimalMaterialThicknessforElectrothermalFlow Conclusion LISTOFREFERENCES...65 VITA...68

9 vi LISTOFTABLES Table Page Table1:Valuesofimportantpropertiesfornumericalsimulation..49 Table2:Propertiesofvariousmaterials...54

10 vii LISTOFFIGURES Figure Page Figure1:RapidElectrokineticPatterning(REP).(a)Velocityprofileof electrothermalflow.(b)a3dtrajectoryoftheelectrothermalflow at200khz.(fig.1bprintedwithpermissionfromkumaretal. Copyright2010Springer)[5].(c[e)Particlesbeingtrappedintoa clusterusingrep Figure2:PhysicsofRapidElectrokineticPatterning.(a)Forcerelationshipona particleduringrep.(b[c)theclusteringofparticleswhenlaseris turnedon.(d)thetemperatureincreseonanitocoverslipand(e) theeffectofdifferentlaserpowersandthetemperatureprofile produced.(fig.2dand2ereprintedwithpermissionfromkumaret al.copyright2010americanchemicalsociety)[6]...6 Figure3:PolarizationofParticles.(b)showsaneutralparticlewhenthereisno appliedelectricfieldand(c)showsapolarizedparticleduetoan electricfield...8 Figure4:ElectricDoubleLayer Figure5:Positiveandnegativedielectrophoresisformorepolarizableparticle andlesspolarizableparticleinuniform(a)andnon[uniformelectric field(b)..13 Figure6:PlotoftheClausius[Mossottifactorfortwodifferentparticles(solid anddottedlines)withrespecttofrequencyandtheseparationof ParticlesusingDielectrophoresis Figure7:DipoleMoment Figure8:AttractionoftwoFixedDipoles Figure9:Particle[particleInteractionwheretwoparticlesarealignedparallel totheelectricfield.(a)hasparticlesthataremoreconductivethan thesolutionand(b)hasparticlesthatarelessconductivethanthe solution Figure10:Particle[particleInteractionwheretwoparticlesthatarealigned perpendiculartotheelectricfield Figure11:Particle[particleInteractionwheretwodifferentparticlesare alignedparallel(a)andperpendicular(b)totheelectricfield...21 Figure12:Chainformingofdifferentparticlesinasolutionappliedwithan electricfield

11 viii Figure Page Figure13:Accumulationof1.0umparticlesattheITOelectrodeoveraperiod of3minforvariouspeak[to[peakvoltages.laserpoweris23mw andacfrequencyis70khz.( IOPPublishing.Reproducedfrom Williamsetal.bypermissionofIOPPublishing)[7]..24 Figure14:Accumulationof1.0umparticlesattheITOelectrodeoveraperiod of3minforvariousfrequencies.laserpoweris23mwandac peak[to[peakvoltageis15.3vpp.( IOPPublishing.Reproduced fromwilliamsetal.bypermissionofioppublishing)[7].25 Figure15:Accumulationof1.0umparticlesattheITOelectrodeoveraperiod of3minforvariousfrequencies.acpeak[to[peakvoltageis15.3vpp. andacfrequenctyis70khz.( IOPPublishing.Reproducedfrom Williamsetal.bypermissionofIOPPublishing)[7] Figure16:Maximumtrappingfrequencyforparticlesofvariousdiametersto clusteratdifferentappliedacpeak[to[peakvoltage.laserpoweris 23mW.[7] Figure17:REPChip Figure18:REPSetup Figure19:TrappingandTranslationofS.oneidensis.(Reprintedwith permissionfromkwonetal.copyright2012royalsocietyof Chemistry)[18] Figure20:SeparationofS.aureusfromS.cerevisae.(Reprintedwith permissionfromkwonetal.copyright2012royalsocietyof Chemistry)[18] Figure21:Plasmonicnanostructures:Thevelocitydistribution(a)andradial velocity(b).(reprintedwithpermissionfromndukaifeetal. Copyright2014AmericanChemicalSociety)[19] Figure22:TwinOpposingMicrovortices(TOMV).(Reprintedwithpermission fromparketal.copyright2012royalsocietyofchemistry)[20] 33 Figure23:ConductingREPusingthinfilmheatersasheatsource(Reprinted withpermissionfromvelascoetal.copyright2012elsevier)[21& 22].34 Figure24:IllustrationofREPwithlaserfocusedatthetopelectrode...36 Figure25:PropagationofLight(ReflectionandRefraction)inMaterials...38 Figure26:ExperimentalSetupandMethod 45 Figure27:(a)Boundaryconditionsofthegeometryinthesimulation.E:AC ElectricFieldModel.H:HeatTransferModel.F:ComputationalFluid DynamicsModel.(b)Vectorfieldofsimulatedelectrothermalflowin logarithmicscale.(c)contourplotofsimulated electrothermalflow 48 Figure28:FinalGeometry..50 Figure29:Freetriangularmeshing(topandbottom)andquadrilateral meshing(center)..52

12 ix Figure Page Figure30:Comparisonofvelocityprofilesofelectrothermalflowof40nmTi fromexperimentalresultsandsimulationatthesamelocation...54 Figure31:TemperatureriseperlaserpowerofNiat(a)andTiat(b)from analyticalresults..55 Figure32:TemperaturegradientofNiat(a)andTiat(b)fromanalytical results.57 Figure33:SteepesttemperaturegradientforITO,NiandTifromanalytical results...58 Figure34:TemperaturegradientofNiat(a)andTiat(b)fromcomputational results Figure35:SteepesttemperaturegradientforITO,NiandTifromcomputation results...60 Figure36:ExperimentalvelocityprofileofelectrothermalflowbyITO,Niand Tiofsimilarthicknessesat30Vpeaktopeakand201.8µmabove thebottomsubstrateinthechannel 60 Figure37:Therelationshipofthemaximumvelocityproducedinthe correspondingthicknessforitoat(a)andniandtiat(b)at30v peaktopeakandat201.8µmabovethebottomsubstrateinthe channel...62

13 x ABSTRACT Khor, Jian[Wei, M.S.M.E., Purdue University, May Optimizing Electrode Thickness and Material for Laser[Induced Electrothermal Flow. Major Professor: StevenT.Wereley,SchoolofMechanicalEngineering Inthisarticle,thematerialdependenceofelectrothermalflowisstudiedtooptimize an optoelectric particle manipulation technique called Rapid Electrokinetic Patterning (REP). REP utilizes a heat source to produce temperature gradients, which, in this study, are induced by the application of a 1064 nm cw laser. The simultaneous application of an uniform AC electric field assemble colloids on the surfaceofanelectrodethatcanbemovedbytranslatingthelaser.thetemperature gradients produced from the heat source are crucial in REP as the higher the temperature gradient the faster the electrothermal flow. Thus, optimization of temperaturegradientswouldenhancerepperformancewithlowerlaserpower.in this study, material properties of the electrode that produces the temperature gradientinrepareinvestigatedtounderstandtheabsorptionoftheradiationfrom the laser. Analytical, computational and experimental aspects of the phenomenon were studied. It was found that the analytical model and the computation model could be used to select the optimum material and thickness of the electrode to

14 xi producethehighesttemperaturegradientperwattoflaserpower.furthermore,the computational model can be used to determine the flow velocity of the electrothermal flow accurately. With this study material selection of the electrode can be calibrated and customized for specific needs when conducting REP. It was foundthatamongthematerialstested(ito,titanium,andnickel),thebestmaterial that produced the highest temperature gradient was titanium followed by nickel and ITO. When comparing the three materials with approximately the same thicknesses, it was observed that the maximum velocity magnitude produced by titaniumandnickelelectrodeswerealmost80timeshigherthantheitoelectrodes.

15 1 CHAPTER1:BRIEFINTRODUCTIONOFMICROFLUIDICS Microfluidics has been an emerging field in the past two decades and there is still potentialforittobecomemoreprominentinthefutureasscientistsandengineers come up with new applications for microfluidics. The research interests toward microfluidicsmostlystemfromthefactthatthestudyofmicrofluidics,whichisthe studyoffluidmechanicsinthemicrokscale,canopenupmanyopportunitiesinthe fieldsofbiomedicalscienceandchemistry[1,2].duetothefactthatitismicroscale, flowsinmicrofluidicsaremostlystokesflow(flowoflowreynold snumber),which isrelativelyeasytostudyphysicallycomparedtohighreynold sflowandalsoeasy tomanipulate. BeingabletocontrolmicroKscalefluidhasmanybenefits.Firstly,tasksthatarenot dependentonvolumecanbeminiaturized.forexample,chemicalanalysesthatare common in drug diagnostics, biological analysis, and etc. do not depend on the volume of the sample but rather ratios of sample volume to reagents to produce results. By being able to control microkscale fluid, the samples can be reduced by 1000 times or more. Consequently, microfluidics can also prevent wastage. Apart frombeingabletoreducethevolumeofthesample,analysescanbecalibratedso that there won t be any excess of samples and reagents when conducting the analysis.

16 2 Thus, microfluidics could be a good tool for many biomedical and chemical applicationsasitreducessampleandreagentsizeandreducewastage. In recent years,a new optoelectric method named Rapid Electrokinetic Patterning (REP) had been developed by Williams et al[3k9]. REP is a particle manipulation technique that is used to differentiate and aggregate particles into clusters from a mixtureofparticlesofdifferentsizes.furthermore,repalsocantranslateparticle clusters from one location to another. This technique could possibly be used in manyapplicationslikebiosensing,nanostructurefabricationandetc.

17 3 CHAPTER2:RAPIDELECTROKINETICPATTERNING 2.1Introduction Inrecentyears,innovationsinmicrofluidicslikeLabKonKaKChip(LOC)havebeenan emergent technology that could have an impact in biomedical application[1, 2]. Simpletaskslikemanipulating,separating,sorting,andmixingarehighlydesirable inthebiomedicalcommunityasthesetaskswouldallowthemtobeabletoconduct laboratoryanalysiswithonlyasmallamountofsamplesandbeabletoconductiton aplatformthatcouldbethesizeofacreditcard[1,2]. REP is an electrokinetic method that can simultaneously accumulate a cluster of particlesofaspecificsizeandmanipulatethecluster[3k9].itrequiresaheatsource like a laser or heaters to provide a temperature gradient and also an AC electrical power source to provide an electric field gradient. There are many benefits to conductingrep.firstofall,repisanonkcontacthighkspeedmanipulationmethod where one can accumulate a cluster of particles in the matter of seconds. Furthermore, REP is a very consistent technique where clusters will consistently form when conducting it. Also, REP chips are very simple to produce and can be fabricatedinanylaboratorythathastherightmaterials.

18 4 (a) (b) (c) (d) (e) Figure1:RapidElectrokineticPatterning(REP).(a)Velocityprofileof electrothermalflow.(b)a3dtrajectoryoftheelectrothermalflowat200khz.(fig. 1bprintedwithpermissionfromKumaretal.Copyright2010Springer)[5].(cKe) ParticlesbeingtrappedintoaclusterusingREP. Ontheotherhand,REPalsohasitslimitations.REPisonlyconductedinasmallarea that is of micrometer length scales. LargeKscale conduction of REP will require higherpowerandamorecomplexsetup.furthermore,thetemperaturegradientor electric field gradient needed to conduct REP could be too unbearable for the samples and could permanently alter the samples structure [10]. However, by havingthisinmind,powerneededtoconductrepcanbecalibratedandreducedto eliminate damage to these samples. As of now, there has been no finding that

19 5 suggestrepcanpotentiallydamageakindofsampleandfurtherstudiesarebeing extendedtoinvestigatethepotentialdamageofrepthatcouldbeappliedontothe samples. 2.2Physics REP consists of three major components: electrothermal flow, particlekparticle interaction, and particlekelectrode interaction. Currently, particlekparticle interactionandparticlekelectrodeinteractioninrephasnotbeenwellstudiedand studies are being done to understand these phenomenon. On the other hand, electrothermalflowhasbeenverywellstudiedanditstheoreticalfoundationiswell established[11k16].

20 6 (a) (b) (c) (d) (e) Figure2:PhysicsofRapidElectrokineticPatterning.(a)Forcerelationshipona particleduringrep.(bkc)theclusteringofparticleswhenlaseristurnedon.(d) ThetemperatureincreseonanITOcoverslipand(e)theeffectofdifferentlaser powersandthetemperatureprofileproduced.(fig.2dand2ereprintedwith permissionfromkumaretal.copyright2010americanchemicalsociety)[6] $Electrothermal$Flow$ Electrothermal flow is a flow that is produced from an AC electric field and a temperature gradient [11K16]. Temperature gradient causes conductivity and permittivity gradients in the fluid which causes microvortices to form due to actuationbyanacelectricfieldonthefluid.electrothermalflowisusedinrepto concentrate particles onto the laser spot to increase the instances of particlek particleinteraction.

21 7 Equation(1)describesthetimeaveragedelectrothermalbodyforce[11K16]. (1) whereεsisthepermittivityofthesolution,eistheappliedacelectricfieldderived from Equation (1), ω is the angular frequency of the AC electric field. The electrothermal body force has two force components. The Coulomb force is the componentontheleftandthedielectricforceisthecomponentontheright.

22 $Particle4Electrode$Interaction$ (a) (b) (c) Figure3:PolarizationofParticles.(b)showsaneutralparticlewhenthereisno appliedelectricfieldand(c)showsapolarizedparticleduetoanelectricfield.

23 9 Aneutralparticleinanelectricfieldbecomespolarizedduetotheelectricfield.This iscalledaninduceddipole.dipoleswillbediscussedinsection2.2.4.thepositive charges would want to migrate to the part of the particle which is parallel to the direction of the electric field while the negative charges would want to migrate to the part of the particle which is opposite of the electric field. Furthermore, by following the same reasoning, the polarized particle would also attract opposite charged ions to its corresponding surface forming a layer. This layer is called the electricaldoublelayer.[17] Figure4:ElectricDoubleLayer.

24 10 Theelectricdoublelayercanbedividedintothreelayers,theinnerHelmholtzlayer (ihp), the outer Helmholtz layer(ohp) and the diffuse layer. The diffuse layer is a regionwheretheconcentrationofchargedionsoppositetothechargeofthesurface charge is higher than the bulk solution. The combined inner and outer Helmholtz layersarecalledthesternlayer.theinnerhelmholtzlayerconsistsofnonkhydrated ions that have specifically adsorbed to the surface and the outer Helmholtz layer consists of a layer of bound hydratedions(hydration not shown in Figure 4). Looking at Figure 4 as an example, the inner Helmholtz layer can consist of both positiveandnegativeionswhiletheouterhelmholtzlayerconsistsofonlypositive ions. This is true for a positive surface charge as well with the exception that positiveionsarereplacedbynegativeionsandviceversa.[17] 2.2.3$Dielectrophoresis$ The polarizability of a particle is governed by the permittivity and conductivity of theparticleandthesolution.thus,thepermittivityandconductivityoftheparticle and the solution has a property called the complex permittivity[17]. The complex permittivityis (2) where ε0 is the permittivity of free space, εr is the relative permittivity, i is the imaginarynumber,σistheelectricalconductivity,ωistheangularfrequencyofthe ACelectricfield.

25 11 The magnitude of the polarizability is frequency dependent. This dependence is describedbytheclausiuskmossottifactorparticle(p)andsolution(s)[17] (3) This decribes the relaxation in the polarizability of the particle with a relaxation timeof[17] (4) The inverse of the relaxation time is often referred to the MaxwellKWagner relaxationfrequency[17]. Dielectrophoresis is the force induced on polarized particles in an electric field. Furthermore,theforceonadipoleinanelectricfieldis[17] (5) WhereEistheelectricfield. Thus,thetimeKaverageddielectrophoreticforceis[17] (6) Where ε is the permittivity, a is the radius of the particle, subscriptp denotes the particlewhilesubscriptsthesolution,andeistheelectricfield.

26 12 It is important to understand that, apart from the electric field, the ClausiusK Mossotti factor is the other parameter that determines the direction of the dielectrophoretic force on the particle. A positive value in the ClausiusKMossotti factor will cause a positive dielectrophoretic force while a negative value in the ClausiusKMossottifactorwillcauseanegativedielectrophoreticforce.

27 13 (a) (b) Figure5:Positiveandnegativedielectrophoresisformorepolarizableparticleand lesspolarizableparticleinuniform(a)andnonkuniformelectricfield(b).

28 14 AscanbeobservedfromFigure5(a),thereisanetzeroforceindielectrophoresisin auniformelectricfield.thisisduetothefactthatthereisnospatialgradientofthe electric field in a uniform electric field. However, there is a net force towards the direction parallel or antikparallel to the electric field when there is a nonkuniform electricfieldascanbeseenfromfigure5(b).ifthedielectrophoreticforceisparallel to the electric field, it is called positive dielectrophoresis and it is called negative dielectrophoresisifitisantikparallel.ascanbeobservedfromthedielectrophoretic force, the ClausiusKMossoti factor plays a role in determining whether the dielectrophoreticforceisnegativeorpositive. FromFigure6,negativedielectrophoresisoccurswhentherealpartoftheClausiusK Mossotifactorisnegativeandpositivedielectrophoresisoccurswhentherealpart oftheclausiuskmossotifactorispositive.[17]

29 15 Figure6:PlotoftheClausiusKMossottifactorfortwodifferentparticles(solidand dottedlines)withrespecttofrequencyandtheseparationofparticlesusing Dielectrophoresis. Dielectrophoresis can be used to separate particles using positive and negative dielectrophoresis[17]. Taking figure 6 as an example, it can be seen that the two particles can be separated between the range whereone(green)isundergoing positivedielectrophoresiswhiletheotherisundergoingnegativedielectrophoresis (red)[17].

30 $Particle$Dipole4Dipole$Attraction$ Figure7:DipoleMoment. ElectricdipoleisformedwhentwochargesofsamemagnitudeQandoppositesigns are distributed into two separate locations[17]. Consequently, this forms a vector fromthenegativechargestothepositivechargescalledthedipolemoment,p$[17]. (7) wheredisthevectorofthedistancefromnegativetothepositivecharge. Furthermore,theforceonadipoleinanelectricfieldis[17] WhereEistheelectricfield. (8)

31 17 Figure8:AttractionoftwoFixedDipoles. Afixeddipoleiswhenamoleculehasadipolemomentinitsnaturalstate.Astwo fixeddipolesareclosetoeachother,thepositiveendofthedipoleisattractedtothe negativeendoftheotherdipoleandthiswouldcausetheparticlestoformachain [17].Polarizedparticlesalsodisplayasimilarbehaviorasthesedipoles. Ontheotherhand,aninduceddipoleiswhenadipolemomentisformedina moleculeorevenagroupofmoleculewiththeintroductionofanelectricfield. Objectslikepolystyreneparticlescanbeinductivelypolarizedunderanelectricfield andthisleadstoaninterestingphenomenonwhereparticlesareattractedor repulsedtoeachotherwhichisobservedinrep.

32 18 (a) (b) Figure9:ParticleKparticleInteractionwheretwoparticlesarealignedparalleltothe electricfield.(a)hasparticlesthataremoreconductivethanthesolutionand(b) hasparticlesthatarelessconductivethanthesolution. InFigure 9(a),bothparticlesare more polarizable than the solution. The electric fieldstrengthinsidetheparticlesandattheedgeoftheparticlesperpendicularto theelectricfieldarelowwhiletherearehighelectricfieldstrengthattheedgeofthe

33 19 particlesthatareparalleltotheelectricfieldespeciallybetweenthetwoparticles due to polarization. Both of these two particles experience positive dielectrophoresisandwillmovetowardshighestelectricfieldstrengthandforma chain. In Figure 9(b), the particles are both less polarizable than the solution. The electric field strength inside the particles and at the edges of the particles perpendiculartotheelectricfieldarehighwhiletherearelowelectricfieldstrength attheedgeoftheparticlesthatareparalleltotheelectricfieldespeciallybetween the two particles duetopolarization. Both of these two particles experience negativedielecterophoresisandwillmovetowardslowestelectricfieldstrengthand alsoformachain.[17] (a) (b) Figure10:ParticleKparticleInteractionwheretwoparticlesthatarealigned perpendiculartotheelectricfield. Next, the pair of particles is rotated 90 clockwise.infigure 10(a),theparticles experiencepositivedielectrophoresisandwillrepeleachotherasthesurrounding

34 20 areaoftheparticles not in between the two particleshasahigherelectricfield strength compared to the area in between the two particles. The same occurs in Figure 10(b) wheretheparticlesexperiencenegativedielectrophoresisand will repel each otherasthesurroundingareaoftheparticlesnotinbetweenthetwo particleshasalowerelectricfieldstrengthcomparedtotheareainbetweenthetwo particles.[17]

35 21 (a) (b) Figure11:ParticleKparticleInteractionwheretwodifferentparticlesarealigned parallel(a)andperpendicular(b)totheelectricfield.

36 22 These behaviors can also extend to particles of different polarizability. For Figure 11(a),themorepolarizableparticle(grey)andlesspolarizableparticle(white)have distributions of the electrical strengths exactly like Figure 9. The less polarizable particleexperiencespositivedielectrophoresisandwantstomovetowardshighest electric field strength while the more polarizable particle experiences negative dielectrophoresisandwantstomovetowardslowestelectricfieldstrength.asthe highest and lowest electric field strength of the particle is parallel to the electric fieldbutnotbetweenthetwoparticles,thetwoparticleswillrepeleachother.on theotherhand, Figure 11(b) has the distribution of electric field strength exactly likefigure 10. The less polarizable particle experiences positive dielectrophoresis and wants to move towards highest electric field strength while the more polarizable particle experiences negative dielectrophoresis and wants to move towards lowest electric field strength. As the highest and lowest electric field strength of the particle is perpendicular to the electric field and is in between the particles(more polarizable particle produce lowest electric field strength and less polarizable particle produce highest electric field strength), the two particles will attracteachother.[17]

37 23 Figure12:Chainformingofdifferentparticlesinasolutionappliedwithanelectric field. Thus, dielectrophoresis can be used to form chains of same particles or even a compositechainofdifferentparticlesjustlikefigure12.

38 $Parameters$for$Particle$Clustering$ Studies have been conducted to study the relationship between different parametersthatcontrolsrepandclustersize[7]. Figure13:Accumulationof1.0umparticlesattheITOelectrodeoveraperiodof3 minforvariouspeakktokpeakvoltages.laserpoweris23mwandacfrequencyis 70kHz.( IOPPublishing.ReproducedfromWilliamsetal.bypermissionofIOP Publishing)[7]. As can be seen from Figure 13, it can be seen that, in general, the number of particles captured increases linearly and then plateaus as the particle cluster reaches steady state. Furthermore, it can also be observed that higher applied voltagecauseslargernumberofparticlescapturedexceptfor15.3vpp.thiscouldbe due to an increased electrothermal body force that is larger then the particlek particleandparticlekelectrodeforceswhichwouldcausetheparticlestobeforced outofthecluster.

39 25 Figure14:Accumulationof1.0umparticlesattheITOelectrodeoveraperiodof3 minforvariousfrequencies.laserpoweris23mwandacpeakktokpeakvoltageis 15.3Vpp.( IOPPublishing.ReproducedfromWilliamsetal.bypermissionofIOP Publishing)[7]. From Figure 14, it can be observed that the particles captured for various frequencies starts out increasing linearly and then plateaus as the cluster reaches steadystate.furthermore,theparticlescapturedincreases,reachesapeakandthen decreasesasfrequencyincreases.thus,itseemsthatthereisanoptimumfrequency thatproducesthelargestparticleclustersize.

40 26 Figure15:Accumulationof1.0umparticlesattheITOelectrodeoveraperiodof3 minforvariouslaserpower.acpeakktokpeakvoltageis15.3vpp.andacfrequencty is70khz.( IOPPublishing.ReproducedfromWilliamsetal.bypermissionofIOP Publishing)[7]. As with Figure 13 and Figure 14, Figure 15 shows that particle clusters increases linearly linear and then plateaus as the particle cluster reaches steady state. The increase in laser power increases the number of particles captured. This could be duetotheelectrothermalflow svelocityincreasesaslaserpowerincreasesduetoa steepertemperaturegradientwhichleadstofasterparticlecirculationandahigher probabilitytohaveaparticleattractedtothecluster.

41 27 Figure16:Maximumtrappingfrequencyforparticlesofvariousdiameterstocluster atdifferentappliedacpeakktokpeakvoltage.laserpoweris23mw.( IOP Publishing.ReproducedfromWilliamsetal.bypermissionofIOPPublishing)[7]. Figure 16 shows the maximum trapping frequency for particles of different diametersbeforeitfailstoformacluster.aparticleofsmallerdiameterwillremain trappedatahigherfrequencywhencomparedtoalargerparticle.furthermore,the maximumtrappingfrequencyofthesameparticleincreasesastheappliedacpeakk tokpeakvoltageincreases. 2.3Methodology Asstatedbefore,fabricationofREPchipsarerelativelysimple.Thechipismadeofa thinfilmelectrodecoatedontoaglasssubstrate.thematerialthatistypicallyused toconductrepisitoasitistransparentandwouldallowforobservationinsidethe chip.

42 28 (a) (b) Figure17:REPChip. With two of these substrates with electrodes facing each other, a spacer that was precuttodefinethemicrochannelsisplacedinbetweenthemtoformthesidewalls. The spacer is any material that can be attached to both electrodes and is nonk reactive to the electrodes and sample. In some cases, a commercial doubleksided tape like Scotch Removable Poster Tape, ¾ x 150 (3M Company, St. Paul, MN) can be used to conduct REP for short amount of time. The typical setup for conductingreprequiresalasersourceandanacfunctiongenerator.thelaserwill beappliedtothedesiredlocationtoconductrep. AtechnicalrequirementforthelaserinREPisthatthelaserhastoprovidepower that range between 0K20 mw. Furthermore, the laser frequency has to be consideredsothattheheatabsorptionofthelaserontothesampleisminimal.ac powersupplyneedstoprovide0k10vwithfrequencyof30k200khz.itisadvised that an oscilloscope is connected to the AC power supply to monitor the power output.[3k9]

43 29 Furthermore,lensesaretypicallychosentobehigherthan10X.Thiswouldensure thatthelaserspotsizeissmalltocreateasteepertemperaturegradient[3k9]. Figure18:REPSetup. The laser and camera is connected to an inverted Nikon TE2000U microscope. An ACfunctiongeneratorisconnectedtotheelecrtricalcontactsoftheREPchip.Asthe laser travels into the microscope, the laser is reflected by a filter cube into the objectivelensandisfocusedontothechip selectrodes.thescatteredlightfromthe particlesintherepchiptravelsbackintotheobjectivelens,throughthefiltercube andintothecamera,whichisbeingcontrolledbyacomputer. Particlesaresuspendedinlowconductivitysolutions(σ 10mS/m).Electricdouble layer becomes thinner which increases the particlekelectrode adhesion when particlesaresubmergedinhighconductivitysolutions[3k9].

44 30 2.4CurrentApplications 2.4.1$Differentiation$of$Biological$Organisms$ In2012,Kwonetal.demonstratedthatREPcouldtrappingofbacteriaShewanella oneidensisandalsocollectionofs.aureusfromas.aureusands.cerevisaemixed solution.thisshowspotentialapplicationinthebiologicalfield[18]. Figure19:TrappingandTranslationofS.oneidensis.(Reprintedwithpermission fromkwonetal.copyright2012royalsocietyofchemistry)[18]. ThemajoraxisofShewanellaoneidensis,a1µmrodKshapedbacteria,wasaligned with the electric field in the REP chip when the AC electric field was turned on. Whenthelaserwasturnedonandfocusedonanelectrode,thebacteriaaccumulates into a sizable cluster at the laser spot. It was also demonstrated that the cluster couldbetranslatedwhenthelasermovedtoanotherlocation.[18]

45 31 Figure20:SeparationofS.aureusfromS.cerevisae.(Reprintedwithpermission fromkwonetal.copyright2012royalsocietyofchemistry)[18]. Furthermore, Kwon et al. also demonstrated trapping and segregating organisms within an REP trap by changing the frequency of the AC electric field. Saccharomyces cerevisae, a 5 µm spherical shaped fungus, and Staphylococcus aureus,a~1 µm spherical shaped bacteria, were both trapped under a 17.5 khz frequency, Vpp and 20 mw laser power. However, when the frequency increased to almost twice of the original frequency, the S. cerevisae were repelled awayfromthetrapwhilethes.aureusremainedinthetrap.[18] Overall, REP had no difficulty manipulating, trapping, and segregating biological organismsandthiscouldopenthedoorspotentialapplicationsinthebiologicalfield especiallyincellculturingandbioksensing.

46 $Plasmonic$Nanostructure$to$Optimize$REP$ Figure21:Plasmonicenanostructures:Thevelocitydistribution(a)andradial velocity(b).(reprintedwithpermissionfromndukaifeetal.copyright2014 AmericanChemicalSociety)[19]. TherehavebeenmanyalternatedesignsforREPtoreducethecostofsettingupa REP setup. One of the alternative designs for REP is to fabricate plasmonic nanostructures to increase the heating effects in a traditional REP chip. The nanostructures, array of gold nanodisks with diameters of 240 nm, that was fabricatedwastunedtoabsorblightoptimallyataspecificwavelength.byexciting the structure at resonance frequency, a temperature gradient is produced and because the temperature gradient is significantly higher than a typical REP chip design, the electrothermal vortex also has a higher flow velocity. This higher flow velocitywouldinturnincreasethechanceofcapturingparticlesintothereptrap. [19]

47 $REP$in$Coplanar$Electrodes$ (a) (b) Figure22:TwinOpposingMicrovortices(TOMV).(Reprintedwithpermissionfrom Parketal.Copyright2012RoyalSocietyofChemistry)[20]. Park et al. investigated twin opposing microvortices (TOMV). Two coplanar ITO electrodes were fabricated onto the glass surface with SUK8 partly covered. An AC electric field was applied while introducing a laser that is focused on the glass surfacewiththeitoelectrode.theorientationofthetomvcanbechangedwithin the chip when the laser is being focused at a different location which leads to a tunablemicrovorticesdevicethatcanhavedifferentandinterestingmixingeffects whenalaserisbeingfocusedatdifferentlocations.[20]

48 $Thin$Film$Heaters$as$Heat$Source$for$REP$ (a) (b) (c) Figure23:ConductingREPusingthinfilmheatersasheatsource.(Reprintedwith permissionfromvelascoetal.copyright2012elsevier)[21,22]. Williams et al and Velasco et al. looked into thin film heaters as an alternative for laserinducedelectrothermalflow.a10µmwidewiretracewasusedastheheating element which could produce 0.5 Kelvin temperature gradient. The system could trap, pattern, and sort particle cluster. This allows for a more affordablesetupto conductrep[21,22]

49 35 2.5ConclusionsandFutureWork REP is desirable compared to other optoelectrokinetic techniques as it does not requireanyfabricationofelectrodegeometry,itiseasytofabricateanditisanonk contact technique that is consistent. Many applications are currently being looked into for REP and REP has the potential for many novel applications in the future.

50 36 CHAPTER3:OPTIMIZINGELECTRODETHICKNESSANDMATERIALFORLASERK INDUCEDELECTROTHERMALFLOW 3.1Introduction RapidElectrokineticPatterning(REP)isanoptoelectrictechniquefornonKinvasive particle manipulation designed by Williams et al. [3K9]. By utilizing the electrothermal vortex, particlekparticle interaction and particlekelectrode interaction,repisabletocapture,translateanddifferentiateparticlesofdifferent sizes. Figure24:IllustrationofREPwithlaserfocusedatthetopelectrode. REP has many potential applications in fabrication and biosensing. It was demonstratedbywilliamsetal.thatrepwasabletomanipulateanddifferentiate particlesofdifferentsizesanddifferentmaterials[3k9].kwonetal.haveusedrep

51 37 to differentiate bacteria from fungus and also demonstrated the translation of the bacteria to another location [18]. There are many advantagestoconductingrep compared with other techniques that require fabrication of electrode geometries. Firstly, REP is a nonkcontact technique where only a heating source and an AC electricfieldsourcearerequiredtoproduceit.secondly,repdoesnotrequireany prior electrode fabrication to manipulate particles. By translating the laser beam, particleclusterscanbeeasilytranslatedtoanotherlocation.inthecurrentsetup, REPrequires4mW/µm 2 totrapparticles.thegoalofthisstudyistoinvestigatethe possibility of reducing the laser intensity even further to minimize the damage to biologicalsamplesandlowerthecostofthelaserbeingutilized[10]. TheelectrodematerialiscrucialforREPaselectrodesarenotonlyusedtoapplyAC electric fieldbuttheyalso absorb the laser that is being focused onto the chip to produce temperature gradients that drive electrothermal vortex. In this study, the material dependence in thermal effects of electrothermal vortex will be accessed and analyzed to determine its role in REP. We have organized this paper in 3 key sections. First we present the theoretical model of laser induced heating of a thin layer on a glass substrate. This simple model provides us insight about the key parameters controlling the laser induced heating. It also shows that Ti and Ni can produce considerably higher temperature gradient than ITO. We then simulate electrothermalflowincomsolmultiphysicsv4.4 tostudylaserinducedheating and electrothermal flow in all its complexity. We mention the theory that governs

52 38 electrothermalflowfollowedbyabriefexplanationofthecomputationalmodel.we thenpresentexperimentalresultsofthevelocityprofileoftheelectrothermalflow fordifferentmaterials.experimentswereperformedtomeasuretheelectrothermal flow velocity profile using MicroKParticle Image Velocimetry (µpiv) for validating thecomputationalmodel.resultsrevealthesametrendasinouranalyticalmodel wheretiandniappeartobefarbettermaterialchoices. 3.2Theory 3.2.1$Reflection$of$the$Laser$at$the$Electrode$Surface$ The reflectance of the laser, R is the fraction of the light reflected to the incident light.byknowingthereflectancer,thelightabsorbedcanbecomputedasitisjust theremainingfractionoflightthatdidn tgetreflectedaway14r. Figure25:PropagationofLight(ReflectionandRefraction)inMaterials.

53 39 ThereflectancecanbecalculatedusingFresnel sequation. (9) where,,,nisthecomplexindexofrefraction, nr is the index of refraction and k is the extinction coefficient, λ is the light swavelength,disthematerialthickness,andφ$istheangleofthelightonthe plane.thesubscriptnumberscorrespondstothematerialsinfigure $Production$of$Temperature$Gradient$in$Electrodes$ The heat equation governs the heat exchange at the surface boundaries of the boundary volume and also the heat generated or dissipated within the boundary volumeinthecourseoftime. (10) wherekisthethermalconductivityofthematerial,gisthepowerdensityorheat generation by the laser, ρ is the density of the material, cpisthespecificheat capacityandtistime. Theheatgenerationinthematerialbythelaserisdefinedas (11) where Risthereflectance, I0isthe intensity at the center of the beam,αisthe absorptioncoefficientoftheelectrode,zisthedistancetravelledbythelaserinthe

54 40 material,wisthebeamwaistofthelaserandristheradialdistancefromthecenter of the beam. The general heat equation is solved to determine the temperature gradientproducedinthechip $Analytical$Solution$of$Laser$Heating$onto$a$Thin$Substrate$ SolvingthegeneralheatequationusingGreen sfunction,thederived analytical solutionofthecwlaserheatingontothethinelectrodeinanrepchip[23k25]is (12) whereais, Pisthelaserpower,k1isthethermalconductivityoftheelectrode,k2isthethermal conductivity of the solution, z1 is the thickness of the electrode,α1 is the optical absorptioncoefficientoftheelectrodeandα2istheopticalabsorptioncoefficientof theglass. From this equation, it can be observed that there are a few important parameters that determine the temperature profile of the electrode. These parametersarethethicknessoftheelectrode,reflectanceoftheelectrode,thermal conductivity of the electrode material and solution, the optical absorption coefficientoftheelectrodeandsolutionandthebeamwaistofthelaser.

55 $AC$Electric$Field$in$Solution$ Equation(13)isderivedfromthechargeconservationequation[15,16]. E = 1 σ s σ s T T E (13) whereeistheappliedacelectricfield,σsistheconductivityofthesolutionandtis the temperature of the solution. This equation describes the AC electric field producedinthemicrochannelwhenconductingrep $Fluid$Mechanics$and$Electrothermal$Flow$ TheNavierKStokesequationgovernsthefluidexchangeattheboundarysurfacesat theboundaryvolumeinthecourseoftime. ElectrothermalflowistheflowproducedbyhavingacombinationofanACelectric field and a heat source that produces a temperature gradient. Electrothermal flow can be solved using the NavierKStokes equation together with an electrothermal bodyforce[11k17]: (14) whereuistheflowvelocity,tistime,fisthegravitationalbodyforce, fet isthe timekaveragedelectrothermalbodyforce,µisthekinematicviscosity.

56 42 Theelectrothermalbodyforceisgivenbythefollowingequation (15) whereεsisthepermittivityofthesolution,eistheappliedacelectricfieldderived from Equation (15), ω istheangularfrequency of the AC electric field. In the electrothermalbodyforce,thefirsttermoftheequationiscalledthecoulombforce while the second term of the equation is called the dielectric force. Take note that thecoulombforceandthedielectricforcesubtracteachother.inthecurrentstudy, Coulombforcedominatesthedielectricforce. 3.3MethodandMaterials 3.3.1$Chip$Design$and$Fabrication$ Experiments are conducted in a microchannel with a height of 280 µm. An electrically insulated double sided tape is trimmed to form microchannel in the microfluidic chip. This spacer was sandwiched between two electrodes that has suitable thermal and optical properties. A thin glass cover slip(thermo Scientific, MA,USA)wasattachedtothesideofthemicrochanneltobeusedasawindowfor viewingintothemicrochannelfromtheside,asshowninfigure26.700nmand70 nm (Indium Tin Oxide) ITO layers (SPI supplies Inc., PA, USA) were used as electrodesfortherepchip.otherthanthat,60nmand80nmnickellayers(drli Inc.,MO,USA)and40nmand50nmTitaniumlayer(DRLIInc.,MO,USA)werealso usedaselectrodesfortherepchip.

57 43 Twoholesweredrilledintothetopelectrodetobethechannel sinputandoutput. Afterthat,theelectrodesandcoverglassweresolventcleanedbyultrasonicationin acetone, isopropanol for three minutes each. The electrodes and glass cover slip werelaterwashed with distilled water and blowkdriedwithnitrogengas.lastly, copper tapes were adhered to the electrodes to act as electrical contacts. The bottomelectrodewasalwaysthe70nmitoelectrodeaslaserwasalwaysfocused at the top electrode and it was preferable that the bottom electrode absorbs minimalradiationfromthelaser $Experimental$Setup$ A Telulex Model SGK100/A signal generator (Berkeley Nucleonics Corp., CA, USA) was connected to the chip and the Agilent 54610B oscilloscope was used(agilent Technologies, CA, USA) to observe the AC electrical signal. A 1064 nm continuous wave Nd:YVO4 laser was equipped onto theinverted Nikon TE2000U microscope. LaserwasfocusedatthetopoftheelectrodewithaNikon20x(0.45NA)objective lens. Furthermore, a Nikon boomkstand microscope was installed to view the microfluidicchipfromthesideusinganikon10x(0.25na)objectivelens.acamera (PCO.1600,CookeCorporation,MI,USA)andafluorescentilluminationlamp(Xcite series 120PC, EXFO Life Sciences & Industrial Division, Ontario, Canada) were installedontotheboommicroscopeforimagerecordingandillumination.fornickel andtitaniumrepchips,aphantomv7.3highspeedcamera(visionresearchinc., NJ,USA)wasusedduetothehigherflowvelocityproducedfromthetwoelectrodes.

58 44 PCOCamerawascontrolledusingthemanufacturerprovidedCamware software. Phantom High Speed Camera was controlled by the manufacturer provided software,pccversionv $Particle$Preparation$ 3.2 µm FluoroKMax TM Red Fluorescent Polymer microsphere Particles (Thermo Scientific, MA, USA) were used as tracers for Particle Image Velocimetry (PIV) analysis. The particles are centrifuged from the native solution and dispensed in a KClsolutionhaving8.8mS/mconductivity $Software$and$Image$Processing$ Image processing to modify the contrast of the recorded image is conducted in MATLAB (MathWorks Inc., MA, USA) and Particle Image Velocimetry (PIV) evaluationisperformedusingdavis7.2(lavisioninc.,mi,usa).

59 45 Figure26:ExperimentalSetupandMethod. AscanbeseeninFigure26,thelaser,whichisincidentfromthebottomofthechip, isfocusedatthetopoftheelectrodewhileanobjectivelensconnectedtoaboomk stand microscope is situated on the side to record the electrothermal flow. After that, PIV is conducted using Davis 7.2. In all the experiments unless otherwise specified, an applied peakktokpeak voltage of 16 V and AC frequency of 100 khz wereused.laserpowerwasmaintainedat41mw. 3.4NumericalSimulations Simulations are conducted in COMSOL Multiphysicsv4.4. Apart from equations introduced in the theory section, there are some additional equations to be consideredtosimulatetheelectrothermalflow.

60 $Production$of$Temperature$Gradient$in$Electrodes$ Temperature gradient in REP is produced by an external heating source, which in thiscaseisalaser.thelaserismodeledasagaussianbeamandincludedasaheat generationterminthesteadykstateheatequation[26]. (16) wherekisthethermalconductivity,gisthepowerdensityorheatgenerationbythe laser. Theheatgenerationinthematerialbythelaserisdefinedas (17) where Risthereflectance, I0isthe intensity at the center of the beam,αisthe absorptioncoefficientoftheelectrode,zisthedistancetravelledbythelaserinthe material,wisthebeamwaistofthelaserandristheradialdistancefromthecenter of the beam. The general heat equation is solved to determine the temperature gradientproducedinthechip.

61 $AC$Electric$Field$in$solution$ Equation(18)isderivedfromthechargeconservationequation[15,16]. E = 1 σ s σ s T T E (18) whereeistheappliedacelectricfield,σsistheconductivityofthesolutionandtis the temperature of the solution. This equation describes the AC electric field producedinthemicrochannelwhenconductingrep $Electrothermal$Flow$ ElectrothermalflowistheflowproducedbyhavingacombinationofanACelectric field and a heat source that produces a temperature gradient. Electrothermal flow can be solved using the NavierKStokes equation together with an electrothermal bodyforce[11k17]: (19) whereuistheflowvelocity,tistime,fisthegravitationalbodyforce, fet isthe timekaveragedelectrothermalbodyforce,µisthekinematicviscosity. Theelectrothermalbodyforceisgivenbythefollowingequation (20) whereεsisthepermittivityofthesolution,eistheappliedacelectricfieldderived from Equation (18), ω istheangularfrequency of the AC electric field. In the electrothermalbodyforce,thefirsttermoftheequationiscalledthecoulombforce

62 48 while the second term of the equation is called the dielectric force. Take note that thecoulombforceandthedielectricforcesubtracteachother.inthecurrentstudy, Coulombforcedominatesthedielectricforce $Simulation$ (a) (b) (c) Figure27:(a)Boundaryconditionsofthegeometryinthesimulation.E:ACElectric FieldModel.H:HeatTransferModel.F:ComputationalFluidDynamicsModel.(b) Vectorfieldofsimulatedelectrothermalflowinlogarithmicscale.(c)Contourplotof simulatedelectrothermalflow.

63 49 A 2KD axisymmetric model is used to conduct the simulation. The NonKIsothermal Flowpackagewasutilizedtosimulatethetemperaturegradientandfluidflow.The coefficientformpdemodulesimulatestheacelectricfieldproducedinthefluidas theelectricfieldisgovernedbythetemperaturegradient. Table1:Valuesofimportantpropertiesfornumericalsimulation. Symbol Property Value P LaserPower 41mW w BeamWaist 2.4µm f ACFrequency 100kHz alpha beta FractionalDifferenceofPermittivitywithrespectto Temperature FractionalDifferenceofConductivitywithrespectto Temperature K0.0041/K 0.021/K er Relativiepermittivity 80F/m sigma Conductivity 8.8mS/m hw ChannelHeight 280µm w ChannelLength 10000µm hg ThicknessofGlass 700µm hito ThicknessofBottomElectrode 60nm In Table 1, the important properties for the numerical simulation are recorded. These properties will be use to compute and formulate the partial differential equationstobenumericallysimulatedinthesoftware.

64 50 Figure28:FinalGeometry. Thegeometryisconstructedusingfiverectanglesstackoneachothertoformapair of glass substrates, a pair of electrodes and the microfluidic channel. These five rectangles are formed in union to each other to transform the geometries into the finalgeometry. The heat generation of the laser is defined as an analytical function to be implementedintheheatequationinthesimulationmodel.materialsaredefinedin thematerialssectionusingthematerialslibrarythatisprovidedinthesoftware. Two simulation modules, NonKIsothermal Flow and Coefficient Form PDE, are utilizedforthesimulationoftheelectrothermalflow.themodelisa2dsymmetrical modelincylindricalcoordinates. TheNonKIsothermalFlowmoduleisthepairingoftheheatmodeltogetherwiththe ComputationalFluidDynamicsmodel.Thisiscrucialforthesimulationoftheheated fluid from the electrodes. The electrodes absorb the laser and the heat absorbed

65 51 fromthelaserisdissipatedintotheliquid.thissimulationtogetherwiththeelectric fieldsimulationisusedtosimulatetheelectrothermalflow. TheelectricfieldismodeledusingtheCoefficientFormPDE.COMSOLMultiphysics v4.4 does have a electrical module but because the electric field produced has a temperature dependence and it was decided that a custom PDE module should be constructedtomodeltheelectricfield. (21) wherethepdeissetuptomatchtheelectricalfieldequationinequation(18). Meshgeometryisacrucialcomponentincomputationalmodeling.Thepropertiesof themeshgeometrylikeelementdensity,shapeoftheelement,sizeoftheelement and etc. determines the stability of the computation and also the accuracy of the results of the computed numerical model. Thus, it is important to ensure that the meshgeometryisasoptimumaspossible.ideally,themeshgeometryshouldbeas fine as possible. However, the computational limit of the computer has to be consideredaswellwhenconstructingthemeshgeometrytoensurethatthemesh construction and the computation in each individual element does not exceed the computationallimit.

66 52 Figure29:Freetriangularmeshing(topandbottom)andquadrilateralmeshing (center). Themeshconstructedonthethinfilmsismappedwithaquadrilateralmeshwhere locationsofindividualnodesarecontrolledtooptimizemeshqualitytotakeaccount for extremely high aspect ratio of thegeometry as it is easy to customize. This is especially important to consider in this simulation as the simulation spans over a large range of length scales from electric field in channel ( in µm range) to temperaturegradientinelectrodethickness(innmrange).thisisthereasonforthe customizationofthedistributionoftheelementshapeandsize.otherpartsofthe meshes are constructed using free triangular meshing. The quality of mesh was ensuredsothatthecomputationalresultsachievemeshindependence.