Magnetic Field-Assisted Finishing of Micropore X-Ray Optics

MISSING IMAGE

Material Information

Title:
Magnetic Field-Assisted Finishing of Micropore X-Ray Optics
Physical Description:
1 online resource (171 p.)
Language:
english
Creator:
Riveros, Raul
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:
Mechanical Engineering, Mechanical and Aerospace Engineering
Committee Chair:
Greenslet, Hitomi
Committee Members:
Taylor, Curtis
Schueller, John K
Schmitz, Tony L
Saab, Tarek

Subjects

Subjects / Keywords:
abrasive -- assisted -- ferrofluid -- field -- finishing -- magnetic -- manufacturing -- nontraditional -- optics -- polishing -- xray
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre:
Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Reducing the size and weight of spaceborne X-ray telescopes minimizes mission costs. In 2006, Ezoe et al. first introduced micropore X-ray optics made from thin silicon wafers called ``MEMS micropore X-ray optics.'' These optics are lightweight (1000 cm^2 effective area per 1 kg) have curvilinear micropores etched by deep reactive ion etching. This process, however, yields reflecting surfaces with an unacceptably high surface roughness of ~30 nm Rq. For efficient X-ray reflection and focusing, a microroughness of 3 nm Rq or better is needed. A suitable finishing process employing alternating magnetic field assisted finishing was developed. This process involves a reciprocating mixture of fine abrasive particles and magnetic fluid actuated by an externally applied alternating and switching magnetic field. The optic is submerged in the fluid mixture which then flows through the optic's micropores, finishing the sidewalls (reflecting surfaces). A finishing machine was constructed to realize this principle on 7.5 by 7.5 mm miniature workpieces called ``mirror chips,'' and another machine was constructed to finish full-sized MEMS micropore X-ray optics.  Investigations into the finishing characteristics were carried out on mirror chips. These studies demonstrated the finishing process' ability to remove material and cause reductions in the micropore sidewall surface roughness. Additionally, improvements in the mirror chips' X-ray reflectance were observed. The finishing process' surface modification mechanism was also studied through the execution of various finishing experiments on flat workpieces. Results from these experiments showed changes in surface texture which suggest that abrasive particles modify the micropore sidewalls by repeated indentation. A possible surface modification mechanism is proposed.
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 Raul Riveros.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Greenslet, Hitomi.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044485:00001


This item is only available as the following downloads:


Full Text

PAGE 1

MAGNETICFIELD-ASSISTEDFINISHINGOFMICROPOREX-RAYOPTICSByRALEDUARDORIVEROSADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2012

PAGE 2

2012RalEduardoRiveros 2

PAGE 3

Idedicatethistomyfamily. 3

PAGE 4

ACKNOWLEDGMENTS Ithankmyfamilyforeverything.Ialsothankmyadvisor,Dr.HitomiYamaguchiGreenslet,whohasgoneaboveandbeyondthecallofdutytoensuremysuccess.Ialsothankmycommitteemembers:Dr.TonyL.Schmitz,Dr.JohnK.Schueller,Dr.CurtisR.Taylor,andDr.TarekSaab.IalsowishtothankourJapanesecollaborators:Dr.YuichiroEzoe,Dr.IkuyukiMitsuishi,TeppeiMoriyama,KensukeIshizu,UtakoTakagi,TomohiroOgawa,andDr.KazuhisaMitsuda.IalsothankallmembersoftheMTRCsinceJuly2006.IthanktheNationalScienceFoundationfortheprojectwhichfundedmydoctoralresearch.IthanktheInternationalSocietyforOpticsandPhotonicsforgrantingmeascholarship.IthanktheAmericanSocietyofMechanicalEngineersandtheUniversityofFloridagraduateschoolformultiplestudenttravelgrants.SpecialthanksgotoDr.JohnGreensletwhoverykindlyeditedmywrittenwork. 4

PAGE 5

TABLEOFCONTENTS page ACKNOWLEDGMENTS .................................. 4 LISTOFTABLES ...................................... 7 LISTOFFIGURES ..................................... 8 ABSTRACT ......................................... 13 CHAPTER 1INTRODUCTION ................................... 15 1.1X-rayAstronomy ................................ 15 1.2X-rayOptics ................................... 17 1.3MEMSMicroporeX-rayOptics ........................ 25 1.3.1Background ............................... 25 1.3.2OpticalDesign ............................. 27 1.3.3ManufacturingMethods ........................ 30 1.3.3.1Deepreactiveionetching .................. 30 1.3.3.2Hydrogenannealing ..................... 33 1.3.3.3Plasticdeformationofsiliconwafers ............ 35 1.3.3.4X-rayLIGA .......................... 36 1.4MagneticField-assistedFinishing ....................... 38 1.4.1Background,Fundamentals,andApplications ............ 38 1.4.2FinishingofOpticalandSmallWorkpieces .............. 44 1.5MEMSMicroporeMirrorChips ........................ 46 2DEVELOPMENTOFAFINISHINGPROCESSFORMEMSMICROPOREOPTICS ........................................ 50 2.1ProcessingPrinciple .............................. 50 2.2FinishingEquipmentforMirrorChips ..................... 55 2.2.1MotionofMagneticAbrasiveFluidMixture .............. 57 2.2.2FinishingEquipmentCharacterization ................ 59 2.3FinishingEquipmentforFourInchOptics .................. 62 3FINISHINGCHARACTERISTICS .......................... 71 3.1EvaluationMethodsforMicroporeOptics .................. 71 3.1.1ObservationofX-rayReectionCharacteristics ........... 71 3.1.2SurfaceProleCharacterization .................... 73 3.1.2.1Surfaceprolingmethods .................. 73 3.1.2.2Surfaceproleanalysis ................... 76 3.1.3OtherCharacterizationMethods .................... 78 3.2LIGA-fabricatedNickelMirrorChips ..................... 78 5

PAGE 6

3.2.1PreliminaryStudy ............................ 78 3.2.2FinishingCharacteristics ........................ 80 3.3DRIE-fabricatedSiliconMirrorChips ..................... 86 3.3.1PreliminaryStudy ............................ 86 3.3.2FinishingCharacteristics ........................ 88 4SURFACEMODIFICATIONMECHANISM ..................... 98 4.1ApproachtoClarifyingtheSurfaceModicationMechanism ........ 98 4.2ObservationofSurfaceModicationsonFlatWorkpieces ......... 99 4.2.1ExperimentalEquipment ........................ 99 4.2.2FinishingofaSputteredMetalSurface ................ 101 4.2.3FinishingofaPatternedSurface ................... 102 4.3ObservationofSurfaceModicationsandFluidFlowthroughaSingleMicropore .................................... 112 4.3.1ExperimentalSetup .......................... 112 4.3.2EffectsofAbrasiveSizeonSurfaceRoughness ........... 117 4.3.3High-ResolutionImagingofSurfaceModications .......... 125 4.3.4ObservationofMAFMMotion ..................... 131 4.3.4.1Microporeowobservationsetup ............. 131 4.3.4.2Visualizationofowdrivenbyanalternatingmagneticeld .............................. 133 4.3.4.3Visualizationofowdrivenbyswitchingdirectcurrent .. 140 4.4DeliberationsontheSurfaceModicationMechanism ........... 143 4.4.1NanoscaleStructuresinMagneticFluids ............... 143 4.4.2PossibleInteractionbetweentheMagneticNanoparticleStructureandAbrasiveParticles ......................... 145 4.4.3IndentationasaSurfaceModicationMechanism .......... 148 5CONCLUSIONS ................................... 151 5.1ConcludingStatements ............................ 151 5.2FutureWork ................................... 153 APPENDIX:MAFMFLOWUNDERANALTERNATINGMAGNETICFIELD ..... 154 REFERENCES ....................................... 161 BIOGRAPHICALSKETCH ................................ 171 6

PAGE 7

LISTOFTABLES Table page 1-1SelectedcosmicstructuresandassociatedX-raygeneratingmechanisms ... 17 1-2Comparisonofselectmirrortechnologies ..................... 26 3-1Preliminarynickelmirrorchipexperimentalconditions .............. 79 3-2LIGA-fabricatedmirrorchipexperimentalconditions ............... 82 3-3PreliminaryDRIE-fabricatedmirrorchipexperimentalconditions ........ 88 3-4AnnealedDRIE-fabricatedmirrorchipgeometryandconditions ......... 89 3-5AnnealedDRIE-fabricatedmirrorchipnishingconditions ............ 91 4-1Sputteredworkpiecenishingexperimentalconditions .............. 101 4-2Flatworkpieceexperimentalconditions ...................... 103 4-3Finishingconditionswithmagneticuidonly .................... 119 4-4Brassworkpiecenishingconditions ........................ 122 4-5Smoothbrassworkpiecenishingconditions ................... 127 4-6Alternatingcurrentowvisualizationconditions .................. 134 4-7Switchingdirectcurrentowvisualizationconditions ............... 141 7

PAGE 8

LISTOFFIGURES Figure page 1-1Schematicdepictingtheanglesofreectedandrefractedraysrelativetotheincidentrayangle ................................... 20 1-2Schematicofaparaxialrayrefractingthroughathinsymmetricallens ..... 20 1-3SchematicofaWoltertypeIgrazingincidencetelescopeconguration ..... 22 1-4Schematicofacut-awayviewofaWoltertypeInestedmirrorarrangement .. 24 1-5Schematicofatwo-reectionMEMSmicroporeX-rayoptic ........... 28 1-6Angularresolutionlimitsforarangeofincidentenergiesandmicroporewidths 29 1-7Schematicofamicroporeetchinglithographyprocess,anICPDRIEchamber,andtheBoschprocess .......................... 32 1-8PhotographofaMEMSmicroporeX-rayopticmadebyDRIE .......... 34 1-9SchematicdepictionoftwomethodsofLIGAfabrication ............. 37 1-10PhotographofaLIGA-fabricatedPMMAmoldandnickelmirrorchip ...... 38 1-11Schematicofapermanentmagnetcontrollingamagnetictoolthroughanonmagneticworkpiece ............................... 39 1-12Photographofironparticlesinachain-likestructurealongthelinesofmagneticuxsurroundingapermanentmagnet .................. 40 1-13Schematicdepictingtheself-recongurationofaexiblebrush ......... 41 1-14ClassicationofmagneticandabrasivematerialsinrelationtoMAF ...... 43 1-15ClassicationofmagneticeldgeneratorsinMAF ................ 44 1-16Schematicofthegeometryofamirrorchip .................... 46 1-17PhotographofaDRIE-fabricatedMEMSmicroporeX-rayopticnexttoamirrorchipandaclose-upoftheamirrorchip ................... 47 1-18OpticalmicrographsofmicroporesidewallsfabricatedbyLIGAandDRIE ... 48 2-1Photographsof200Lofwater-basedandoil-basedmagneticuidinnon-magnetizedandmagnetizedstates ...................... 52 2-2Photographsof200Lofabrasiveslurry,water-basedalkalinecolloidalsilica,MAFM,andmagnetizedMAFM ........................... 54 2-3Schematicofthemirrorchipnishingsetupandtheprocessingprinciple .... 55 8

PAGE 9

2-4Photographofthemirrorchipnishingmachine .................. 56 2-5Schematicofthepoletipgeometry ......................... 56 2-6Three-dimensionalmodelofthemirrorchipnishingmachinewithmagneticyokecomponentsdelineatedinred ......................... 57 2-7Multipleviewsofthemirrorchipholder ....................... 58 2-8Aschematicoftheelectriccircuitandthesimulationresults ........... 59 2-9Fluidbehaviorwithoutdiodesandwithdiodes ................... 60 2-10Poletiptemperatureandmagneticuxdensitymeasurementsetup ...... 60 2-11Changesinpoletiptemperatureandcurrentwithtime .............. 61 2-12Changeinmagneticuxdensitywithtime ..................... 62 2-13Fourinchopticnishingsetupmachineconceptandperipheraldevices .... 63 2-14SchematicrepresentationofarotatingfourinchopticsubmergedinmagnetizedreciprocatingMAFM .......................... 65 2-15Schematicofthefourinchopticnishingmachinewithmagneticyokecomponentsdelineatedinred ............................ 65 2-16Magneticeldsimulationsetup ........................... 67 2-17Variousgeometricmodicationsappliedtothepoletip .............. 68 2-18Plotscomparingthesimulatedmagneticuxdensity ............... 69 2-19ACADmodelandphotographofthenalpoletipdesign ............ 70 2-20CADmodelandphotographofthenalpoletipsassembledwiththemachine 70 3-1MirrorchipX-rayreectancemeasurementsetup ................. 72 3-2SchematicofX-rayreectancemeasurementapparatus ............. 73 3-3SchematicsoftheSWLIopticalpath,Mirauobjective,andMichelsonobjective 74 3-4FringecontrastversusZ-positionoftheSWLIobjective ............. 75 3-5SchematicofanatomicforcemicroscopesetupandtheLennard-Jonespotential ........................................ 76 3-6Separationoftheroughnessprolefromthemeasuredsurfaceprole ..... 77 3-7Surfaceproleltertransmissioncurves ...................... 77 3-8Schematicofsurfaceprolestatistics ....................... 78 9

PAGE 10

3-9SEMmicrographsofthemicroporesidewallsat60fromthemirrorchipface 80 3-10SWLImeasurementsofanunnishedandnishedsidewall ........... 81 3-11Photographofamirrorchippreparedforthenishingexperiment ........ 81 3-12X-rayreectancetestresults ............................ 82 3-13SurfaceroughnessmeasuredbySWLIof52unnishedand52nishedsidewalls ....................................... 83 3-14OpticalandSEMmicrographsofunnishedandnishedsidewalls ....... 84 3-15EDSspectra ..................................... 85 3-16Higher-magnicationSEMmicrographsofnishedandunnishedsidewalls .. 86 3-17AFMimagesofunnishedandnishedsidewalls ................. 87 3-18AFMimagesofbump-likefeaturesonasidewall ................. 87 3-19X-rayreectancetestresults ............................ 89 3-20PhotographoftheannealedDRIE-fabricatedmirrorchippreparedforanishingphase .................................... 90 3-21FESEMmicrographsofmicroporefeaturesbeforeandafternishing ...... 92 3-22SWLImeasurementsoftheofrepresentativesurfaceproles .......... 92 3-23SurfaceroughnesschangeswithsurfaceconditionbySWLIusinga100msamplinglength .................................... 94 3-24AFMimagesofrepresentativemicroporesidewallsafterDRIE,hydrogenannealing,andhydrogenannealing+MAF ..................... 95 3-25SurfaceroughnesschangeswithsurfaceconditionbyAFM ........... 96 3-26FESEMimagesofmicroporefeatureonaDRIE-fabricatedmirrorchipbeforeandafternishing .................................. 97 4-1Possibleabrasiveactionsonasurface ....................... 98 4-2Schematicrepresentationoftheprocessingprincipleforatworkpieces .... 99 4-3Theatworkpieceholderdesign .......................... 100 4-4Changesobservedinthesurfaceshapeafternishing .............. 102 4-5Schematicoftheworkpiecesurfacepattern .................... 103 4-6AFMimagesofthetrackedpocket ......................... 105 10

PAGE 11

4-7Changesinthemeasuredheightdifference,h,withnishingtime ........ 105 4-8Schematicofregionschosenforobservation ................... 106 4-9MeasuredvaluesofRzandRqwithnishingtime ................ 107 4-10Comparisonofsurfaceprolesinthethreetrackedregionswithpolishingtime 108 4-11Low-magnicationFESEMandBSEimagesoftheworkpiecesurface ..... 109 4-12High-magnicationFESEMandBSEimagesoftheworkpiecesurface ..... 109 4-13ComparisonofEDSspectra ............................. 110 4-14Phasecontrastimagesofthethreeregionsofinterestwithnishingtime .... 111 4-15Resultsofexperimentrepeatedwithoutabrasive ................. 113 4-16Schematicandcross-sectionofthemicroporereplicationsetup ......... 114 4-17Photographsofthemicroporereplicationsetupcomponents ........... 115 4-18Measurementsofthereplicatedmicroporecross-sectionaldimensions ..... 116 4-19AphotographshowingthemicroscopeandvideocameraplacementsforMAFMmotionobservation .............................. 117 4-20Photographsofthemicroporereplicationsetupusingabrasssheetastheworkpiecematerial .................................. 118 4-21SWLImeasurementsofthebrassworkpiecesurfacenishedbymagneticuidonly ....................................... 120 4-22SWLImeasurementsofthesurfacenishedwith50nmdiamondwater-baseddiamondslurry ............................. 123 4-23SWLImeasurementsofthesurfacenishedwith0.2mdiamonduniversalbasediamondslurry ................................. 124 4-24SWLImeasurementsofthesurfacenishedwith0.5mdiamonduniversalbasediamondslurry ................................. 125 4-25SWLImeasurementsofthesurfacenishedwith6mdiamondwater-baseddiamondslurry .................................... 126 4-26MeasuredchangesinRaversusmeanabrasiveparticlediameter ........ 126 4-27Detailsoftheregionsofinterestonthenishedsmooth-brassworkpiece ... 128 4-28Heightmapsoftheunnishedandnishedsmooth-brasssurfacemeasuredbyAFM ........................................ 130 11

PAGE 12

4-29Three-dimensionalAFM-measuredimagesoftheunnishedandnishedsurfacesandtheircorrespondingPSDcurvesandSavalues .......... 131 4-30ComparisonofPSDcurvesoftheunnishedandnishedsurfacesfromAFMandSWLImeasurements .............................. 132 4-31AFM-measuredSavaluesoftheunnishedandnishedsurfaces ........ 132 4-32Microporereplicationsetupforowvisualization ................. 133 4-33StillvideoframesofMAFMpreparedwith50nmwater-baseddiamondslurry 142 4-34StillvideoframesofMAFMpreparedwith0.5muniversalbasediamondslurry ......................................... 144 4-35Approximatevaluesfortheinnerdiameterofdipolarparticleringsandthedifferenceinpotentialenergybetweenlinearchainsandringcongurationsofdipoles ....................................... 147 4-36AFMimageofsingle3nm-deepnanoindentationinsilicon(100) ........ 149 A-1MAFMthroughamicroporeinazeroeld. ..................... 154 A-2MAFMthroughamicroporeinanalternatingmagneticeldat1Hz. ...... 155 A-3MAFMthroughamicroporeinanalternatingmagneticeldat5Hz. ...... 155 A-4MAFMthroughamicroporeinanalternatingmagneticeldat10Hz. ...... 156 A-5MAFMthroughamicroporeinanalternatingmagneticeldat15Hz. ...... 156 A-6MAFMthroughamicroporeinanalternatingmagneticeldat20Hz. ...... 157 A-7MAFMthroughamicroporeinanalternatingmagneticeldat25Hz. ...... 157 A-8MAFMthroughamicroporeinanalternatingmagneticeldat30Hz. ...... 158 A-9MAFMthroughamicroporeinanalternatingmagneticeldat35Hz. ...... 158 A-10MAFMthroughamicroporeinanalternatingmagneticeldat50Hz. ...... 159 A-11MAFMthroughamicroporeinanalternatingmagneticeldat45Hz. ...... 159 A-12MAFMthroughamicroporeinanalternatingmagneticeldat50Hz. ...... 160 12

PAGE 13

AbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulllmentoftheRequirementsfortheDegreeofDoctorofPhilosophyMAGNETICFIELD-ASSISTEDFINISHINGOFMICROPOREX-RAYOPTICSByRalEduardoRiverosAugust2012Chair:HitomiYamaguchiGreensletMajor:MechanicalEngineeringReducingthesizeandweightofspaceborneX-raytelescopesminimizesmissioncosts.In2006,Ezoeetal.rstintroducedmicroporeX-rayopticsmadefromthinsiliconwaferscalledMEMSmicroporeX-rayoptics.Theseopticsarelightweight(1000cm2effectiveareaper1kg)havecurvilinearmicroporesetchedbydeepreactiveionetching.Thisprocess,however,yieldsreectingsurfaceswithanunacceptablyhighsurfaceroughnessof30nmRq.ForefcientX-rayreectionandfocusing,amicroroughnessof3nmRqorbetterisneeded.Asuitablenishingprocessemployingalternatingmagneticeldassistednishingwasdeveloped.Thisprocessinvolvesareciprocatingmixtureofneabrasiveparticlesandmagneticuidactuatedbyanexternallyappliedalternatingandswitchingmagneticeld.Theopticissubmergedintheuidmixturewhichthenowsthroughtheoptic'smicropores,nishingthesidewalls(reectingsurfaces).Anishingmachinewasconstructedtorealizethisprincipleon7.57.5mmminiatureworkpiecescalledmirrorchips,andanothermachinewasconstructedtonishfull-sizedMEMSmicroporeX-rayoptics.Investigationsintothenishingcharacteristicswerecarriedoutonmirrorchips.Thesestudiesdemonstratedthenishingprocess'abilitytoremovematerialandcausereductionsinthemicroporesidewallsurfaceroughness.Additionally,improvementsinthemirrorchips'X-rayreectancewereobserved.Thenishingprocess'surfacemodicationmechanismwasalsostudiedthroughtheexecutionofvariousnishing 13

PAGE 14

experimentsonatworkpieces.Resultsfromtheseexperimentsshowedchangesinsurfacetexturewhichsuggestthatabrasiveparticlesmodifythemicroporesidewallsbyrepeatedindentation.Atheoryfortheabrasiveparticleactuationmechanismisproposed. 14

PAGE 15

CHAPTER1INTRODUCTION 1.1X-rayAstronomyInauguralNobelprizewinnerWilhelmConradRntgenwasthersttoexplicitlydetectandimageX-raysin1895[ 1 ].SinceRntgen'searlyobservations,theunusualnatureofX-rays,specically,theirabilitytoeasilytransmitthroughmaterialsopaquetovisibleandultravioletlight,theirabsorptionproperties,andtheirreluctancetorefractpiquedtheinterestsofmanyinthescienticcommunity.Asaresult,X-raysnowplayanintegralroleinmaterialanalysis,medicaldiagnosis,andastronomicalobservations.X-raysareahighenergyformofelectromagneticradiation;theirenergyroughlyrangesbetween100and100,000eV;forcomparison,visiblelightrangesbetween1.7and3.2eV[ 2 ].Inthisbriefintroduction,vefundamentalX-rayproductionmechanismsarepresented.Themostprevalentistheprocessofionization.Whentheenergyofaparticleorphotonincidentonanelectroninorbitexceedsitsionizationenergy,theelectronwillbeknockedoutoftheatomaprocesscalledinelasticscattering[ 3 ].Theatomisnowenergizedaboveitsgroundstate.Theelectronvacancyislledbytransitionofanotherelectron,restoringthegroundstateoftheatom.Thisdecayoftheatom'senergyisassociatedbytheemissionofaphotonhavingacharacteristicenergyequaltothedifferenceinenergybetweentheexcitedandgroundstates.Iftheejectedelectronoriginatedfromtheatom'sinnermostelectronshells,theenergyofthesubsequentemittedphotonwillbeintheX-rayrange[ 3 ].AnotherpotentialX-raygeneratingmechanismisblackbodyradiation.Matterwithtemperaturesaboveabsolutezerowillemitelectromagneticradiationoverawell-knowncontinuousspectrumofenergieswhosepeakemissionisdependentsolelyonitstemperature.Photonsareemittedfromthethermalmotionofchargedparticlesinthematter.Ifthetemperatureis106K,theblackbodyradiationfromthesurfaceofthematterwillbeintheX-rayrange[ 2 ]. 15

PAGE 16

Ifthetemperatureofagasishighenoughtoionizetheatomsaplasmaisformed,whereenergeticnegativelychargedelectronsandpositivelychargedionscomposethegas.Inthisstate,collisionsareprevalent,andenergyistransferredontheiroccurrence.Whenanelectronpassesnearanion,however,itsdirectionischangedduetoelectriceldinteraction.Thisaccelerationyieldselectromagneticradiation.Ifthegastemperatureis106K,theemittedradiationwilllieintheX-rayrange[ 2 ].Radiationgeneratedinthismannerisreferredtoasbrakingradiation,or,inGerman,bremsstrahlung[ 2 ].Synchrotronradiationisemittedwhenrelativistic(movingnearthespeedoflight)electronsaredeectedfromtheirlinearpathbyamagneticeld[ 4 ].ThistypeofradiationiscreatedinsynchrotrondevicesonEarthandisobservedinviolentcosmicevents.InverseComptonscatteringisaprocessthatgeneratesX-rayswhenanultra-relativisticelectronscollidewithamicrowavephotons(whicharepresentthroughoutallspaceinthecosmicmicrowavebackground).ThescatteredphotonsareshiftedtoX-rayenergylevelsbytheDopplereffect[ 2 ].AfterthediscoveryofX-rays,researchersbeganinvestigatingtheirpropertiesandwaystomanipulatethem[ 5 6 ].ThestudyofX-rayswasstrictlylimitedtoEarthlysourceshowever,sinceX-raysareabsorbedbytheatmosphere[ 2 ].AdvancesintechnologyandphysicaltheorypromptedcertainscientiststoattempttodiscoverthecompletesolaremissionspectrumincludedemissionintheX-rayband,whichdid,infact,exist[ 2 7 ].AnattempttodetectX-rayemissionfromtheEarth'smoonserendipitouslyresultedinthedetectionofabrightsourceofX-raysinthemoon'sbackground;thesourcewasnamedScoX-1[ 8 ].Thisdiscoverytriggeredgreatinterestamongastronomersandmoreexperimentswereplannedandexecutedtodiscoverothersources[ 2 ].In1970,theUhurusatellitebeganasurveyoftheentireskyintheX-rayspectralregionwhichbroughttolighttheexistenceofextremelypowerfulcosmicX-raysources. 16

PAGE 17

DatacollectedfromUhuruandsubsequentmissionshasbeenusedtodiscoverhighenergycosmicstructuresandthemechanismsthatproducetheX-raysthatreachEarth[ 2 ].X-rayemissionthushasbeendiscoveredfromalargevarietyofstructures;someofthesearelistedinTable 1-1 .TheassociatedX-raygeneratingmechanismsinTable 1-1 arenotnecessarilytheonlymechanismspresentinthelistedstructures.ThetwentiethcenturyhassinceseenthebirthanddevelopmentoftheeldofX-rayastronomy. Table1-1. SelectedcosmicstructuresandassociatedX-raygeneratingmechanisms StructureX-raygeneratingmechanisms AuroraBremsstrahlungCometsIonizationGaseousplanetsSynchrotron,IonizationNeutronstarsBlackbody,Synchrotron,SupernovaremnantsBlackbody,IonizationQuasarsInverseComptonscattering,SynchrotronBlackholesSynchrotron 1.2X-rayOpticsThedevelopmentoftheeldofX-rayastronomyiscoupledwiththeevolutionoftheX-rayimagingsystemsused.EarlyX-rayimagingsystemstypicallyinvolvedsomeformofacollimatormadeofahighatomicnumber(highZ)material;higherZmaterialsabsorbmoreX-raysduetotheirhigherelectrondensity[ 3 ].ThecollimatorwasplacedaboveadetectorwhichreactstotheincidentcollimatedX-rays.Thedetectorsinvolvedalayerorlmofagelatinouscolloidofphotosensitivecrystalswhichundergoachemicalchangeunderincidentlight;thesewerecalledphotographicplates[ 9 ].SubsequentX-raydetectorsincludedGeigercountersandtheirlower-voltagecounterpartscalledproportionalcounters,gasscintillationproportionalcounters,andsolidcrystal-basedscintillationcounters.Other,moreadvanceddetectorswerelaterdevelopedandusedonhighimpactX-rayobservatories;theseincludesolidstatedetectors,channelelectronmultipliers,microchannelplates,andcalorimeters[ 2 ].AdetaileddiscussionofX-raydetectorsliesbeyondthescopeofthiswork;however,relevantdetailsoftheX-raydetectorsusedinlaterchaptersshallbepresentedaccordingly. 17

PAGE 18

Comparedtovisiblelight,itisrelativelydifculttochangethedirectionofanX-ray;consequently,achievingfunctionalX-rayopticsisalsodifcult.Fortunately,anumberofadvanceshaveallowedfortherealizationofX-rayoptics;infact,theirdevelopmenthasburgeonedintoawellestablishedeldofstudy.Tounderstandtheirfunction,abasicgraspofthephysicsoflight-matterinteractionsisnecessary;thusthissectionwillreviewthephysicsanddesignofcertainX-rayopticalsystems.Therefractiveindex,n,ofamaterialistheratioofawave'sphasevelocitythroughareferencemedium(speedoflightinavacuum)tothephasevelocitythroughthemediumofinterest.Thisrelationshipisdescribedbythefollowingequation: n=c v(1)wherecisthephasevelocityofthewavethroughareferencemediumandviswave'sphasevelocitythroughthemediumwithwhichitisinteracting.Thephasevelocity,andthustherefractiveindex,ofawavethroughamediumdependsonboththemedium'spropertiessuchastheatomicscatteringfactoranddensityandthewave'spropertiessuchasitsenergyandelectric-eldvectororientation[ 10 ].Therefractiveindexforamediuminteractingwithradiationisdescribedbythefollowingrelation: n=1)]TJ /F9 11.955 Tf 11.95 0 Td[()]TJ /F2 11.955 Tf 11.96 0 Td[(i=1)]TJ /F2 11.955 Tf 13.16 8.09 Td[(r02 2Nat(f1)]TJ /F2 11.955 Tf 11.96 0 Td[(if2)(1)whereisdifferencebetweentherefractiveindexand1,isthematerial'sabsorptionindex,f1andf2aretherealandimaginarycomponentsofthematerial'satomicscatteringfactor,Natisthematerial'satomicdensity(atomsperunitvolume),r0istheradiusofanelectron,iistheimaginarynumber,andisthewavelengthoftheX-rayradiation[ 11 ].TherealpartofEquation 1 maybeapproximatedby: nr=Rer 0=r 1)]TJ /F9 11.955 Tf 13.15 8.94 Td[(!2p !2=1)]TJ /F9 11.955 Tf 11.95 0 Td[((1) 18

PAGE 19

whereisthedielectricconstantofthematerial,0isthevacuumdielectricconstant,!isthephotonfrequency,and!pistheplasmafrequencyofthematerial[ 12 ].Theplasmafrequency,!pisdescribedby: !2p=Ne2 m0(1)whereNistheelectrondensityofthematerial,eisthechargeoftheelectron,andmisthemassoftheelectron.TheplasmafrequencyofmostmaterialsrangeswithinthetensofeV[ 12 ].Forexample,theplasmafrequencyofgold(ametalcommonlyusedinX-rayapplications)is2.1831015Hz[ 13 ];asimplecalculationusingthePlanck-Einsteinequation: E=h(1)whereEisthephotonenergy,hisPlanck'sconstant,andisthefrequency(inthiscase=!p),revealsthattheequivalentenergyofthegoldplasmais9.0158eV.X-rayphotonenergiestypicallyrangebetween300eVto100,000eV;clearly,materialplasmafrequenciesarefarlowerthanX-rayphotonfrequencies.ThisdisparityresultsintherefractiveindexalwaysbeingslightlylessthanunityforX-raysinallmaterials[ 10 12 ].Whenelectromagneticradiationstrikesasurface,itwilleitherreect,refract,orbecomeabsorbedintothemedium.Figure 1-1 depictssuchinteractions.Theangleatwhichtheraywillrefractdependsontherefractiveindicesofthetwomediumsattheinterface.Snell'slawrelatestheincidentrayandrefractedrayanglesrelativetothesurfacenormalatthepointofincidence;itisshownhere, n1sin1=n2sin2(1)wheren1andn2aretherefractiveindicesoftwoadjacentmediumsand1and2aretheincidentandrefractedrayangles[ 14 ].ItisobservedthatincidentX-rays,havingrefractiveindicesofnearly1,willresultinminisculedifferencesbetween1and2.Figure 1-2 showsaschematicofaraypassingthroughaconvexlens.Asimplied 19

PAGE 20

versionofthelensmaker'sequationisshownhere, f=R=2 n)]TJ /F4 11.955 Tf 11.95 0 Td[(1(1)wherefistherearfocallength,andRistheradiusoflenssurface;thisequationistrueforathinsymmetricallens.Therefractiveindex,n,forX-raysresultsinexceedinglylongfocallengths,renderingrefractiveopticsunfeasible[ 12 ].Acompoundrefractivelenswasproposedforfocusing5keVX-rayswhichiscomposedofalineararrayofsmallholescausingrepeatedsuccessiverefraction;however,theseopticshaveverysmall(250m)aperturesandarelimitedintheirapplications[ 15 ].SincethediscoveryofX-raydiffractionincrystalsin1913[ 16 ],thedevelopmentofdiffraction-basedopticssincegrown;however,theseopticsarelimitedtospecicwavelengthsandarethuslimitedintheirapplications. Figure1-1. Schematicdepictingtheanglesofreectedandrefractedraysrelativetotheincidentrayangle Figure1-2. Schematicofaparaxialrayrefractingthroughathinsymmetricallens 20

PAGE 21

Insteadofrefractiveordiffractiveoptics,reectiveopticsmaybeused.Unfortunately,X-raysareeasilyabsorbedbymostmaterialsthusreectionisalmostnonexistent,exceptinthecaseoftotalreection.Totalreectionisaphenomenonthatoccurswhenanelectromagneticwavestrikesaninterface(suchasthatofFigure 1-1 ).Iftheangleofincidenceislargeenoughatapoint,nolightwillbeabsorbed,andallthelightwillbetotallyreected.ThisangleiscalculatedbyEquation 1 ,bysettingtherefractionangle90andsolvingfortheincidentangleasshowninEquation 1 .Thisangleiscalledthecriticalangleandraysincidentatangleslessthanthecriticalanglearesaidtobeatgrazingincidence.2=90sinc=n2 n1c=arcsinn2 n1 (1)ThetotalreectionofX-rayswasobservedin1923byArthurCompton[ 6 ].CriticalanglesforX-raysareusuallynolargerthan2;however,thisisenoughtobuildreasonably-sizedX-rayimagingsystemswithmultiplereections.In1948,PaulKirkpatrickandAlbertVinicioBaezpublishedanX-rayopticdesignusingsphericalsurfacestoimageX-raysbygrazingincidencereections[ 17 ].Theopticssufferedfromsevereaberrations,however.ItisofgreatinteresttoastronomerstoclearlyimageX-raysourcesfromthesky.AnimagingsystemmustsatisfytheAbbesineconditiontoproducesharpimages[ 14 18 ].Fordistantobjects(objectsatinnitywhichemitparaxialrays)throughacoma-freelens,theAbbesineconditionisasfollows, sin=)]TJ /F2 11.955 Tf 10.49 8.09 Td[(h f(1)whereistheangleoftherefractedraytotheopticalaxis,histheheightoftherayfromtheopticalaxis,andfistherearfocallength,asshowninFigure 1-2 [ 14 ].Thiscondition 21

PAGE 22

relatesthefocallengthtothelenscurvaturebyaconstantratioofsinesbetweentheincidentandrefractedraywithrespecttoh.Thus,broad-energy-bandX-rayimagingopticsmustchangethedirectionofX-raysbytotalreectionandmustalsosatisfytheAbbesinecondition.In1952,HansWolterpublishedthreecongurationsofparaboloidalandhyperboloidalmirrorsintendedtofocusX-raysbygrazingincidencereection[ 19 ].ThesecongurationssatisfytheAbbesinecondition.Wolter'sTypeItelescopecongurationworksparticularlywellasitallowsforsmall(<1)reectionangles.ThustheWoltertypeIopticdesignisthemostsuitableforimagingcelestialX-raysources.AschematicoftheWoltertypeIopticdesignisshowninFigure 1-3 .ThefocusofaparabolacoincideswiththefocusF1ofthehyperbola.Thedimensionsoftheparabolaandhyperbolaarerelatedbyaspecicsetofequations[ 20 21 ].Theraysentertheopticandreectrstofftheparabola,thenoffthehyperbola,focusingtothehyperbolicfocusF2(inreality,thefocallengthisslightlylonger)[ 12 ].Whenrevolvedabouttheopticalaxis,asetofparaboloidal(primary)andhyperboloidal(secondary)surfacesresults.Mirrorsareguredtoconformtothesesurfacesoversomespanalongtheopticalaxis. Figure1-3. SchematicofaWoltertypeIgrazingincidencetelescopeconguration 22

PAGE 23

Theeffectiveareaistheareaonasurface(thesky)normaltotheopticalaxisthroughwhichX-raysentertheopticandarefocused;inotherwords,itistheprojectionofthereectingmirrorsurfacesontothesky.TheWoltertypeIdesignshowninFigure 1-3 ,thoughfunctional,doesnothaveenougheffectiveareatoachievehighsensitivity(doesnotcollectenoughphotonstoimagedimobjects).Asolutiontothisisnestingseveraltelescopesofdiminishingradiihavingcoincidentfocalpointsandopticalaxes;itshouldbenotedthatthereareotherimportantfactorswhichcontributetotheeffectivearea[ 12 ].Figure 1-4 showsacut-awayviewofthistypeofanestedcongurationoffourtelescopes.Themirrorthicknessaffectstheeffectiveareaandisoftenminimized.WoltertypeImirrorshavebeenfabricatedoutofglassshellsofafewcentimetersradialthickness;theyareusuallycoatedwithathinmetalliclayer.ThistypeofnestedX-rayopticalsystemwasusedontheEinstein[ 22 ],RntgenSatellite(ROSAT)[ 23 ],andChandraobservatories[ 24 ].Theseglassshellopticsexhibithighangularresolution;butareheavy(Chandra'smirrorassemblyweighs1600kg[ 25 ])andcostlytomanufacture.EffortstoincreasetheeffectiveareaofX-rayopticsoftencompromisetheimagingsystem'sresolution.Mirrorsformedviaelectroformednickelreplication(ENR)havelargereffectiveareasandmoderateresolution[ 26 ].SegmentedfoilmirrorsformedbymandrelshapereplicationarelightweightcomparedtoglassshellorENRoptics;however,theyhavemuchlowerresolutionduetotheirconicalWoltertypeIapproximatedgeometry[ 27 ].Thermallyslumpedglassmirrorshavehighreectivity,however,theirresolutionislimitedbythemirrorgureerrorswhichareachallengetocontrol[ 28 ].Pore-basedopticsaremeanttobelightweightX-rayopticalternatives.Glassmicroporeopticsaremadefromfusingthousandsofthin(hundredsofmicrons)square-crosssectionleadglassberstogether;theyaremuchlighterinweightthanotherX-rayoptics;butarelimitedintheirresolutionandsubsequentlytheirapplications 23

PAGE 24

[ 29 30 ].Siliconporeoptics(SPO)aremadefromstackingsiliconwaferswithanetchedarraysofsquarechannels(11mm).SPOspromisehighreectivityandgreatlyincreasedeffectiveareas;however,waferalignmentissuesandmountinghardwarenegativelyaffecttheresolutionandweightrespectively[ 31 ].OpticaldesignsofX-rayopticsaremodieddependingontheirapplications(wide-eld,spectroscopy,etc.)aswell,whichheavilyinuencestheirmechanicaldesignapproach;thus,thereisnosingleopticsuitableforallapplications. Figure1-4. Schematicofacut-awayviewofaWoltertypeInestedmirrorarrangement TherearegeneraldifcultiesassociatedwithWoltertypeIopticsaswell.Theimageproducedwillhavesevereeldcurvatureforoffaxisobjects;thisisaconsequenceofitsdesign[ 12 21 ].Asalludedtoearlier,thealignmentofthemirrors,detectors,andothersensorsisalwayschallenging.Thedesignofthemirrorsupportsisalwayscarefullyconsidered,asthemirrorswillsaganddeformundertheirownweight.Thetotalweightoftheopticassemblymaybealimitingdesignfactor.ThegureandnishspecicationsofX-raymirrorsareamongthemostchallengingtoachieve;forexample,themirrorsontheChandratelescopeareconsideredbysomeasthemostprecisemirrorseverfabricatedbymankind[ 32 ].Asaresult,WoltertypeIopticsaredifcultandconsequentlycostlytobuild.Table 1-2 comparestheperformanceofafewX-rayoptic 24

PAGE 25

mirrortechnologies;themirrortechnologiesshowninTable 1-2 aregenerallyrestrictedtothosehavingastronomicalapplications. 1.3MEMSMicroporeX-rayOptics 1.3.1BackgroundFiguringannularmirrorstoconformtoparaboloidalandhyperboloidalsurfacesoftenrequiresaminimummirrorthicknesstosupportthedesiredshape.Thisnotonlyincreasestheweightoftheopticsbutalsoreducestheeffectiveareaforagiventelescoperadius.Reducedeffectiveareadiminishestheoptic'sabilitytoimageweakX-raysources.PetreandSerlemitsosproposedaconicalapproximationofaWoltertypeIopticin1985[ 33 ].Theyarguedthatnestingthinconicalcoatedaluminumshellsinplaceoftheparaboloidalandhyperboloidalmirrorsurfaceswouldresultingreatlyincreasedeffectiveareaandreducedcost.ForhigherenergyX-rays(>4keV),aconicallyapproximatedWoltertypeIopticdesignwouldexhibitonlyminimallyreducedresolvingpower[ 33 ].SchmidtproposedawideeldofviewX-rayopticbasedoffaone-dimensionalarrayofatreectorsin1975[ 34 ].Thesewouldbeexpandedtotwostackedorthogonallyorientedreectorarraysknownaslobstereyeopticsduetotheiranalogousnesswiththeeyestructuresofmacrurancrustaceans[ 35 ].WilkinsandStevensonproposedtheuseofbentleadglassmicrochannelplates(MCP)asX-rayfocusingdevicesin1985[ 36 ].Thesedevicesofferedgreatexibilityinapplicationandlowweight,andtheycouldbearrangedaseithercollimators,lobstereyeoptics,orconicalapproximationsofWoltertypeIoptics[ 36 ].In1999,Peelepresentedlobstereyeopticsfabricatedfromelectrochemicallyetchedsiliconwaferswithsquareporewidthsof4.8.0m;however,thereectingsurfacesexhibitedexcessivesurfaceroughness[ 37 ].Peele'sinvestigationswereamongtheearliesttoemploymicromachiningtechnologiestofabricateX-rayoptics. 25

PAGE 26

Table1-2. Comparisonofselectmirrortechnologies MirrortechnologyResolutionAdvantagesDisadvantagesAssociatedmissions Glassshell0.5arcsecHighresolution,rigidSmalleffectivearea,heavy,costlyEinstein,ROSAT,ChandraElectroformednickelreplication15arcsecLargeeffectivearea,moderatelyhighresolution,moderatelyrigidCostly,mirroralignmentdifculty,heavyXMM-NewtonSegmentedreplicatedfoil2arcminLargeeffectivearea,lightweight,broadenergybandLowresolution,mirroralignmentdifcultySuzakuSlumpedglass5arcsecBroadenergyband,highreectivity,moderateweightMirrorgurecontroldifculty,thinmirrorsagNuSTARGlassmicropore1arcminMultipleopticalcongurationsandapplicationspossible,lightweight,smallsizeManufacturingdifculty,mirroralignment,possibleunwanteddiffractionBepiColombo,ISSSiliconpore5arcsecHighreectivity,verylargeeffectiveareaMirrorgurecontroldifculty,heavy,mirroralignmentXEUS,IXO 26

PAGE 27

In2006,Ezoeetal.proposedanultra-lightweightpore-basedX-rayfocusingoptic[ 38 ].Theoriginalconceptinvolvedanannulararrangementofmanysmallsections(calledmirrorchips)ofasiliconwaferonaspeciallydesignedholder.Thesemirrorchipswere550.3mmtrapezoidaldicedp-typesilicon(110)wafers.Anarrayoflinearmicrometer-sizedpores(micropores)wasetchedintothemirrorchips.Aftersomeexperimentation,theetchingwasaccomplishedusingpotassiumhydroxide(KOH)(wet-etching)solutionandawell-alignedsiliconnitridemask,patternedbyadry-etchingprocess.Thequalityofthemicroporesidewallsurfacewasimprovedbytheadditionofultrasonicwavesduringtheetchingphase(reducedfrom2.1nmto0.5nmrootmeansquareroughness(rms),ina33marea)[ 39 40 ].Themicroporesidewallscoincidedwiththe(111)planes,andsoftX-rayreectionofftheporesidewallswasdemonstrated[ 38 40 ].Thesurfaceroughnessofthesidewalls,asestimatedbyaspecializedX-rayreectometrymeasurement,wasfoundtobe6nm[ 41 ].Sincetheaforementionedphotolithographyandwet-etchingtechniquesareusedinthefabricationofmicroelectromechanicalsystems(MEMS),theseopticswerecalledMEMSmicroporeX-rayoptics.Thoughextremelylightinweight,theseopticshadatheoreticalangularresolutionlimitofseveralarcminutes,sincetheWoltertypeImirrorsurfaceswereapproximatedbyatdiscontinuousplanes[ 42 ].Additionally,theireffectiveareawassmallcomparedtootherpore-basedopticsofsimilarradius.Toresolvethelimitationsoftheseoptics,theopticwasredesignedtoconformtoaconicallyapproximatedWoltertypeIconguration,andanewmanufacturingprocesswasdeveloped;thisdesignwasrstpresentedin2009andwillbepresentedinthefollowingsection[ 42 ]. 1.3.2OpticalDesignFigure 1-5 showsasimpliedschematicofatwostageMEMSmicroporeX-rayopticsystem.Curvilinearmicroporesareetchedthroughthethicknessofathin(300m)substrateinanaxisymmetricradialpatternofvaryingporewidth(5m).Inthe 27

PAGE 28

congurationofFigure 1-5 ,atwostagesystemisshownwheretheprimarystagecontainsnestedconicallyapproximatedparaboloidmirrorssurfaces(poresidewalls),andthesecondarystagecontainsnestedconicallyapproximatedhyperboloidmirrorsurfaces.Figure 1-5 showsonlyfournestedsetsofmirrors;however,arealdesignmaycontainhundreds.NotlabeledinFigure 1-5 arethefourradialmirrorsupports(at45,135,225,and315);inarealdesign,moresupportswouldbeincludedaswell.Theprimaryandsecondarystagesaresphericallycurvedtocorrectlypositionthemirrorsthattheirintendedinclinationangleswhichmayrangebetween0.8and1.7[ 39 ].Thefocallengthisapproximatelyhalfthesphericalradius. Figure1-5. Schematicofatwo-reectionMEMSmicroporeX-rayoptic Ifrealized,thisMEMSmicroporeX-rayopticsystemwouldbethelightesttypeofX-rayopticyet.ItsmaximumachievableresolutionistheoreticallylimitedbyX-ray 28

PAGE 29

diffractionthroughthemicropores.Thislimitisdescribedbythefollowingrelationship, d(1)whereistheangularresolution,isthewavelength,anddisthewidthofthemicropore[ 43 ].Figure 1-6 showscalculationsofpotentialangularresolutionsforarangeofincidentenergiesandmicroporewidths.UsingEquation 1 ,anincidentphotonenergyof1keV,anda20mporewidth,anangularresolutionof10arcsecispossible.ItisclearfromFigure 1-6 thatthereexistssomelowermicroporewidthlimit,dependingonthedesiredenergybandandresolutiongoalsforthespecicsetofoptics;thelowermicroporewidthlimitmaybedependentontheopticmanufacturingprocessesaswell.Itisthereforedesirabletochooseareasonablysmallporewidthtomaximizetheoptic'seffectiveareabyallowingtheinclusionofmorenestedmirrors. Figure1-6. Angularresolutionlimitsforarangeofincidentenergiesandmicroporewidths Anotherfactorwhichaffectstheresolutionistheconicalapproximation;fortunately,ashorteraxialmirrorlengthreducestheeffectoftheconicalapproximation.Forexample,ifasubstratethicknessof1mmischosen,thelossinresolutionwouldbe1arcsec[ 38 ].Therefore,itisdesirabletolimitthethicknessofthesubstratetominimizeeffectoftheconicalapproximation. 29

PAGE 30

Thefabricationofthisopticisatechnicallychallengingtaskrequiringtheuseofthelatestmicrofabricationtechniques.TherearethreebasicconcernsinthefabricationofaMEMSmicroporeX-rayoptic;theseare, 1. Productionofthehigh-aspectratiomicroporestructures 2. Settingthemirrorinclinationangles 3. Controlofthemirrorsurfacequality.Theprocesschosentocreatetheporestructuresmustyieldporeswithsidewalls90tothesubstratesurfaceandhighshapeaccuracy.Themirrorinclinationanglesmustbesetbysphericaldeformationoftheoptic;thesuccessofthisprocesswillhaveasignicanteffectontheoptic'sabilitytosharplyfocusobjects.Itisinterestingtonotethat,incomparisontootherX-rayopticssystemssuchasthoseofshowninTable 1-2 ,themirrorswillbefabricatedoutofthesamesubstrate(monolithicstructure);thismayreducecertainimageaberrationsduetomirrormisalignments.Themirror(microporesidewall)surfacequalitymustbeasclosetoaperfectlysmoothsurfaceaspossibleforefcientX-rayreectionandfocusing. 1.3.3ManufacturingMethodsTodate,MEMSmicroporeX-rayopticshavebeenfabricatedoutofsinglecrystalsiliconandnickel.SiliconMEMSmicroporeopticsareetchedfromsiliconwafersusingdeepreactiveionetching;themicroporesidewalls(mirrors)aresmoothedbyhydrogenannealing,andtheirinclinationsaresetbyhigh-temperatureplasticdeformationoftheoptic.NickelMEMSmicroporeopticsmaybeformedbyX-rayLIGA.Thefollowingsubsectionswilldetailthesemanufacturingprocesses. 1.3.3.1DeepreactiveionetchingReactiveionetching(RIE)isawellestablishedmicrofabricationprocessinwhichfreeradicalsandionsfromaplasmaareacceleratedtowardsaworkpiece.Asaresult,avarietyofchemicalandphysicalprocessoccurwhichanisotropicallyremoveworkpiecesurfacematerial[ 44 ].Thetermdeepreactiveionetching(DRIE)referstoanRIE 30

PAGE 31

processwhichhasbeenmodiedsuchthatdeepandhigh-aspectratiostructuresmaybesuccessfullyfabricated.DRIEisapotentialmethodoffabricatingthemicroporesonMEMSmicroporeX-rayopticsonasiliconsubstrate.Figure 1-7 (A)showsaschematicofanappropriatelithographyprocessforfabricatingmicroporesinsilicon.Athin(200nm)layerofaluminumisdepositedonthesurfaceofasiliconwafer.Alayerofphotoresistisappliedtothealuminumsurfaceandexposedtoultraviolet(UV)radiationthroughapatternedUVlightmask.Thephotoresististhendevelopedtodissolvetheexposedareas.Anetchantisappliedtotheexposedaluminumsurfacetodissolveit.Thiswetetchingprocessistimedependentandmayexhibitundercut;thusthethicknessofthealuminumlayerinstep(i)shouldbecontrolledaswellaspossible,sinceinconsistencieswillresultinover/underwidthmicropores.Thephotoresistisremoved,andthewaferisplacedintheDRIEchamber.AftertheDRIEisnished,thealuminumlayermaybepeeledofdissolved,andthesiliconwafermaythenbecleanedtoremoveanyprocessdebris.AschematicoftheDRIEchamberisshowninFigure 1-7 (B).EithersulfurhexauorideSF6oroctauorocyclobutaneC4F8owsintothechamberandisexcitedandionizedbyamagneticeldwhichoscillatesatradiofrequencies(RF),forminganinductivelycoupleplasma(ICP).TheplasmaisacceleratedbyanelectriceldtowardsthesiliconwaferworkpiecewhichsitsatopaatelectrodewhichisconnectedtoanRFpowersupply.Thealternatingelectriceldoftheelectrodeinteractswiththechargedplasma.Therearemultiplemechanismswhichoccurduringthisprocesswhichcombinetoachievematerialremoval.Figure 1-7 (Bi)attemptstovisuallyportraythesemechanisms[ 44 ].Theetchingofthesurfaceoccursinthefollowingmanner.Amagneticeldformsaplasmafromtheincomingchambergasbydissociatesthegasmoleculesintoions,freeradicals,andothermolecules.Freeelectronsfromtheplasmaareremovedeitherbyconductionthroughthechamberwallorbyadsorptionontothewafersurface.The 31

PAGE 32

Figure1-7. SchematicofaA)microporeetchinglithographyprocess,B)anICPDRIEchamber,andB)theBoschprocess plasmasubsequentlydevelopsapositivenetchargewhiletheelectrode/workpieceformsanegativecharge;thus,adirectcurrent(DC)biasisformed.TheDCbiasacceleratesthepositivelychargedplasmacomponentstowardstheworkpiece.Thesecomponentsarereactivespeciesandincludefreeelectrons,ions,anddissociatedmolecules.Thereactivespeciesareadsorbedontotheworkpiecesurfaceandreactwiththeworkpiecesurfaceatoms.WhenSF6isfedintothechamber,compoundsformedwiththesiliconsurfacewillbeSiFxwherex4.Theproductsofthesereactionsare 32

PAGE 33

eithervolatileandleavethesurfaceimmediately,ortheyaresputteredoffbyotherincidentreactivespecies.Thesedesorbedproductsexitthechamberviathegasoutlet/pumpsystem[ 44 ].Sulfurcompoundsareunlikelytobedepositedonthesiliconsurfaceunderetchingconditions;theyaremostlyevacuatedfromthechamber[ 45 ].InDRIE,thisetchingprocessishaltedandasurfacepassivationstepisexecuted.AstheshowninFigure 1-7 (Cii),theowofSF6iscutoffandtheowofC4F8islet.WhenthisgasisdissociatedbytheICP,freeradicalsstriketheexposedsiliconsurface,formingapolymericlayer.ThegasisthenswitchedtoSF6andtheetchingisrepeated;however,theremainingpolymerlayeronthesidewallsurfacesprotectsthesiliconsidewallsurfacesfromfurtheretching(undercut)[ 45 ].Thesuccessivealternationoftheetchingandpassivationphasesallowsfordeepstructurestobefabricated;thisisknownastheBoschprocess,namedaftertheGermancompanyRobertBoschGmbHwhereitwasdeveloped[ 45 ].TherstMEMSmicroporeX-rayopticsfabricatedbyDRIEwerereportedin2009[ 42 43 ].Theseopticsexhibitedanunacceptablyhighmicroporesidewallsurfaceroughness(20nmrms)[ 43 ].Figure 1-8 showsaphotographofafourinchMEMSmicroporeX-rayopticfabricatedbyDRIE;theopticshownis300mthickanditsmicroporesare20minwidth.Thoughtheroughnessofthemirrorswashigh,thisDRIEprocess,byfabricatingallmirrorsatonce,promisesarelativelyinexpensivemethodofX-rayopticfabrication. 1.3.3.2HydrogenannealingHydrogenannealingmaybeappliedtoimprovethemicroporesidewallsurfaceofasiliconMEMSmicroporeopticetchedbyDRIE.Inthisprocess,theopticisplacedinanovenhavingapressurizedowofpurehydrogenH2gas.Thetemperatureisraisedbetween900Cand1300C,justbelowthesiliconmeltingtemperatureof1414C.Whenannealinginanultrahighvacuumenvironment,thesurfaceatomsself-diffusetolowersurfacechemicalpotentialpositions,effectivelysmoothingthesurface[ 46 47 ]. 33

PAGE 34

Figure1-8. PhotographofaMEMSmicroporeX-rayopticmadebyDRIE(PhotographcourtesyofRaulRiveros) However,whenhydrogengasisintroducedintotheheatedenvironment,itdissociatesandadsorbsontothesiliconsurface[ 48 ].Thepresenceofhydrogenatomsonthesurfaceisbelievedtoslowtherateofsurfaceatomself-diffusion[ 47 49 ].Aslowerrateofsurfaceatomself-diffusionallowsforbettercontroloftheannealingtimeandsubsequently,microstructuregeometry.Hydrogenannealingessentiallyroundscornergeometries;thus,microstructurespresentonasiliconsurfacemightbeinadvertentlyerasedordefunctionalizedbyover-annealing.Accordingly,thereductionofthesurfaceroughnessonavarietyofsiliconworkpieceshasbeendemonstrated.SatoandYoneharashowedthathydrogenannealingreducedtheroughnessoftheirsilicon-on-insulator(SOI)wafersfrom10nmto0.09nmrms(11marea)[ 50 ].Gaoetal.showedtheeffectsofhydrogenannealingonthesidewallsofareactiveionetched(RIE)siliconopticalwaveguide;theyfoundthatthesidewallroughnessreducedfrom12.6nmto0.3nmrms(22marea)[ 51 ].Kuribayashietal.demonstratedthesurfaceroughnessreductionofthesidewallsofRIEetched0.7mwideand3mdeeptrenchesonsiliconfrom1.72nmto0.11nmrms(1.51.5marea)[ 47 ]. 34

PAGE 35

ThemicroporesidewallsurfaceroughnessofMEMSmicroporeX-rayopticshasbeenreducedbyhydrogenannealingfrom34nmto0.72nmrms(11marea)[ 52 ].Thisreductioninsurfaceroughnesswasconrmedbyscanningelectronmicroscopy[ 43 52 ].Itmaybepossibletofurtherreducethesidewallsurfaceroughnessbyhydrogenannealing;however,excessiveroundingofthemicroporeedgeswouldoccur,affectingtheoptic'sfunctionality. 1.3.3.3PlasticdeformationofsiliconwafersTointroducetheappropriateinclinationangles(totheopticalaxis)tomicroporesidewallsurfaces(mirrors)onaMEMSmicroporeoptic,asphericalcurvaturemustbeintroducedintotheatsiliconwafer.Elasticdeformationmightbeapossibilityforlargerradii;however,theopticisseverelyweakenedbythepresenceofmicropores.Itisassumedthatstressconcentrationswouldcauseundesiredyieldingandfailureoftheopticstructure.In2004,Nakajimaetal.demonstratedthatitispossibletoplasticallydeformsinglecrystalsiliconwaferswhilestillpreservingtheircrystalstructure[ 53 ].Intheirprocess,asiliconwaferispressedbetweentwosphericallyshapedgraphitedies,oneconvexandoneconcave,witha200Nforce.Thediesandwaferareinahydrogengasenvironment(topreventoxidation)at1380C.Apparently,becauseofthecrystallographicsymmetryofsilicon,themostaccuratedeformationoccurswithsilicon(111)wafers,forthe{111}planesdevelopaspecicslip-bandsystem[ 53 ].ThisprocessisfeasiblefordeformingsiliconwafersforuseasX-rayreectionmirrors,asdemonstratedbyseveralstudies[ 54 56 ].Mikaetal.showedthedeformationofsiliconwafersintonon-sphericalgeometry[ 56 ].ThisprocesshasalsobeenappliedtoDRIEetchedsiliconMEMSmicroporeX-rayoptics.Asphericalradiusof1033mmwasachievedonaMEMSmicroporeX-rayoptic[ 57 ],andX-rayfocusingtestrevealedthattheoptichadaresolutionof20arcmin[ 58 ].AnotherMEMSmicroporeX-rayopticwastestedwhichhadameasuredangularresolutionof14arcmin[ 59 ]. 35

PAGE 36

1.3.3.4X-rayLIGAIn1972,SpearsandSmithrstpresentedapatterningtechnique,calledX-raylithography,capableofcreating1m-sizedfeatureswithouttheuseofcomplicatedfocusingoptics[ 60 ].Intheirprocess,X-raysfromarelativelydimX-raysourcewereshonethroughpatternedthinheavy-metalmaskwhichwasplacedafewmicrometersabovethetargetworkpiecemadeofpolymethylmethacrylate(PMMA).TheexposedpolymerchainswerescisedbytheincidentX-rayradiationandwereeasilyremovedbyasolventmixtureofmethylisobutylketoneandisopropylalcohol,leavingasetoftrencheswhichpreciselyreplicatedthemaskgeometry[ 60 ].In1982,Becker,Ehrfeld,andMnchmeyercombinedX-raylithographywithelectrodepositioninasuccessfulefforttoproducehigh-aspectratiomicroscalegasseparationnozzlesfortheenrichmentofuranium-235isotopes.TheirmethodutilizesX-raylithographytocreatepolymericprotrusionsonametallicsubstratewhichsubsequentlyactsasamoldforgalvanoplasticnozzlefabrication,aprocesstheynamedLIGA[ 61 ].TheabilityoftheLIGAprocesstocreatedeep(hundredsofmicrometers)structuresisduetotheuseofhigh-uxsynchrotronradiation(X-rays)whichpenetratesdeepintothethicknessofthePMMA.Figure 1-9 showstwomethodsofLIGAfabrication.Method1,showninFigure 1-9 (A)beginswiththeexposureofabulkofPMMAtosynchrotron-generatedX-raysthroughapatternedmask.Thetimeofexposureandthebeamintensitydistributioniscarefullyconsidered[ 62 ].TheexposedPMMAisdamagedbytheradiationandiseasilydissolved,leavinganegativemold.Themoldisthensputteredwithathinmetallayer,introducinganelectricallyconductivesurfacetothemold.Electroformingonthesputteredlayersurfaceisdoneandtheexcessisremovedbynexedabrasives.Nickelisoftenchosenasthestructurematerialbecauseitiscorrosionresistantandtheelectrolyticuid(nickelsulphamate)exhibitshighmicrothrowingpower[ 63 ].TheremainingPMMAisdissolved,renderingthedesiredstructurewiththesputteredmetallayerontheverticalwallfeatures,whichmaynotbedesired.Thereisthepossibilityof 36

PAGE 37

H2gasbeingtrappedinthestructureduringelectroformingsincematerialwilldepositinsideatrenchinthreesimultaneousdirectionsbutnotnecessarilyatequalrates[ 64 ].Additionally,itispossibletodamagethestructureduringgrinding(step5)aslargecuttingforcesmaycausestructuraldeformations. Figure1-9. SchematicdepictionoftwomethodsofLIGAfabrication Method2ofFigure 1-9 (B)issimilartothatusedbyBeckerin1982.APMMAlayerisdepositedinliquidformontoametallicsubstrate.Itisexposedtosynchrotronradiationanddeveloped.ThenickeliselectroformedontotheexposedmetallicsubstrateandgrowsaroundthePMMAmoldwalls.ThestructureisthendislodgedfromthemetalsubstrateandthePMMAisdissolved.Method2islesspotentiallydamagingtothestructurethanmethod1;however,duringimmersionofthemoldintotheelectrolyticuid,thesurfacetensionoftheuidmightapplyamomentatthebaseofthefree-standingPMMAstructures.IfthebondbetweenthePMMAmoldstructureand 37

PAGE 38

themetallicsubstrateissomehowawed,itmightseparate;inthisrespect,method1issuperiorasthePMMAstructureismonolithic[ 64 ].StructurescreatedbyLIGAtypicallyexhibitafeaturesidewallsurfaceroughnessbetween10nmand30nmrms[ 65 ].LIGAhasbeenusedtoproducenickelMEMSmicroporeX-rayopticsbymethod1withmicroporeasidewallroughnessof10nmrms[ 66 ];Figure 1-10 showsaminiatureX-rayoptic(calledamirrorchip)andamoldfabricatedbymethod1.Method2wasappliedtofabricateMEMSX-raymicroporeopticswithhighgeometricaccuracy[ 64 ].SinceLIGArequiresasynchrotronsource,productionofLIGA-fabricatedopticsislimitedwhichunfortunatelylimitstheirdevelopment.IfafullsetofMEMSmicroporeX-rayopticswerefabricated,themirrorinclinationanglescouldbesetbyelasticdeformationsincenickelisnotabrittlematerial. Figure1-10. PhotographofaLIGA-fabricatedA)PMMAmoldandB)nickelmirrorchipbymethod1ofFigure 1-9 (A)(PhotographscourtesyofRaulRiveros) 1.4MagneticField-assistedFinishing 1.4.1Background,Fundamentals,andApplicationsAmagneticeldgeneratedononesideofanonmagneticobjectwilllikelypermeateitandreachtheotherside.Amagnetictoolonthesideoppositetothemagneticeldgeneratorwillinteractwiththemagneticeld.SuchaninteractionisdepictedinFigure 1-11 whereapermanentmagnetforcesamagnetictoolagainstthesurface 38

PAGE 39

ofanonmagneticworkpiece.ReferringtoFigure 1-11 ,verticaldisplacementofthepermanentmagnetwouldresultinasimilardisplacementofthemagnetictool.Ifthetoolpossessesabrasiveproperties,abrasionoftheworkpiecesurfacemayresult.Thisremotemanipulationofamagnetictoolbyamagneticeldisthebasisforasurfacenishingtechniquecalledmagneticeld-assistednishing(MAF)[ 67 ]. Figure1-11. Schematicofapermanentmagnetcontrollingamagnetictoolthroughanonmagneticworkpiece ItappearsthattheearliestconceptionofanMAF-typeprocessoccurredduringthe1930s,astherstknownpatentonsuchaprocesswasappliedforin1938[ 67 68 ].Itdidnot,however,developintoasubjectofscienticresearchuntilthe1960s[ 69 70 ].MAFhassincebeenappliedtoawidevarietyofapplicationsinmanufacturing,electronics,medicine,aerospace,andotherindustries.MAFofferssomeadvantagesoverconventionalsurfacenishingtechniques;thesewillbeexplainedinthefollowingparagraphs.PerhapsthemostprominentadvantageofMAFistheabilitytomachinesurfaceswhichareinaccessibletoconventionalnishingprocesses.Ifthemagnetictoolsaresmall,manysuchtoolsmaybeintroducedontothetargetsurface.Thesmallmagnetictools,underanappliedmagneticeld,willformchain-likestructuresalongthelinesofmagneticuxasvisualizedinFigure 1-12 .Ifthegradientofthemagneticeldintensityisgreatenough,themagnetictoolswillgrouptogether.ShinmuraandYamaguchiappliedthisbehaviortotheinternalsurfacenishingofcleangascontainers.Fortygramsofamixtureofmagneticparticlesandabrasivemagneticparticleswasintroduced 39

PAGE 40

intoacleangascontainer'ssmallopening.Thecontainerwasrotatedwhilestationarypermanentmagnetsoutsideofthecontainerheldthemagnetictoolmixturesteadilyagainsttheinnercontainersurface.Theabrasivetooledgescombinedwithmagneticforce-derivedpressureabradedtheinnercontainersurface,resultinginapeaktovalleyheightreductionfrom7mto0.2m(0.8mmprolelength)[ 71 ].Thisinternalsurfacenishingmethodhasbeensimilarlyappliedtotheinternalnishingoftubeswhereinthetubeisrotatedandstationarymagnetsholdmagnetictoolssteadyagainsttherotatinginnertubesurface[ 72 74 ].Byrotatingthemagnetsaroundthetube,curved/complextubesmaybenishedinternallyaswell[ 75 ]. Figure1-12. Photographofironparticlesinachain-likestructurealongthelinesofmagneticuxsurroundingapermanentmagnet(PhotographcourtesyofRaulRiveros) MAFiscapableofnishingfree-formsurfaces;thisabilityismainlyattributedtothechain-likecongurationofmagnetictools,sometimesreferredtoastheexiblebrush.Thisexiblebrushiscapableofconformingtothetargetsurfaceshapebybreakageandrecongurationofthemagneticparticlechainsofwhichitiscomposed.Asaresult,MAFofcurvedsurfacesispossible.ExternalnishingofrollerelementshasbeenstudiedwithMAFinrotatingtubeandstationarymagnet-typecongurationswithsignicant 40

PAGE 41

reductionsinsurfaceroughness[ 76 78 ].Effortshavealsobeenaimedatnishingdiesandmoldsandotherthree-dimensionallycurvedsurfaces[ 79 80 ].Figure 1-13 illustratesanotheradvantageofMAF'sexiblebrush.AcomplexworkpiecesuchastheoneshowninFigure 1-13 oraharddiskaccessarm[ 81 ]maybenishedbyintroducingitstargetsurfacesintoamassofmagnetizedtools.Duetothenatureofthemagnetictoolchainconguration,normalforceisexertedonallhorizontalworkpiecesurfacesincontactwiththemagnetictools.Theintroductionofmechanicalvibrationsintotheworkpiece,themagnets,orboth,wouldencouragemagnetictoolchainbreakageandreconguration,creatinghorizontalnormalforceswhichwouldinturnnishtheverticalworkpiecesurfaces. Figure1-13. Schematicdepictingtheself-recongurationofaexiblebrusharoundacomplexworkpieceA)before,B)during,andC)aftercompletecontact InMAF,themagneticeldintensityanddistributioncontrolnishingforcesandpressure.Thisisaparticularadvantageoverconventionalsurfacenishingprocessesinwhichnishingforcesmaybedifculttocontrol.Forexample,inabasicpadpolishingprocess,aatorcurvedworkpieceisnishedwithasimilarlyshapedsolidlaporpad.Abrasivesintroducedbetweenthepadandtheworkpiecearemechanicallyforcedbythepadontotheworkpiece.Additionally,parallelmotionsareintroducedbetweenthepadandworkpiecesurfaces.Thenishingforcesexertedbytheabrasiveparticlesontotheworkpiecearedependentontherelativemechanicalmotionsofthepadand 41

PAGE 42

workpiece,makingthemhighlysusceptibletoundesiredmechanicalvibrationsandmisalignments.Asaresult,undesirablyhighforcesmaydevelop,resultinginrandomworkpiecesurfacedefectslikescratchesorpits;thisoccurrenceoftenlimitsthelowestachievablesurfaceroughnessofconventionalnishingprocesses.InMAF,however,themagnetictoolsarenottypicallymechanicallyconnectedtothemachinestructure.Vibrationsandmisalignmentsexperiencedbythemagneticeldgeneratorarenotfaithfullyreproducedbythemagnetictools.Additionally,ifthemagnetictoolsareofapowder-likestructure,theyarepliable,andvariationsintheappliedmagneticeldwouldlikelycausemagneticchainstructurereconguration,whichwouldnotnecessarilytranslatetosuddenincreasesinsurfacenishingforces.Inpractice,thereareavarietyofmaterialsthatmaybemanufacturedorusedincombinationstoformmagnetictoolsforMAFprocesses.Theirselectiondependsdirectlyontheintendedapplication.AsshowninFigure 1-14 ,therearethreebasiccategoriesofmaterialsinMAF:magneticmaterials,abrasivematerials,andmagneticabrasivematerials.Themagneticforce,~F,exertedbyamagneticentityinanappliedmagneticeldisdescribedby, ~F=VHrH(1)whereVisthevolumeofmagneticmatter,isitsmagneticsusceptibility,Histhemagneticeldintensity,andrHisthegradientofthemagneticeld.InMAF,themagneticmaterialmaybeintheformofasolid(asinthecaseofamagneticpolishingpad),apowder,orauid.AbreakdownoftheclassicationofmagneticmaterialsinMAFisshowninFigure 1-14 .Magneticpowdersarecommerciallyproducedinavarietyofsizesrangingfromtensofnanometerstomillimetersandareavailableinavarietyofmaterialsasaremagneticsolids.Magneticuidsaresuspensionsofferrouspowdersinaliquidmedium.Magneto-rheologicaluid(MRF)isasuspensionof5mironparticlesinahydrocarbonoil.Magneticcompounduid(MCF)iscomposedofnmsizedferrousparticlesandmsizedironparticlessuspendedinwaterwithotheradditivesto 42

PAGE 43

controlitsviscosity.Magneticuidisasuspensionof10nmmagnetite(Fe3O4)particlesinwateroroil.Manyapplicationsoftheseuidshavebeenstudied[ 67 ];however,acompletediscussionoftheseisbeyondthescopeofthissection.Magneticabrasiveparticlesaretypicallyironpowders(occasionallycobalt)thathaveundergoneaprocesstobindabrasive(whitealumina(WA),tungstencarbide(WC),ordiamond)particlestothesurfaceoftheironparticles;thisisaccomplishedviaspecializedsinteringorplasmaspraysetups[ 82 84 ].Thesizeofmagneticabrasiveparticlesmayrangefromtensofmtoseveralhundredm.AnMAFprocessdesignerapplicationwouldselectanappropriatecombinationofmagneticandabrasivematerialsdependingonhowmuchforceisrequired,thedesiredsurfaceroughness,andotherrelevantcriteriaasmaybeconceived. Figure1-14. ClassicationofmagneticandabrasivematerialsinrelationtoMAF DependingontheparticularcongurationofanMAFprocess,themagneticeldgenerator(s)maybeeitherpermanentmagnetsorelectromagnets;abreakdownoftheclassicationofmagneticeldgeneratorsinMAFisshowninFigure 1-15 .MagneticelduxdensitiesforMAFprocessesmayrangebetweenthe0.01Tand1T.Permanentmagnetsmaybeheldstationaryorinmotion.Electromagnetsmaybe 43

PAGE 44

employedtogenerateastaticeldoratransienteld.Transienteldintensityprolestaketheformofthesuppliedcurrentwaveformtotheelectromagnets.Likepermanentmagnets,electromagnetsmaybeheldstationaryorinmotion.Additionally,magneticeldgeneratorsaredesignedinanMAFprocesstorealizeafunctionalmagneticcircuitwhereinthelinesofmagneticuxdirectlyinteractwiththemagneticmaterials[ 85 ];thegoalisalwaystoachieverelativemotionbetweentheabrasiveparticlesandthetargetworkpiecesurface. Figure1-15. ClassicationofmagneticeldgeneratorsinMAF 1.4.2FinishingofOpticalandSmallWorkpiecesDuetoabilityofMAFtonishfreeformsurfaceswithahighdegreeofnishingforcecontrol,MAFhasbeenappliedtoopticguringandpolishing.IntheMAFofglassoptics,anarrowlayerorribbonofMRFmixedwithconventionalabrasivesisappliedtoamovingsurface,typicallyawheel.AmagneticeldgeneratorpositionedoppositetotheMRFribbonmagnetizestheportionoftheribbonthatliesinthepresenceoftheeld,increasingtheMRF'sviscosity.Anoptic'stargetsurfaceisbroughtincontactoverthisregion,causinglocalizedmaterialremovalandsurfacenishing[ 86 87 ].Acomputercontrollediterativesystemwheretheopticalsurfaceisaccuratelyproledandamachinepathisgeneratedfordeterministicremovalofgureandsurfaceerrorsis 44

PAGE 45

employed[ 88 ].Thisprocessisknowntoreduceopticalsurfacegureerrorstoaslowas30nm(peaktovalley,fullaperture)and1nmrmssurfaceroughness.AnMAFprocessdesignedtogurefreeformopticalsurfacesinvolvesajetofMRFandabrasiveswhichisstiffenedbyanappliedmagneticeld.Thisjetimpingesonthefreeformopticalsurface,causinghighlylocalizedmaterialremoval.Theappliedmagneticeldforcesthejetowtostabilize,meaningthatitwillnotbreakupduetoturbulencebeforestrikingthetargetsurface.Asimilariterativemeasurement/toolpathgeneration/machiningprocessisemployedinthisprocessaswell.Usingthisprocess,Dumasetal.reducedgureerrorinafreeformopticfrom223nmto44nm(peaktovalley)[ 89 ].MAFisalsocapableofnishingthinatwaferssuchasthoseusedforopticallters.Yamaguchietal.usedarevolvingsolidmagnetictoolincombinationwithanMRF-basedabrasiveslurrytonisha60mthickquartzwafer.Asurfaceroughnessreductionof1.5nmto0.6nm(averageroughness)wasachievedwithminimaldisturbancetothethicknessofthewafer[ 90 ].Kimetal.havedevelopedacongurationofanMAFprocessfornishingofsmallgrating-typedevices.Asilicongratingwith200mwideby53.2mdeepwet-etchedtrencheswasnishedusingarotatingmagnetizedribbonofMRFandabrasives.Atrenchsidewallsurfaceroughnessreductionfrom184.6nmto18.1nm(averageroughness)wasachieved[ 91 ].TheinternalnishingoftubesbyMAFhasbeenstudiedforseveraldecadesnow;however,recenteffortsintheinternalnishingofnarrowtubessuchascapillarytubeshasbeenachievedviaMAF.Yamaguchietal.havedemonstratedtheinternalsurfacenishingof400minnerdiameterstainlesssteeltubes;asurfaceroughnessreductionfrom1.26mto0.02mwasobserved[ 92 ].Theinternalnishingof419minnerdiametertubeswithlaser-machinedaxialhelicalgrooves(fortubeexibility)yieldedasurfaceroughnessreductionfrom0.65mto0.12m(averageroughness).The 45

PAGE 46

nishingofsmallinternalsurfacesispossibleduetotheavailabilityofsmallmagneticpowdersandabrasives.Amajordifcultythatariseswhenattemptingtonishasmallinternalsurfaceistheincreaseddifcultyingeneratingasuitablemagneticeldgradient[ 92 ].AsdescribedbyEquation 1 ,amagneticforce,andconsequentlyanishingforce,willdeveloponlyifanonzerogradientexists. 1.5MEMSMicroporeMirrorChipsEzoeetal.haddevelopedmirrorchipsaspartoftheirrstMEMSmicroporeX-rayopticdesign[ 38 ].TheuseofmirrorchipshassincechangedfrombeinganintegralpartoftheMEMSmicroporeX-rayopticdesigntoatoolfortestingthemanufacturingprocessesofsubsequentMEMSmicroporeX-rayopticdesigns[ 43 ].Figure 1-16 showsthebasicdesignofamirrorchip.Theyareofsquaregeometry,andtheirthickness,th,isdeterminedbysiliconwaferfromwhichtheyaredicedoretched.Parallelarraysofcurvilinear(radiusR)orrectilinearthrough-the-thicknessmicroporesareetchedintothesubstrate.Themicroporewidth,D,spacing,s,lengthL,andthespacebetweenthelengthsofthemicropores(notshown)determinetheopenareapercentageofthemirrorchip.Figure 1-16 showsonlyverowsofmicropores;however,realdesignsmightincorporatebetween70and100rowsspaced10to50mapart. Figure1-16. Schematicofthegeometryofamirrorchip 46

PAGE 47

MirrorchipsaresimplyminiatureMEMSmicroporeX-rayoptics;asizecomparisonisshowninFigure 1-17 (A).Mirrorchipsmaybefabricatedbythesameprocessesasfull-sizedMEMSmicroporeX-rayoptics.Figure 1-17 (B)isaclose-upphotographofanactualmirrorchipfabricatedbyDRIE;thedimensionsofthisare7.57.50.3mm;themicroporesare50mwide.Theopticalfunctionofmirrorchips,astheirnamewouldsuggest,istoreectanincidentbeamofX-raysoffthemicroporesidewalls,changingitsdirection.Accordingly,thegureandnishofthemicroporesidewallsdeterminestheiropticalperformance. Figure1-17. PhotographofA)aDRIE-fabricatedMEMSmicroporeX-rayopticnexttoamirrorchipandB)aclose-upoftheamirrorchip(PhotographscourtesyofRaulRiveros) ThemicroporesidewallsurfaceofmirrorchipsfabricatedbyDRIEorLIGAisunfortunatelytoorough.Figure 1-18 showsopticalmicrographsofrepresentativemicroporesidewalls.ThesidewallshowninFigure 1-18 (A)isofaLIGA-fabricatedsidewall.Thissidewallismainlycomposedofnickel;howeverithasathingoldlayerwhichisseeninyellow.Thetextureofthesidewallischaracterizedbyaperiodicstructureinthedirectionofsynchrotronlightexposure;thisismostlikelyduetosimilarerrorspresentintheabsorptionmaskduringexposure.Small(<1m)pitsarepresentthroughoutthewall;theseareresultofthepresenceofH2gasbubblesformedduringelectroforming.Additionally,thereappeartobeoddly-shapedsmoothregionspresent 47

PAGE 48

throughouttheroughareas;thecauseofthesestructuresisunclear;however,itisspeculatedthatthevariationinthemicrostructureofthePMMAsubstratemayhavecausedthisvariation.Generally,thesidewallsurfaceroughnessofLIGA-fabricatedstructuresis10nmrms. Figure1-18. OpticalmicrographsofmicroporesidewallsfabricatedbyLIGAandDRIE ADRIE-fabricatedsiliconmirrorchipsidewallisshowninFigure 1-18 (B).ThesurfacetextureofDRIE-fabricatedstructuresidewallsisprimarilycontrolledbytheetchingmaskgeometryandtheDRIEconditionswhichincludethepowerandfrequencyofplasmageneration,gasowrates,DCbiasvoltage,etchingandpassivationcycletimes,andelectrode/wafertemperature.Whenetchingfeaturesofvaryingaspectratios,itispossiblefortheretobelessmaterialremovalfromnarrowfeaturesthanfromwiderfeatures.ThesidewallsurfacetextureshowninFigure 1-18 (B)ischaracterizedbysmoothplateausfor50mbeginningfromtheetchingdirection.Nextarelativelyconsistentisotropicsurfacetexturecreatedbyion/freeradicalreactionsduringetchingcoverstheremaining250mofthickness.Longtrenchstructuresareseenthroughoutthesidewallthickness;thesearelikelyduetogeometricerrorsintheetchingmask.Overall,thesidewallsurfaceroughnessofDRIE-fabricatedstructuresis30nmrms. 48

PAGE 49

Asurfaceroughnessof<1nmrmsisacceptableforX-raymirrors[ 37 ];thesidewallsshowninFigure 1-18 aresimplytoorough.AsdescribedinSection 1.3.3.2 ,thesidewallsurfaceroughnessofsiliconmirrorchipsmaybeimprovedbyapplyingahydrogenannealingprocess;however,annealingtimeislimitedtopreventover-modicationofthemicroporefeatures.Nosuchprocessexistsfornickelmirrorchips.Thus,aneedforamicroporesidewallsmoothingprocessexisted.NosurfacesmoothingprocesshadbeenshowntocontrollablyandselectivelynishthesidewallsurfacesofporesassmallasthoseofMEMSmicroporeX-rayoptics.AsdescribedinSection 1.4.2 ,MAFhadbeenappliedtothenishingofoptics,MEMSdevices,andinternalsurfacesofmicroscalecomponents.ThescalablenatureofMAFandpromisingpriorstudiespromptedtheinvestigationofMAFasapotentialmicroporesidewallsurfacenishingtechnique. 49

PAGE 50

CHAPTER2DEVELOPMENTOFAFINISHINGPROCESSFORMEMSMICROPOREOPTICS 2.1ProcessingPrincipleTonishthemicroporesidewallsurfacesofMEMSmicroporeX-rayopticsusingMAF,magnetictoolssmallenoughtoenterthe20mwidemicroporesarerequired.Carbonylironpowdersassmallas1.1mdiameterarecommerciallyavailable;these,however,arelikelytoformmagneticparticlechainsoflargediametersrelativetothemicroporewidth;magnetorheologicaluidwouldlikelyexhibitasimilarlimitation.Compositemagneticabrasiveparticlesarenottypicallyavailableinsinglemicronorsmallersizeseither.Magneticuid(MF),however,isasuspensionof10nmmagnetite(Fe3O4)particlesinwateroroil.MFwasthereforeconsideredapossiblemagneticmaterialfornishingMEMSmicroporeX-rayoptics.Amagneticuidwasrstpatentedin1965asasuspensionofsubmicron-sizedmagnetiteparticlessuspendedinrocketfuelforuseinzerogravityenvironments[ 93 ].Sincethen,magneticuidshavebeenproducedbychemicalprecipitationandotherprocesses.Commerciallyavailablemagneticuidsaresoldwithavarietyofselectableconcentrations,surfactants,anddispersivemediums.Figure 2-1 showsphotographsof200Lofwater-basedandoil-basedMFinnon-magnetizedandmagnetizedstates.Figures 2-1 (A)and 2-1 (B)showadrasticdifferenceinthesurfacetensionofthetwouids;theoil-baseduidwetsamuchlargerareaofthepetridishsurfacethandoesthewater-basedMF.Intheirmagnetizedstates,bothwater-basedandoil-baseduidsareattractedtothepermanentmagnet;however,theoil-baseduidexhibitsabarbstructure,showninFigure 2-1 (D).Thisbarbstructurenotobservedinthemagnetizedwater-basedMF(Figure 2-1 (B))becauseofthehighersurfacetensionofwaterandlowerconcentrationofmagnetiteparticles(3.6wt%inthewater-basedMF,10wt%intheoil-basedMF). 50

PAGE 51

Magneticuidsexhibitlittletonoresidualmagnetization.ThesmallmagnetiteparticlesizeallowsformagneticparticlechainrelaxationmechanismsgovernedbyNelrotationorBrownian(thermal)rotationtobreakdipole-chainspreviouslyformedbyanappliedmagneticeld[ 94 ].SlightchainformationhasbeenobservedinzeromagneticeldenvironmentsduetoVanderWallsandmagneticdipolemomentinteraction,dependingontheparticlesizeandsurfactantlayerthickness[ 95 ].Theforce,~F,exertedbymagneticuidmaybewrittenas, ~F=1 0JfVfrB(2)where0isthemagneticpermeabilityofavacuum,Jfisthemagneticpolarizationofthemagneticuid,Vfisthevolumeofmagneticparticles,andrBisthegradientofthemagneticinduction[ 85 ].Themagneticpolarization,Jf,isdependentonmagnetiteparticlesintheMFvolume.MagneticuidhasbeenusedinMAFapplicationspreviously.Aprocesscalledmagneticeld-assistednenishing(MAFF)wasrstpresentedin1981[ 96 ].Inthisprocess,arubbersheetlapisstiffenedbyanunderlyingvolumeofmagnetizedmagneticuid.Abrasivesabovetherubbersheetcontacttheworkpiecesurface,andrelativemotionisintroducedbetweenthelapandtheworkpiece.Theforceexertedbythestiffenedmagneticuidistransferredtotheworkpieceviatherubbersheetandtheabrasiveparticles,nishingtheworkpiece[ 97 98 ];thisprocesshasbeenappliedtosphericalsurfacesaswell[ 99 ].Twosimilarprocessesknownasmagneticoatpolishing(MFP)[ 100 ]andmagneticuidgrinding(MFG)[ 101 ]involveamixtureofMFandabrasiveparticles.AmagneticeldappliedbeneaththeMFandabrasivemixturedrawsthemagneticuiddownwards.TheMFbuoyancy,createdbythemagneticeld,tendstoaccumulatetheabrasiveparticlesabovetheMF.Aworkpieceplacedjustbelowtheuidsurfacecomesincontactwiththeabrasiveparticleswhichexertasmallforceontheworkpiece 51

PAGE 52

Figure2-1. Photographsof200LofA),B)water-basedandC),D),E)oil-basedmagneticuidinA),C)non-magnetizedand(b),(d),(e)magnetizedstates(PhotographscourtesyofRaulRiveros) 52

PAGE 53

surface.Whenrotated,theworkpiecesurfaceincontactwiththeabrasiveparticlesisnished[ 102 ].Thecontrolledlownishingforcesappliedbythisprocesshavealloweditssuccessfulapplicationtothenishingofceramicbearingballs,internaltubesurfaces,androllers[ 103 105 ].TonishthesidewallsofMEMSmicroporeX-rayoptics,smallabrasivesmustbeusedaswell.Commerciallyavailableabrasivesareavailableinpowders,slurries,andsolids.Fornishingmicropores,onlypowdersandslurriesaresuitable;thoughtobettercontrolabrasiveparticleagglomeration,abrasiveslurriesarepreferred.Figure 2-2 (A)isaphotographof200Lofdiamondabrasiveslurry;thisparticularslurryhas0msizeddiamondabrasiveparticlesdispersedinaliquidmediumwhichisabletomixwithbothwaterandoil.Figure 2-2 (B)shows200Lofalkalinecolloidalsilicawhichisachemicallyactiveabrasiveusedinthechemicalmechanicalpolishing(CMP)ofcommercialsiliconwafersforelectronics.Colloidalsilicahasaparticlesizeof20-60nmandiscapableofnishingsilicon;therefore,itisconsideredaviableabrasivematerialnishingMEMSmicroporeX-rayoptics.SimilartotheaforementionedMFPandMFGprocess,MFandabrasiveslurryorcolloidalsilicamaybemixedtogether,creatingasubstancewithbothmagneticandabrasiveproperties.Thisuidmixtureisherebyreferredtoasmagneticabrasiveuidmixture(MAFM).Figure 2-2 (C)and(D)shows200LofMAFM(100mL+100mLabrasiveslurry)innon-magnetizedandmagnetizedstates,respectively.MAFMappearstohavethenecessarypropertiestonishMEMSmicroporeX-rayopticsasitisamagneticandabrasiveuidwithparticlesneenoughtoenterandexitthemicroporeswithoutclogging.UsingMAFM,itwaspossibletodesignanMAFprocesstonishMEMSmicroporeX-rayoptics.TheprincipleofthisprocessisshowninFigure 2-3 (A).TheconceptinvolvestheuseofanalternatingmagneticeldwhichhasbeenusedinMAFprocessespreviously[ 106 107 ].Theopticissubmergedandxedinanadequatevolumeof 53

PAGE 54

Figure2-2. Photographsof200LofA)abrasiveslurry,B)water-basedalkalinecolloidalsilica,C)MAFM,andD)magnetizedMAFM(PhotographscourtesyofRaulRiveros) MAFM.AcontainerholdstheMAFMandoptic.Twoinward-facingelectromagnetsconnectedtoanalternatingcurrent(AC)powersupplygenerateanalternatingmagneticeld.MagneticpoletipsconcentratethemagneticeldaroundtheMAFMandoptic.ThealternatingmagneticeldisintendedtoactuatetheMAFM,causingittoowthroughtheopticsmicropores.ThisactionisillustratedschematicallyinFigure 2-3 (B)Insidethecontainerandaroundtheopticisamixtureofferrousparticlesandabrasiveparticles 54

PAGE 55

inaliquiddispersivemedium.Upontheapplicationofamagneticeld,theferrousparticlesareacceleratedinthedirectionofmagneticux.Thismovementdisplacesthesuspendedabrasiveparticlesandforcesthemontothemicroporesidewalls,abradingthem.Inthisway,nishingofthemicroporesidewallsMEMSmicroporeX-rayopticscanbeexecuted. Figure2-3. Schematicofthemirrorchipnishingsetupandtheprocessingprinciple 2.2FinishingEquipmentforMirrorChipsThemachineshowninFigure 2-4 wasbuilttorealizetheprocessingprincipleofFigure 2-3 .Sincemirrorchipswerethetestingplatform,thismachinewasdesignedtonishmirrorchipsandwillbereferredtoasthemirrorchipnishingmachine.InFigure 2-4 ,twohand-woundcoils(1.08mmdouble-coatedcopper,2060turns)having30158.1mm12L14steelcoreswith19.05mm-thickpolycarbonatebobbinwallsaremountedonlinearlyguidedstages.Thestagesareactuatedbyaleftandright-handedsquarethreadmanualleadscrewwhichclosesoropensthegapbetweenthecoils.Adigitallinearscalereadsoutthegap.Magneticpoletipsareconnectedtotheelectromagnetsandmadeof12L14steelaswell.ThepoletipdimensionsareshowninFigure 2-5 .Theyweredesignedtoconcentratethemagneticeldinanareaapproximatelythesizeofamirrorchip.Thespacebetweenthepoletipsisreferredtoasthepole-polegap. 55

PAGE 56

Figure2-4. Photographofthemirrorchipnishingmachine(PhotographcourtesyofRaulRiveros) Figure2-5. Schematicofthepoletipgeometry Toincreasetheintensityofthemagneticeldatthepole-polegap,amagneticyokewasintegratedintothemachinesdesign.ThemagneticyokesidesarelabeledinFigure 2-4 ;however,forclarity,themagneticyokecomponentsandtheelectromagnetcoresaredelineatedbyredlinesinacomputeraideddesign(CAD)modelshowninFigure 2-6 .Byplacingasetofmechanicallyconnectedferromagneticbodiesfromoneendofanelectromagnettoanother,amagneticcircuitiscompleted.Thisreducesthe 56

PAGE 57

resistancetothemagneticcircuitandincreasesthemagneticeldintensityatthepole-polegap. Figure2-6. Three-dimensionalmodelofthemirrorchipnishingmachinewithmagneticyokecomponentsdelineatedinred Tonishthemicroporesidewallsofamirrorchip,thechipismountedonaspecializedaluminumholdershowninFigure 2-7 (A).Theholderissimplytwohalveswithagroovemachinedontheinnersurfaceinwhichthemirrorchipsits.Figure 2-7 (B)isaphotographofaDRIE-fabricatedmirrorchipclampedinsidetheholderwithitsmicroporesexposed.TheholderisnotdesignedtostrictlyrestricttheowofMAFMthrougheverywherebutthemirrorchipsmicropores;itwasinsteaddesignedtosimplyholdthemirrorchipinsideaconical-nosedpolymerictesttubeasshowninFigure 2-7 (C).Ideally,theowwouldberestrictedtoonlyowthroughthemicropores;however,aholderwiththiscapabilitywouldbechallengingtomanufacture. 2.2.1MotionofMagneticAbrasiveFluidMixtureAbasicprocedureforperforminganishingexperimentonamirrorchiptakesthefollowingstepsgeneralsteps.First,amirrorchipissecuredinthemirrorchipholderofFigure 2-7 .Second,2mLofMAFMispreparedinsideatesttube.Thirdlytheholderisplacedinsidethetesttube,submergingthemirrorchipintheMAFM.Thetesttubeisthenheldbetweenthemagneticpoletips(15mmpole-polegap)ofthemirrorchip 57

PAGE 58

Figure2-7. Multipleviewsofthemirrorchipholder(PhotographscourtesyofRaulRiveros) nishingmachine.TheACpowersupplyisactivated,commencingthenishingprocess.Afteracertainlengthoftime,theACpowersupplyisswitchedoffandthemirrorchipisremovedandcleaned.Inearlytrials,noapparentchangesinthemicroporesidewallsurfacesweredetected.Thiswasthoughttobecausedbythedynamicuidmotionduringnishing.IfthecoilsarewiredinparalleltotheACpowersupply,theMAFMexhibitssymmetricmotion.Althoughthedirectiontheeldalternates,theMAFMwhichlikeMF,exhibitsnoresidualmagnetization,issimplyattractedtoanynon-zeromagneticeld.ToachievesidetosidemotionandproperlyrealizetheconceivedprincipleofFigure 2-3 ,thecoilswouldneedtobeactivatedindividually.ThustheparallelcircuitwasmodiedthroughtheadditionofopposingdiodesasshowninFigure 2-8 (A).Usingthiscircuit,eachcoilreceivesonlyahalf-wave,effectivelyswitchingtheeldsourcefromsidetoside.Themagneticeldusedinthisprocessisthereforedescribedasalternatingandswitching. 58

PAGE 59

TheelectriccircuitsimulationsoftwarePSPICE(CadenceDesignSystems,Inc.,SanJose,CA,USA)wasusedtosimulatethecurrentowthroughthecoils,theresultsareplottedinFigure 2-8 (B).Inductorsoftenexhibitsomecapacitance,thereforethenon-zerocurrentperiodsoverlapbetweencoils. Figure2-8. Aschematicoftheelectriccircuitandthesimulationresults Figure 2-9 (A)showsanexampleofsymmetricuidmotion,andFigure 2-9 (B)showsanexampleoftheasymmetricuidmotioncreatedbythemodiedelectriccircuitofFigure 2-8 (A).Whenexposedtothealternatingandswitchingmagneticeld,theMAFMinsidethetesttubeoscillatesfrompoletiptopoletip(sidetoside).Atlowfrequencies(1Hz),theMAFMrespondslinearly.However,ahigherfrequencies(6Hz)theuiddevelopsmultiplemodesofvibration,achievingsomemaximumamplitudeat22Hz.Atfrequenciesabove30Hz,theamplitudeoftheMAFMsurfacevibrationisnotperceptiblebythehumaneye.Earlynishingexperimentssuggestedthatthehighermaterialremovalrateswerepositivelycorrelatedwithhigheruidagitation;aparametricstudytoverifythisobservationhasnotyetbeenperformed. 2.2.2FinishingEquipmentCharacterizationOncebuilt,itwasofscienticinteresttoobservethethermal,electrical,andmagneticbehaviorofthemirrorchipnishingmachine.Todothis,aninfraredthermometerwasaimedatoneofthemachinesmagneticpoletips;ithadameasurementspotof 59

PAGE 60

Figure2-9. FluidbehaviorA)withoutdiodesandB)withdiodes(PhotographscourtesyofRaulRiveros) 25mmanditspositionwasxedbyatripodstand.Aninfraredthermometerwasusedsinceitwasnotexpectedtobeaffectedbythealternatingmagneticeld.TheACpowersupplyhasacurrentmeasurementreadoutwhichupdatesat1secondintervals.Themagneticuxdensity,B,wasmeasuredwithaHalleffectsensorprobe(1.78mmsensingarea)andcontrollerasystemcalledaGaussmeter.TheHalleffectsensorwasclampedtoathree-axisstageandpositionedatthecenterofthepole-polegap,coincidentwiththecoilaxis.AphotographoftheGaussmeter,infraredthermometer,andmirrorchipnishingmachineisshowninFigure 2-10 Figure2-10. Poletiptemperatureandmagneticuxdensitymeasurementsetup(PhotographcourtesyofRaulRiveros) 60

PAGE 61

Thepowersupplywasactivatedandrunfortwohours.Readingsfromallthreeinstrumentsweretakenfromtherstsecondofoperationandevery5minutesthereafter.Themachineallowedtocooldownovernight.Theprocedurewasrepeatedtwomoretimesandthevaluesfromeachreadingfromeachinstrumentwereaveraged.Figure 2-11 showstheaveragedreadingsfromtheinfraredthermometerandtheACcurrentreadout.Thesurfacetemperatureofthepoletiprosefrom25to40C.Thecurrentdroppedfrom1.00to0.96A.Thedropincurrentowiscommonlyduetoanincreaseincoilwireresistancecausedbyariseinthetemperatureofthecoil. Figure2-11. Changesinpoletiptemperatureandcurrentwithtime Themagneticuxdidnotvaryenoughtoregisteradifference;accordingtheGaussmeter(0.01mTresolution),themagneticuxdensitywasconstantat67.2mTasshowninFigure 2-12 .Althoughthemagneticuxdensitydidnotchangesignicantly,theriseinthepoletipsurfacetemperaturecouldaffectthenishingprocesscharacteristics.Duringanishingtrial,heatfromthepoletipswillconductthroughthetesttubeandintotheMAFM.ThetemperatureriseinMAFMhasnotbeenmeasured;however,itdoesnotappeartoriseassharplyasthatofthepoletips.ThisriseinMAFMtemperaturehasbeenobservedtoencouragetheevaporationofvolatileuidsintheMAFM;evaporatedvaporsoftenwillcondenseontheinnertesttubesurface.This 61

PAGE 62

evaporationofuidscorrespondstoavisibleseparationoftheMAFMcomponents.ThemaximumnishingtimeislimitedtotwohoursasaresultofmachinetemperatureriseandMAFMdegradation. Figure2-12. Changeinmagneticuxdensitywithtime 2.3FinishingEquipmentforFourInchOpticsRecentadvancementsintheproductionMEMSmicroporeX-rayoptics[ 59 ]havemotivatedtheproductionofamachinecapableofapplyingthisMAFprocesstofull-sizedMEMSmicroporeX-rayoptics.Thenewmachinewouldneedtoreproducethenishingprocessprincipleofthemirrorchipnishingmachine;however,itshouldreproduceitoveralargerarea.Additionally,becauseofthelargeareaofmicroporesonafull-sizedMEMSmicroporeX-rayoptic,themachinewouldneedtorunforamuchlongertime.Figure 2-13 (A)showsanoriginalconceptforthismachine;itisherebyreferredtoasthefourinchopticnishingmachine.Twoinward-facingelectromagnetsareconnectedbyanappropriatelysizedmagneticyoke.Poletipsaremountedontheinnerbobbinwalls.Ahollowcircularopticholderispositionedbyaspeciallydesignedslidewhichallowsfortheoptictobemountedoutsideofthemachineforeaseofuse.TheopticholderisdesignedtocontainbothMAFMandoptic.Asteppermotorrotatestheopticholdersuchthatnishingoccursonallportionsofoptic. 62

PAGE 63

Figure2-13. Fourinchopticnishingsetupmachineconceptandperipheraldevices 63

PAGE 64

Across-sectionalviewofthemachineconceptisshowninFigure 2-13 (B).Inthisgure,theopticisvisibleinsidetheopticholder.Themagneticyokecomponentsarealsovisible.AsrevealedinFigure 2-11 ,themirrorchipnishingmachinedevelopsaconcerningamountofheatduringoperation;inadditiontooverheatingoftheMAFM,italsoisexpectedthatthemachinewouldoverheat,causingdamagetoitspolymericcomponentsandcoilwindings.Tomitigatetheseissues,thefourinchopticnishingmachinedesignfeaturesacirculatingwatercoolingsystem,showninFigure 2-13 (B).Coolwaterenterscoilcoresfromtheinnerfaces,thentravelsthroughthecoreandexitsintoasump.Thewaterispumpedfromthesumpintoawaterchiller,andthecyclerepeats.Thesimplisticdesignofthiswatercoolingsystemallowsforarobustsystemtobebuiltwithcommerciallyavailableparts.Itiscertainlypossibletohavemoreefcientdesigns,however,theywouldbemoredifculttomanufactureandintegrateinthedesign.Anotherheat-basedconsiderationisthephysicalseparationofthepole-tipsfromtheopticholder;heatcannotdirectlyconductintotheholder.Thecoils,whicharelargerthanthoseofthepreviousmachine,areconnectedtoanACpowersupplyusingtheelectriccircuitofFigure 2-8 (A).ThewayinwhichtheMAFMinteractswiththeopticisschematicallydepictedinFigure 2-14 .TheMAFMwillllhalfoftheopticholder.ThemagneticpoletipswillbedesignedtoconcentratethemagneticeldatthesurfaceoftheMAFM,causingthemaximumagitationofthesurface.Therefore,onlyathinstripthewidthoftheopticistargetedfornishing.Thesteppermotorrotatestheopticwhichallowsforthenishingoftheotherareasoftheoptic.TheopticholderrotationhastheaddedbenetofcontinuallyremixingtheMAFM,keepingtheabrasivesfromsettling.Theexactcommandedsteppermotorrotationschemewillbedeterminedduringinitialtesting.Themagneticpoletipdesignwillhaveagreatinuenceonthenishingcharacteristics.Figure 2-15 showsthemagneticyokeandelectromagnetcoresdelineatedbyredlines. 64

PAGE 65

Figure2-14. SchematicrepresentationofarotatingfourinchopticsubmergedinmagnetizedreciprocatingMAFM Additionally,Figure 2-15 showssomegeometricdesignconstraintsbywhichthemagneticpoletipsaretoconform.Theseincludeabrasselbowttingforthecoolingwaterentrance,apulleyforthedrivebeltconnectiontotheopticholder,anda80mmholepatternontheinnerbobbinwalls. Figure2-15. Schematicofthefourinchopticnishingmachinewithmagneticyokecomponentsdelineatedinred Figure 2-16 (A)showsaphotographofaMEMSmicroporeX-rayoptic,theannularregioncoveredbymicroporeshasawidthof30mm.Itisthoughtthatthemagneticuxdensityinthisregionshouldbeofconstantmagnitudetoachieveuniformnishing.This30mmsectionisherebyreferredtoasthenishingarea.Adesignforthemagnetic 65

PAGE 66

poletipswasconceivedbasedonthegeometricconstrainsofFigure 2-15 theneedtoconcentratethemagneticeldinthenishingarea;thisinitialdesignisshowninFigure 2-16 .Magneticeldsimulationsoftwarewasusedtodeterminethedistributionofthemagnitudeofthemagneticuxinthenishingarea.Sincethepoletipshavetwo-foldsymmetryandthemachineissymmetricaboutacentralplaneperpendiculartothecoilaxis,aneighthofthemagneticmachinecomponentscouldbemodeled,omittingthemagneticyoketoeconomizecomputermemory.Figure 2-16 (C)showsthemodelusedtoperformthesimulation.Theoptictobenishedisplaced15mmfromthepoletipduetogeometricdesignrestrictions.Thecoilissuppliedwith1AofDCcurrent.ACwasnotusedbecausethefocusofthesimulationisnottoextractexactmagnitudevalues,buttodeterminethemagneticelddistributioninthenishingarea.Sincethepoletipsarenottodesignedtobemagneticallysaturated,themagneticelddistributioninthenishingareaisexpectedtobesimilarlyshapedwithinatherangeofmagneticuxdensitiesapplicabletothisMAFprocess;thisrangeisbetween20mTand100mTandisdeterminedbythesaturationuxdensitiesoftheMFusedintheMAFM(10mT).Figure 2-17 (A)showsaviewoftheinitialpole-tipdesignandashadedmodelofthemagneticuxdensitydistributiononthesurfaceofthepole-tipsimulationmodel.Figure 2-18 (A)showsthemagnitudeofthemagneticuxdensityalonga100mmlinebeginningattheorigin(indicatedinFigure 2-16 (C))whichiscoincidentwiththenishingarea.Figure 2-18 (B)showsthemagnitudeofthemagneticuxdensityinthenishingarea.Theinitialpoletipdesignshowsanarch-likedistributioninthenishingarea.Itisdesiredtohaveaatdistributioninthenishingarea;therefore,aseriesofgeometricmodicationsareappliedthepoletips.Figure 2-17 (BE)illustratesthesemodications.Theanglemodicationistherecessionoftheinnercornersofthepoletipclosesttotheoptic;itisincludedtoequalizetheendpointsoftheuxdistributioninthenishingarea.Thecurvaturemodicationisacurvatureaddedtothefaceofthepoletipclosesttotheoptic;itismeanttocounterthearch-likeshapeoftheuxdistributioninthe 66

PAGE 67

Figure2-16. MagneticeldsimulationsetupinvolvingtheA)areaofmicropores,B)poletipdesign,andC)the1/8-geometryFEMmodel(PhotographcourtesyofRaulRiveros) nishingarea.Theradiusmodicationreducesapparenttiltofthedistributionleftafterthecurvaturemodication.Finally,anedgeroundisaddedtoalledgestosimulaterealisticmanufacturingdetails.ThesimulatedmagnitudeofthemagneticuxdensitydistributioninthenishingareaisplottedinFigure 2-18 (B).Thepeak-to-valleyvaluesofthedistributioninthenishingareaversusgeometricmodicationisplottedinFigure 2-18 (C);thenaldeviation,aftertheedgeroundmodicationis<1mT,whichisconsideredacceptable. 67

PAGE 68

Figure2-17. Variousgeometricmodicationsappliedtothepoletip 68

PAGE 69

Figure2-18. PlotscomparingthesimulatedmagneticuxdensityintheA)horizontalradialdirectionandintheB)nishingareaandtheC)variationinthenishingareaversusmodication 69

PAGE 70

Havingattainedthisnaldesign,thepoletipsweremanufacturedoutof1000-seriescarbonsteel.Figure 2-19 showsacomparisonbetweenaCADmodelandaphotographofoneoftherealpoletips.Figure 2-20 showsaCADmodelandaphotographofthenalpoletipdesignassembledintothefourinchopticnishingmachine.ThemagneticelddistributioninthenishingareahasnotyetbeenmeasuredasitrequiresaspeciallydesignedholderfortheHalleffectsensorprobe;effortsarecurrentlyunderwaytorealizesuchasystem. Figure2-19. ACADmodelandphotographofthenalpoletipdesign(PhotographcourtesyofRaulRiveros) Figure2-20. ACADmodelandphotographofthenalpoletipsassembledwiththemachine(PhotographcourtesyofRaulRiveros) 70

PAGE 71

CHAPTER3FINISHINGCHARACTERISTICS 3.1EvaluationMethodsforMicroporeOptics 3.1.1ObservationofX-rayReectionCharacteristicsTypically,X-rayopticsareevaluatedbasedontheirabilitytoresolveobjectsinview.ThesurfacequalityofX-raymirrorsmaybecharacterizedbyobservingtheirspecularreectionproperties.ThisissimilartothewayinwhichtheX-rayreectionperformanceofmirrorchipsischaracterized.Essentially,themirrorchipmicroporesidewallsareexposedtosoftX-rayradiationinamodiedX-rayreectometry(XRR)setup.AcollimatedbeamofX-raysisaimedatamirrorchipandwillgenerallyexposethetargetareaasindicatedinFigure 3-1 (A).AsdepictedinFigure 3-1 (B),theopticisrotated,byanangle,untiltheX-raysaregeometricallyobscured;thisangleisreferredtoasthecollimationangle.AnX-raydetectorplacedbehindthemirrorchipmeasurestheintensityofthetransmittedX-rays.Figure 3-1 (C)portraystherelationshipbetweenthemeasuredX-rayintensityandtheangulardisplacement,ofthemirrorchip.IfthemicroporesidewallsdonotreectanyX-rays,themeasuredintensityprolewillbeastraightline,asindicatedbythecollimatedproleinFigure 3-1 (C).Ifreectionoffthemicroporesidewallsdoesoccur,anexcessintensitywillbemeasuredasshownbythemeasuredproleinFigure 3-1 (C).Thisexcessinintensityduetoreectionischaracteristicofthegureandnishofthemicroporesidewallsurfaces.Theintensity,I,ofperfectlycollimatedX-raystransmittedthroughamirrorchipwithat-guredmicroporesidewallsmaybedescribedbythefollowingequation, I=R(,)exp")]TJ /F11 11.955 Tf 11.29 16.86 Td[(4 nsin2#(3)whereistheincidentX-raywavelength,istheincidentangletothemicroporesidewalls,R(,)describesthereectionproleoffperfectlysmoothandatmicropore 71

PAGE 72

sidewalls,nistheindexofrefractionfromEquation 1 ,andisthestandarddeviationofthemirrorsurfaceheight(assumingaGaussianheightdistribution).TheexponentisknownasthestaticDebye-Wallerfactorandquantiesthelossinreectedintensity,orreectanceduetointerference[ 108 ].Ifatmicroporesidewallsareagoodassumption,thesurfaceroughness,,ofthesidewallsmaybeinferredbyttingEquation 3 tothemeasuredprole. Figure3-1. MirrorchipX-rayreectancemeasurementsetup AschematicofthereectancemeasurementapparatusisshowninFigure 3-2 .TheX-raysaregeneratedbyanelectronbeamstrikinganaluminumanodewithaDCvoltageof4kV.X-raysareemittedfromexcitedAluminumatoms1mbelowthesurfaceinalldirections.SomeX-rayswilltravelthrougha2.5mmapertureinanopticallythick(3mm)aluminumstop.SomeofthoseX-rayswillcontinuethroughasimilaraluminumstopwhichis1.75mfromthealuminumanode.ThedispersionangleoftheX-raysbeyondthesecondstopis0.13.Thebeamthenstrikesthemirrorchipwhichismountedonasingleaxisgoniometerwhichtiltstheopticinthedirectionduringreectancemeasurements.AnX-raydetectorisplaced3mmbehindthemirrorchip;thedetectorconsistsofa0.3mm-thickberylliumwindowandareversebiasvoltagesiliconPINdiode(p-typesemiconductor,intrinsicsemiconductor,andn-typesemiconductorlayers)whichcreatesanavalancheofelectrons(currentpulseproportionaltoincidentphotonenergy)whenanX-rayphotonstrikeschargecarrier-depletedintrinsiclayer.Thecomponentsarehousedinavacuumchamber(0.2 72

PAGE 73

Pa)tominimizeX-rayabsorptionbyair.ThebeamdispersionangleisconsideredintheR(,)termofEquation 3 whenttingintensityprolesattainedwiththissetup;additionally,aspecicwavelength,(inthisresearch,AlK,1.49keV)isselectedforconsiderationbythereadinganappropriatebinofthedataacquisitionsystem[ 38 ]. Figure3-2. SchematicofX-rayreectancemeasurementapparatus 3.1.2SurfaceProleCharacterization 3.1.2.1SurfaceprolingmethodsTheX-rayreectancetestingmethoddescribedaboveisanondestructivemicroporesidewallcharacterizationtechniquewhichcancollectivelyassessthemicroporesidewallsurfacecondition.Wheneffortsarefocusedonthereductionofthemicroporesidewallsurfaceroughnesshowever,itbecomesnecessarytodirectlyprolethesidewallsurfaces.Itisunfeasibletoinsertasurfaceprolingprobeintoasinglemicropore,astheirwidthsoftenlieinthetensofmicrometers.Themonolithicgeometryandnarrowmicroporesofmirrorchipsnecessitatetheirdestructionforthesakeofdirectsurfaceprolometry.Typically,themirrorchipsaremanuallysnappedinhalf,exposingthemicroporesidewalls.Inthisresearch,twoinstrumentswereemployedtoprolemicroporesidewallsurfaces:ascanningwhitelightinterferometer(SWLI)andanatomicforcemicroscope(AFM).Bothinstrumentsallowforthegenerationofthree-dimensionalimagesofthemeasuredsurfaces.Figure 3-3 (A)showsthebasicdesignofaSWLI.Itsdesignissimilartothatofamonocularmicroscopewithon-axislighting(duetothebeamsplitter).Alightsourcepassesthroughalter(550nmcenterwavelength,80nmbandwidth) 73

PAGE 74

whichisreectedbyamirrorandathenabeamsplitter.Thelightthenpassesthroughaninterferometricobjective,reectsoffthesamplesurfacebackupwardsthroughtheinterferometricobjectiveandontoaCMOSarray.Theopticalcomponentsaremountedontoapiezoelectrictransducerstage(calledthescanningstage)whichtranslatesthefocalplanethroughthefullheightofthesurface.Duringasinglecommandedtranslation(ascan),theCMOScamerarecordsavideooftheimages.Computersoftwarethendecodestheimagesobservedandinferstheshapeofthesamplesurface.TwotypesofinterferometricobjectivesareschematicallyrepresentedinFigures 3-3 (B)and 3-3 (C).Themainfunctionoftheinterferometricobjectivesistosplitthelightintotwopaths,reectingoneoffareferencemirrorandtheotheroffthesamplesurface.Thetwolightpathsthenrecombineandinterfereifthesurfacelieswithinthedepthofeld,forminganimageknownasaninterferogramwhichischaracterizedbydarkandbrightareasknownasfringes.InamonochromaticMichelsoninterferometer,achangeinsurfaceheightequalto=4wouldcauseachangeinbrightnessfromlighttodark.Ifthesurfacefeatureheightrangeis>=4,itbecomesbedifculttodetermineheightchangefromastillinterferogram.Theuseofwhitelightorarangeofwavelengthsreducesthisambiguity. Figure3-3. SchematicsoftheA)SWLIopticalpath,B)Mirauobjective,andC)Michelsonobjective 74

PAGE 75

Asthescanningstageistranslated,thefocalplaneispositionedabovethesurfaceandscanneddownwardsthroughandbeyondthesamplesurface.Throughoutthescan,thecontrastbetweenthefringesincreasesanddecreasesasschematicallyplottedinFigure 3-4 .Usingmethodsthatarebeyondthescopeofthisdiscussion,thepeakofthefringecontrastcanbedetermined,allowingfortheprolingofsurfacesoflargeheightranges.Thevertical(surfacefeatureheight,Z)resolutionofaSWLIcanbe<0.1nm.Thelateral(XandY)resolutionisdeterminedbytheopticalmagnicationandtheCMOSarraypixelsize;inthisresearchthelateralresolutionisaslowas0.276m.BecausetheSWLIdecodesinterferograms,itiseasilydeceivedbythin(<10m)translucentlmsordebriswhichmaycausetheirownlightinterference;thuspropersamplepreparationisnecessarytoobtainreliableresults. Figure3-4. FringecontrastversusZ-positionoftheSWLIobjective TheAFMfunctionsbyfeelingthesamplesurfacewithanextremelysharptip.AschematicofanAFMisshowninFigure 3-5 (A).Acantileverwithasharp(2nmradius)tipisbroughtincloseproximitytothesurfaceatoms.Apiezoelectricstagesystemrastersthecantileveredtipoveranarea.Thetip-surfaceinteractionscausetheendofthecantilevertodeect.Afour-quadrantphotodiodemeasuresthepositionofaninfraredlaserreectedoffthetopofthedeectedcantileversurfacebycomparingthelightintensitiesinthefourquadrants.Theheightdeectionofthecantilevertipisgeometricallyrelatedtothemeasuredlaserposition.Apositionfeedbacksystem 75

PAGE 76

indicatesthelateralpositionofthecantilevertip;thusathree-dimensionalimageofthesurfacemaybegenerated.TheforcesthatdeecttheAFMcantilevertiparedescribedbytheLennard-Jonespotential(showninFigure 3-5 (B))whichdescribestheattractionandrepulsionphenomenonthatoccurswhentwoatomsarebroughtincloseproximitywitheachother.TheZresolutionofanAFMcanbeaslowas0.01nm(noisegoverned),anditslateralresolutionisdeterminedbythepositionsamplingrateandrasterspeed,whichcanbeaslow1nmorless. Figure3-5. SchematicofanatomicforcemicroscopesetupandtheLennard-Jonespotential 3.1.2.2SurfaceproleanalysisMeasuredsurfaceprolesmustbeprocessedtoobtainausefulrepresentationofthesurfacetexture.Figure 3-6 representsthewayinwhichasurfaceproleisgenerallyprocessed.Theinitialsurfaceheightdataisattainedbythesurfaceproler.Usingsoftware,thenominalformofthesurfaceistandsubtracted.Thesurfaceproledataisthenlteredtoisolateeithertheso-calledroughnessorwavinessproles.Mostcommonly,theroughnessproleisusedtodescribethesurfacetexture.TheproleltersusedintheisolationoftheroughnessprolefromthesurfaceprolearedescribedbyISOstandardsISO16610-21andISO13565-1.Theyarenormallyappliedtoalineofsurfaceheightpoints,andtheirwavelengthcutoffsare 76

PAGE 77

Figure3-6. Separationoftheroughnessprolefromthemeasuredsurfaceprole afunctionoftheline'slength(samplinglength)andtheapproximateheightrangeoraveragedeviationfromtheprolecenterline(iftheproleisconsiderednon-periodic).Figure 3-7 showsthetransmissioncharacteristicsofthesurfaceproleltersandtheirrespectivecutoffwavelengths,whichhaveaGaussiankernelfunction. Figure3-7. Surfaceproleltertransmissioncurves Thelteredroughnessprolemaythenbeanalyzedusinganumberofstatistical,spatial,andspectralmethods.Statisticalmeasuresofthesurfaceprolegenerallyinvolveparameterstodescribetheroughnessproleamplitudedensityfunction,whichissimplyahistogramofthesurfaceheightasshowninFigure 3-8 .TheseparametersincludethestandarddeviationRq,rangeRz,skewnessRs,andkurtosisRku.The 77

PAGE 78

spatialroughnessparametersincludetheaveragesurfaceheightfromtheprolecenterlineRaandmeanwidthoftheroughnessproleelementsRSm.Spectralmethodsforanalyzingtheroughnessproleinvolvethevisualizationoftheprole'sfrequencycontentviadigitalFouriertransforms(DFT),autocovarianceplots,andpowerspectraldensityplots(PSD).Spectralmethodsareparticularlyusefulwheninferringthesurface'slightscatteringpropertiesfromsurfaceroughnessproles. Figure3-8. Schematicofsurfaceprolestatistics 3.1.3OtherCharacterizationMethodsInadditiontothemethodsdescribedabove,thetextureofthemicroporesidewallsurfacesandthegeometryofthemirrorchipsmaybeobservedandanalyzedusingvariousimagingtechniqueslikeopticalmicroscopyorscanningsecondaryelectronmicroscopy(SEM).Thechemistryofthesurfacesmaybeanalyzedbybackscatteredelectronmicroscopy(BSE),energydispersiveX-rayspectroscopy(EDS),andphasecontrastmicroscopybyAFM. 3.2LIGA-fabricatedNickelMirrorChips 3.2.1PreliminaryStudyIntheearlystagesofthisresearch,bothLIGA-fabricatednickelmirrorchipsandDRIE-fabricatedsiliconmirrorchipswereavailable.Numeroustrialswereperformedinitiallyusingaheuristicapproachtondingasetofworkingprocessparameterswhichinclude:abrasivesize,magneticuidconcentration,abrasiveslurrytomagneticuidratio,alternatingcurrent,andnishingtime.Fromtheapparenttrendsobservedinthese 78

PAGE 79

earlytrials,processparameterswerechosenwhichappearedtoyieldmeasureablechangestothemicroporesidewallsurfaces.Thus,apreliminaryexperimentwasperformedonanickelmirrorchip.TheexperimentalconditionsforthisnishingexperimentareshowninTable 3-1 .Thenickelmirrorchiphad20mwidemicroporesandwas0.2mmthick.ThemirrorchipwasmountedasshowninFigure 2-7 .MAFMwaspreparedbymixing1mLof1.8wt%Fe3O4water-basedmagneticuidand1mLof0.5mdiamondabrasiveslurry(250nmmeanparticlesize).AnACfrequencyof25Hzwaschosen,foritcausedagreatamountofuidagitation.AnACcurrentof1Awasselectedtonotelectricallyoverloadthepowersupply.Anishingtimeof60minappearedinearlytrialstocausemeasureablechangestothemicroporesidewallsurfaces;thusitwaschosenforthisexperiment. Table3-1. Preliminarynickelmirrorchipexperimentalconditions WorkpieceNimirrorchip7.50.2mmPoresize202000mAbrasiveslurryDiamondslurry0-0.5mSuppliedamount:1mLMagneticuidWater-basedmagneticuid,1.8wt%Fe3O4Suppliedamount:1mLPole-poledistance15mmMagneticuxdensity65.5mTatnishingareaAlternatingcurrent1A,25HzPolishingtime60min Afternishing,themirrorchipwasimagedinanscanningelectronmicroscope.Figure 3-9 (A)showsanSEMimageofamicroporesidewallsurfacewastakenat30fromanickelmirrorchipfacenormal.Itappearsthatathinlmispresentonthesidewallsurfaceandappearstobepeeling.ThisisthoughttobethegoldlayerwhichhasbeendamagedbythegrindingphaseoftheLIGAprocess.Figure 3-9 (B)showsasimilarimageofthenishedmirrorchip.Thereappearstobepartialremovalgoldsurfacelayerandsmoothingofthesidewallsurface. 79

PAGE 80

Figure3-9. SEMmicrographsofthemicroporesidewallsat60fromthemirrorchipface Amicroporesidewallwasextractedfromanunnishedmirrorchipandanotherwasextractedfromanishingmirrorchip;bothweremeasuredwiththeSWLI.Figure 3-10 showsthree-dimensionalsurfacesobtainedbytheSWLIoftheunnishedandnishedmicroporesidewalls;ahigh-passGaussiansplinelterwithan8mcutoffwasappliedtothesurfaces.Theunnishedmicroporesidewallsurfaceroughnesswas13.2nmSq(standarddeviationofthetwo-dimensionalarrayofsurfaceheights).Thenishedmicroporesidewallhadaroughnessof6.7nmSq.ThisdifferenceinmeasuredroughnessvaluesandthedifferencesinsidewallsurfaceconditionsshowninFigure 3-9 motivatedtheexecutionofacarefullydesignedmirrorchipnishingexperimentdescribedinthefollowingsection. 3.2.2FinishingCharacteristicsAnishingexperimentwasperformedonasinglenickelmirrorchip.Sinceitisnotfeasibletocompareasinglemicroporesidewallsurfacebeforeandafternishing,thisexperimentwasdesignedtonishingonlyhalfofthemicroporesontheoptic,allowingforacomparisonbetweenunnishedandnishedmicroporesfromthesamemirrorchip.Toachievethis,themirrorchipwaspreparedasshowninFigure 3-11 .An 80

PAGE 81

Figure3-10. SWLImeasurementsofanunnishedandnishedsidewall appropriatelysizedpieceofadhesivepolyimidetapewasstucktothemirrorchipsfrontandbacksurfaces,obscuringasinglecolumnofmicropores.ThemirrorchipwasthennishedusingtheconditionsshowninTable 3-2 .AslightdifferenceintheselectedACfrequencyyieldedaslightlydifferentmagneticuxdensityfromtheconditionsusedinthepreliminaryexperiment. Figure3-11. Photographofamirrorchippreparedforthenishingexperiment(PhotographcourtesyofRaulRiveros) Oncethemirrorchipwasnished,theadhesivetapewasremovedandthechipwascleanedbyaseriesofultrasonicbathsinpurewater,acetone,andethanol.AnX-rayreectancetestwasperformedonboththeunnishedandnishedcolumnsofmicropores.TheresultsofthereectancetestsareshowninFigure 3-12 .Theoreticallinesarettothedatasets.AsshowninFigure 3-12 ,thereectancedataforthe 81

PAGE 82

Table3-2. LIGA-fabricatedmirrorchipexperimentalconditions WorkpieceNimirrorchip7.50.2mmPoresize202000mAbrasiveslurryDiamondslurry0.5mSuppliedamount:1mLMagneticuidWaterbasedmagneticuidSuppliedamount:1mLPole-Poledistance15mmMagneticuxdensity67.2mTatpolishingareaAlternatingcurrent1A,22HzPolishingtime60min unnishedcolumnofmicroporescoincideswitha1.8nmRqsidewallroughnesswhereasthenishedcolumnalignswellwitha0.7nmRqsidewallroughness.BecauseofvariationinthebrightnessofX-raygenerator,a2nmRqerrorispossibleinthesidewallroughnessestimation;however,themeasureddatasuggeststhatsomeimprovementinthesidewallX-rayreectionpropertiesoccurred. Figure3-12. X-rayreectancetestresults Next,thechipwassectionedalongcutlinesasindicatedinFigure 3-11 .Thisallowedmanymicroporesidewallstobeextractedandplacedonaspecimenholderfordirectsurfaceproling.Inall,52microporeunnishedsidewallsweresuccessfullyextracted;thus52nishedsidewallswerealsoextracted.Thesurfaceroughnessina3030mareawasobservedbySWLIoneachsidewall.Becausethesidewalls 82

PAGE 83

weredeformedduringsectioningandextraction;a4th-orderpolynomialsurfacewastandsubtractedfromtheareatoremovethedeformation.Toavoiddataspikes,a30msamplinglengthprolelineinthedirectionofsynchrotronlightexposurewasusedtocalculatetheRqroughness.Figure 3-13 showsthecalculatedroughnessvaluesforallSWLImeasurements.Thenishedsidewalls(8.73.2nmRq)wereunexpectedlyrougherthantheunnishedsidewalls(2.20.4nmRq),contradictingtheX-rayreectancetestresults. Figure3-13. SurfaceroughnessmeasuredbySWLIof52unnishedand52nishedsidewalls Tofurtherinvestigatetheprocesseffects,arepresentativeunnishedmicroporesidewallandanishedmicroporesidewallwasselectedforhighresolutionimaging.Thesesidewallshadatypicalsurfacetexturefortheirgroups.Figure 3-14 showsoptical,SEM,andBSEmicrographsoftheselectedunnishedandnishedsidewalls.Thedottedrectanglesindicatethesameareaoneachsurface.Theopticalmicrographsindicatethatpartialremovalofthegoldlayeroccurredbythedifferenceincolor;thisoccurrenceiscorroboratedbytheincreasedquantityofdarkregionsintheBSEimageofthenishedsidewallwherethedarkregionscorrespondatomically 83

PAGE 84

lighternickelexposedbypartialremovalofthegoldlayer.Andgenerally,thenishedsidewallsappeartohavearoughertexture.Figure 3-15 aretheresultsfromEDSmeasurementstakeninthelightanddarkregionsasindicatedFigure 3-14 (Aii).Thesemeasurementsconrmthepresenceongoldonbothregions;thedifferenceincolorintheSEMimagesislikelyduetodifferencesinthesampletopography. Figure3-14. OpticalandSEMmicrographsofunnishedandnishedsidewalls Figure 3-16 showshighermagnicationimagesintheregionsindicatedbythesolidwhitesquaresinFigure 3-14 (ii).ComparingFigures 3-16 (Ai)and 3-16 (Bi),thenishingprocessclearlycauseddamageandremovalofthegoldlayer,whichappearsasthelightergraycolor.HighermagnicationimageswithintheregionsareshowninFigure 3-16 (ii).Theseshowlittledamagetothegoldlayeredgesontheunnishedsurface,butthegoldlayeredgesareslightlypeeledandblurredonthenishedsurface.Theseobservationssuggestthatthenishingprocessindeedcreatessomenishingpressureonthemicroporesidewallsurfaces.Itshouldbenotedthatthepit-likefeatures 84

PAGE 85

Figure3-15. EDSspectraofdarkandlightareasasindicatedinFigure 3-14 (Aii) seeninFigure 3-16 areduetoH2gasbubbleformationduringtheelectroformingstepoftheLIGAprocesss.Itwasthentheorizedthatthenishingprocessexhibitedmicroporesidewallsurfacenishingonaverysmallscale,removingsmallfeatures.Perhaps,theslightpressureappliedbytheMAFprocesssimultaneouslycausedsmoothingandpeelingofthegoldlayer.Totestthis,ThesurfacesweremeasuredwithanAFM;themeasuredsurfacesareshowninFigure 3-17 .Inthe3030mAFMmeasurementsthenishedsurfacecertainlyexhibitsahigherheightrange;however,thenetextureofthelargerbumpfeaturesappearstobesmootheronthenishedsurface.55mmeasurementsrevealasmoothersurfacetextureonthenishedsurface.Themeasuredroughnessvalueofthesidewallsinthe55mscanswere9.32.5nmRqfortheunnishedsurfaceand5.70.7nmRqofthenishedsurface.ItthereforeappearsthatthenishedsurfaceweresmoothedbytheMAFprocess,however,lowerlateralresolutionsurface 85

PAGE 86

Figure3-16. Higher-magnicationSEMmicrographsofnishedandunnishedsidewalls measurementsappearedasiftheroughnesshadincreasedduetopeelingofthegoldlayer.ThisimprovementinsmallscaleroughnessisconsistentwiththeimprovementinX-rayreectanceofthenishedmicroporesidewalls.Figure 3-18 showsAFMscansofrandomlyscatteredbumpfeaturesobservedonthemicroporesidewalls.ThesmallscalesmoothingobservedinFigure 3-17 isalsoobservedonthesebumpfeatures.Theunnishedbumphasasharpshape;whereasthenishedbumpappearstohavebeenerodedbythenishingprocess. 3.3DRIE-fabricatedSiliconMirrorChips 3.3.1PreliminaryStudyGenerally,inxedandlooseabrasiveprocesses,theuseoflargeabrasivesresultsinhighermaterialremovalduetolargerabrasivecuttingedges.Highernishingforces 86

PAGE 87

Figure3-17. AFMimagesofunnishedandnishedsidewalls Figure3-18. AFMimagesofbump-likefeaturesonasidewall 87

PAGE 88

Table3-3. PreliminaryDRIE-fabricatedmirrorchipexperimentalconditions WorkpieceSiliconmirrorchip7.57.50.3mmPoresize201600300mAbrasiveslurryDiamondslurry6m-mean,water-basedSuppliedamount:1mLMagneticuidWater-basedmagneticuidSuppliedamount:1mLPole-poledistance15mmMagneticuxdensity65.5mTatnishingareaAlternatingcurrent1A,25HzPolishingtime60min duetofewercuttingedgesincontactwiththeworksurfaceresultaswell.Thus,inanattempttoachievealargeamountofmaterialremoval,anexperimentwascarriedoutusingtheconditionsofTable 3-3 onasiliconmirrorchip.Inthisearlynishingexperimentonasiliconmirrorchip,arelativelylargeabrasivesizeof6mwasused.Afternishing,anX-rayreectancetestwasperformedonthemirrorchip.Forreference,anX-rayreectancetestwasperformedonanunnishedmirrorchipfromthesameproductionbatch.TheresultsfromthistestareshowninFigure 3-19 .ThenishedmirrorchipexhibitedimprovedX-rayreectioncharacteristics(3.8nmRq)thantheunnishedmirrorchip(6.2nmRq).Surfaceprolemeasurementsofanishedmicroporesidewallandanunnishedsidewallfromthesameproductionbatchwerecompared;however,noclearchangewasobserved.Theresultsfromthisexperimentmotivatedtheexecutionofanexperimentinwhichthemorphologicalchangeinmicroporesidewallsurfacefeatureswascarefullyobserved;thisexperimentisdetailedinthefollowingsection. 3.3.2FinishingCharacteristicsThissectionwilldetailtheresultsofamultiphaseexperimentperformedonasiliconmirrorchipwhichhadpreviouslyundergoneahydrogenannealingtreatment;Table 3-4 showsthemirrorchipinformation.TheuseofahydrogenannealedmirrorchipwouldallowfortheobservationoftheeffectsoftheMAFprocessonsmoothsurfaces, 88

PAGE 89

Figure3-19. X-rayreectancetestresults whichcouldcontributetoanunderstandingofthenishingmechanism.Unfortunately,only1suchmirrorchipwasavailablefornishingexperiments.Thus,oneobjectiveofthisexperimentwastosimplyshowevidenceofmaterialremovalduetothenishingprocess.Morphologicalchangesinthemirrorchipgeometrywouldbeindicativeoftheprocessnishingcharacteristics. Table3-4. AnnealedDRIE-fabricatedmirrorchipgeometryandconditions Material Silicon(110) Geometry 7.57.50.3mm Poregeometry WidthD20m(seeFigure 1-16 ) SpaceS30m CurvatureR150mm LengthL1600m Thicknessth300m Hydrogenannealing Pressure6.66kPa H2gasowrate1slm Temperature1300C Annealingtime30min TheannealedmirrorchipwaspreparedsimilarlytothechipnishedinSection 3.2.2 ;aphotographofthechipwithadhesivepolyimidetapecoveringasinglecolumnofmicroporesisshowninFigure 3-20 .Sinceonly1workpiecewasavailable,anditwasdesiredtodemonstrateevidenceofmaterialremoval,a5-phasenishingexperimentwasoutlined.Table 3-5 showsthenishingconditions.Threedifferentabrasiveswere 89

PAGE 90

usedastheywereobservedtoyieldmeasureablechangestothemicroporesidewallssurfacesinpriorexperiments[ 109 ].Phases1use0.5mdiamondabrasiveslurry,whichhasbeenusedinstudieswithnickelmirrorchips.Atotalof6hrofnishingwascarriedoutwiththisabrasivesuchthatarelativelylargeamountofmaterialwouldberemoved.Phase4uses50nmpolycrystallinewater-baseddiamondslurrywhichwasexpectedtofurthersmooththemicropores.Phase5involvestheuseofwater-basedalkalinecolloidalsilicawhichisachemicallyactiveabrasiveslurryusedinthenalnishingofcommercialsiliconwafers.Alkalinecolloidalsilicaformssiloxanebondswiththesiliconsurfaceatomswhichhavehigherbondstrengththandothesiliconbulkatoms.Anexternalmechanicalforcebasicallyripsthesilicaparticlesandanyattachedsiliconsurfaceatomsalongwithitoffthesurface;thisisthemechanismofthechemicalmechanicalpolishing(CMP)process[ 81 ].ItisthoughtthatmagnetiteparticlesinMAFMpreparedwithalkalinecolloidalsilicawouldexerttheaforementionedmechanicalforce. Figure3-20. PhotographoftheannealedDRIE-fabricatedmirrorchippreparedforanishingphase(PhotographcourtesyofRaulRiveros) Thenishingphaseswereperformedonthemirrorchip.Inbetweeneachphasethechipwascleanedbyultrasonicationandnewtapewasusedtocovertheunnishedmicroporesbetweennishingphases.Afterbeforeanynishingprocesswasapplied,theannealedmirrorchipwasimagedusingeldemissionscanningelectronmicroscopy(FESEM).Afterthenishingphases,thechipwasagainimagedwithFESEM,allowingforthecomparisonofindividualfeaturesbeforeandafternishing.Figure 3-21 shows 90

PAGE 91

Table3-5. AnnealedDRIE-fabricatedmirrorchipnishingconditions FinishingphasePhase1Phase2Phase3Phase4Phase5MagneticuidWater-based,1.8wt%10nmFe3O4particles:1mLAbrasiveslurryTypePCD*PCDPCDPCDAlkalineCS**Size0.5m0.5m0.5m50nm20nmSuppliedamount1mLFinishingtime120minPole-poledistance15mmAlternatingcurrent22Hz,1AMagneticuxdensity67.2mTatnishingarea *Polycrystallinediamond,**Colloidalsilica suchcomparisons.Figures 3-21 (A)and 3-21 (B)showmicroporesimagedfrom30fromthemirrorchipfacenormal.Astreakedtextureonthesurfaceofthemirrorchipfaceappearstohavebeenremovedbythenishingprocess.Figures 3-21 (C)and 3-21 (D)showthesamemicroporeendbeforeandafternishing;structuresontheinnermicroporesidewallarenolongerpresentafternishing.Figure 3-21 (E)showsabridgedefectleftonthebacksideofthemirrorchipduetoincompleteetching.Figure 3-21 (f)showsthenishedbridgedefect;itappearssmootherandsmallbump-likestructuresontheupperandloweredgesofthefeaturehavebeensmoothed.Figures 3-21 (g)and 3-21 (h)compareanareaontheundersideofthebridgedefect;roughsurfacefeaturesinthisareahavebeenremoved.TheFESEMimagesofFigure 3-21 showevidenceofmaterialremovalbytheMAFnishingprocess.Themirrorchipwasbrokentoexposethemicroporesidewalls.Anishedmicroporesidewallandanunnishedmicroporesidewallwereselectedfordirectedsurfaceprole.Forcomparison,anun-annealedandunnishedDRIE-fabricatedmicroporesidewallwasalsoextracted.Figure 3-22 showsSWLImeasurementsperformedonthethreesamplesurfaces.Adifferenceintextureisapparent;specically,lowwavelengthcomponentsofthesurfaceasperitieshavebeenremovedbythehydrogenannealingprocessandfurtherremovedbythenishingprocess.ThelargescalewavinessofcreatedduringDRIEappearstoremain. 91

PAGE 92

Figure3-21. FESEMmicrographsofmicroporefeaturesbeforeandafternishing Figure3-22. SWLImeasurementsoftheofrepresentativesurfaceproles 92

PAGE 93

Ahigh-passGaussianlterwitha14mcutoffwasappliedtothedatasetsofFigure 3-22 .Ten100msamplinglengthprolelinesweretakenintheoff-axisdirectionasindicatedinFigure 3-22 .SeveralroughnessparameterswerecalculatedfromtheseproleslinesandplottedinFigure 3-23 .Hydrogenannealingappearstohavegreatlyreducedboththepeaktovalley,Rz,androot-mean-square,Rq,roughnessoftheDRIEsurface.ThenishingprocessappearstohaveonlyslightlyreducedthesurfacesRz,anditreducedtheRqvalueby1nm.Themeanwidthoftheprolespacing(meanspacingofprolezeroheight-crossingpoints),RSmappearstoincreasewithbothsurfacetreatments.Theskewnessofthesurfaceamplitudedensityfunction,Rsk,appearstotrendtowardsmorenegativevalues.ThehigherRSmandlowerRskvaluesofthenishedsurfacesuggestthattheMAFprocesspreferentiallyremovesmaterialfromthepeaksoflowwavelengthasperities;itmaybereasonedthattheprocessremovesmaterialbymechanicalactionasopposedtochemicalaction,sincematerialunevenlyremovedfromthesurface.TobettervisualizetheeffectsofhydrogenannealingandMAFonthemicroporesidewalls,thesurfaceswerescannedbyAFM.Figure 3-24 showsthemeasuredsurfaces.The55mscansshowninFigure 3-24 (A).ThehydrogenannealedsurfaceappearstohaveanisotropictexturesimilartotheDRIEsurfacewithgreatlyreducedfeatureheight;thisisnotunexpectedbecausehydrogenannealingalonedoesnotintroduceanydirectionalityintoannealedsurfaces.ThenishedsurfaceofFigure 3-24 (Aii)suggeststhattheMAFprocessremovedtheisotropichydrogenannealedsurfacetexture.Asetof11mscansofthesurfacesisshowninFigure 3-24 (b).Thesmallerscansizeallowsforaclearvisualizationofthesurfaces.Thehydrogenannealedsurfacetextureexhibitsanon-directionalsurfacetexturewithperiodicitysimilartothatoftheDRIEsurface.Thenishedsurfaceappearstohaveadirectionalsurfacetexture;ithasnotbeenveriedwhetherthesemarksareduetoabrasivescratchingoratomicsteps. 93

PAGE 94

Figure3-23. SurfaceroughnesschangeswithsurfaceconditionbySWLIusinga100msamplinglength 94

PAGE 95

Figure3-24. AFMimagesofrepresentativemicroporesidewallsafterDRIE,hydrogenannealing,andhydrogenannealing+MAF Figure 3-25 showsthemeasuredpeaktovalleyheightdifferenceinthefullscanarea,SzandanRqcalculatedfromtenproleslinesaslongasthescansidelength.ThedifferencesinSzandRqvalueswarrantedtheuseoflogarithmicverticalscalesinFigure 3-25 .Surfaceroughnessreductionswereclearlyobtainedbybothsurfacetreatments.ThemeasuredRqvaluesatthe5msamplinglengthare35.45,1.72,and0.18nmfortheDRIE,hydrogenannealed,andhydrogenannealed+MAFsurface,respectively.Thesurfaceroughnessachievedbythenishingprocessof0.18nmRqiswellwithinthesurfaceroughnessrequirementsofmodernX-raytelescopes.MAFprocessedmicroporesidewalls,althoughsmoothonasmallscale,arenotsuitableasX-raymirrorsduetotheirgureandwaviness.Effortsareunderwaytocorrectthesehowever[ 59 ].Afterthecompletionofthisexperiment,anotherhydrogenannealedsiliconmirrorchipwasobtained.TochecktherepeatabilityoftheMAFprocess,the5-phasenishingprocedurewasappliedtothisnewchip.Nomicroporeswereobscuredinthis 95

PAGE 96

Figure3-25. SurfaceroughnesschangeswithsurfaceconditionbyAFM experimenthowever.FESEMmeasurementsweretakenbeforeandafterthenishingprocess.Figure 3-26 showsacomparisonofthesameorsimilarfeaturesbeforeandafternishing.Figures 3-26 (i)and 3-26 (ii)showaslightlycleanersurfaceafternishing.Figure 3-26 (iii)showsnoclearlydiscernibledifferences;however,Figure 3-26 (iv)showsremovalofstructuresfromamicroporeendsidewallafternishing.Theseimagessuggestthattheeffectsofthenishingprocesswererepeatedonthesecondhydrogenannealedmirrorchip. 96

PAGE 97

Figure3-26. FESEMimagesofmicroporefeatureonaDRIE-fabricatedmirrorchipbeforeandafternishing 97

PAGE 98

CHAPTER4SURFACEMODIFICATIONMECHANISM 4.1ApproachtoClarifyingtheSurfaceModicationMechanismTheexperimentsdetailedinChapter 3 conrmedthefeasibilityoftheMAFprocessdevelopedinthisresearch.Theresultsfromvariousexperimentsdemonstratedthenishingprocessabilitytoremovematerial,reducethemicroporesidewallsurfaceroughnessofnickelandsiliconmirrorchips,andmostimportantly,improvetheX-rayreectanceofmirrorchips.Nevertheless,furtherknowledgeaboutthenishingprocessisrequiredtoengineerasurfaceusingthisnishingprocess.Thespecicmechanismbywhichthemicroporesidewallsaresmoothenedbytheprocesshasnotbeenclaried.Figure 4-1 depictspossiblewaysinwhichabrasiveparticlescanmodifyasurfaceinalooseabrasiveprocess.Dependingonthecharacteristicsofforce-exertingagentinthenishingprocess,abrasiveparticlesmayabradesurfacebyscratching,rolling,indenting,oracombinedaction. Figure4-1. Possibleabrasiveactionsonasurface Carefulobservationofchangestoasurfacebeforeandafternishingcouldrevealwhichtypeofabrasiveactionhasoccurred.SincetheMAFprocessstudiedinthisresearchinvolvesneabrasiveparticlesintendedastheprimarysurfacemodicationtool,itishypothesizedthattheseparticlesmodifythetargetsurfaceinwayswhicharecharacteristicoftheprocessnishingmechanism.Thus,bypreciselycharacterizingindividualsurfacemodications,insightmaybegainedintothemagnitudeanddirectionofforcesactingontheabrasiveparticlescreatingsuchsurfacemodications.Subsequently,regardingtheMAFprocessinthisresearch,anunderstandingofhowan 98

PAGE 99

appliedmagneticeldresultsinsurfacemodicationscouldbedeveloped,potentiallyleadingtothedevelopmentofatheoreticalprocessmodel.ThischapterwilldetailexperimentsaimedatstudyingsurfacemodicationsandMAFMmotiontondthesurfacenishingmechanism. 4.2ObservationofSurfaceModicationsonFlatWorkpieces 4.2.1ExperimentalEquipmentItisnotpracticallyfeasibletoobservesurfacemodicationscausedbythenishingprocessonmicroporemirrorchipsidewallsduetotheirmonolithicstructure.Themostpracticalworkpiecewouldbeasmoothatsurface,allowingthedirectapplicationhigh-resolutionsurfacecharacterizationtechniquesonspecicareasbeforeandafternishing.OnepossibilityforapplyingtheMAFprocesstoistosimplyaddavolumeofMAFMabovetheatworkpieceandactivatethemagneticeldgenerators.Figure 4-2 showsthismodiedprocessingprinciple.Themagneticeldacceleratessuspendedferrousparticleswhichinturndisplacesabrasiveparticles.Acceleratedabrasiveparticlesincloseproximitytotheatworkpiecesurfacemaycontactthesurfaceandcausesurfacemodication. Figure4-2. Schematicrepresentationoftheprocessingprincipleforatworkpieces TorealizethenishingprocessingprincipleofFigure 4-2 ,aworkpieceholderwasdesignedforthemirrorchipnishingmachinetohold550.5mmworkpieces,showninFigure 4-3 (A).Theworkpieceispositionedinsideasimilarlysizedsquarepocketonanaluminumdisk;theworkpiecemaybesecuredwithbitsofadhesivetape.Thedisk 99

PAGE 100

andworkpieceareplacedatthebottomofapolymericcontainerintowhichMAFMmaybeadded.Fourausteniticstainlesssteelrodsconnectthedisktoarigidstructureonthemachinetocontroltheworkpiecealignment.Figure 4-3 (B)isaphotographofa550.5mmsilicon(100)workpiecesecuredinthealuminumdiskwithadhesivetape.Figure 4-3 (C)showsaphotographofthemirrorchipnishingmachinewiththeholder,polymericcontainercontaining2mLofMAFM.Thepole-polegapis22.5mmandissetbythecontainerwidth. Figure4-3. TheA)atworkpieceholderdesignconceptandB)theactualholder,andC)aphotographofthemirrorchipnishingmachineretrottedtonishatworkpieces(PhotographscourtesyofRaulRiveros) 100

PAGE 101

4.2.2FinishingofaSputteredMetalSurfaceThoughmaterialremovalwasobservedinthemirrorchipnishingexperiments,individualsurfacemodicationswerenotclearlyidentiedbecauseoftheroughunnishedmicroporesidewallsurfacetexture.Itwasthoughtthatusingasmoothlow-hardnessmetallicsurfacelayerasatargetsurfacewouldallowclearvisualizationsofsurfacemodicationsformedbyincidentabrasives.Inanearlyexperiment,aatsiliconchipwassputter-coatedwitha50nmthicklayerofgold.Thechipwouldbenishedasdescribedintheprevioussection.TheexperimentalconditionsusedareshowninTable 4-1 .Theyaresimilartotheconditionsofpreviousmirrorchipnishingexperiments. Table4-1. Sputteredworkpiecenishingexperimentalconditions Workpiece5.650.38mm,Silicon(100),single-sidepolished50nmthickAu(sputtered)AbrasiveslurryUniversal-basedpolycrystallinediamond0.5mdiameter,1mLMagneticuidWater-based,anionicsurfactant,3.6wt%Fe3O4,1mLPole-poledistance22.5mmAlternatingcurrent22Hz,1ANumberofcycles79,200(1hr) A1010mareawasmeasuredbySWLIbeforeandafternishing;themeasurementsarecomparedinFigure 4-4 .Themeasurednishedsurfaceexhibitsindent-likedeformationsof2mdiameterand3nmdeep.Pooradhesionofthegoldlayeronthesiliconchipsurfacecausedlargesectionsofthegoldlayertoliftoffthesurfaceduringpost-nishingultrasonication.Oncenoticed,thechipwasremovedfromtheultrasonicbath.Thus,itispossiblethatthedeformationsonthenishedsurfacewerecausedbyeithertheultrasoniccleaningorincompleteremovaloftranslucentcrystallinediamondparticleswhichwouldcausesuchSWLIimageartifactstooccur.Theresultsfromthisexperimentarethereforeinconclusive.Anishingexperimentonaatworkpiecewithamorerobustsurfaceisdetailedinthefollowingsection. 101

PAGE 102

Figure4-4. Changesobservedinthesurfaceshapeafternishing 4.2.3FinishingofaPatternedSurfaceThenishingexperimentdetailedinthissectionwasperformedonacommerciallyavailableworkpieceusedforthecalibrationofscanningprobe-typemicroscopeslikeAFM.Itisa550.5mmsiliconchipwitha100nmthicksilicondioxide(SiO2)layeronitstopsurface.Thisoxidelayerhasbeenpreciselypatternedbythemanufacturer;aschematicoftheoxidelayerpatterisshowninFigure 4-5 .Atwo-dimensionalarrayof5msquarepocketsspacedata10mpitchhavebeenetchedthroughthe100nmthicknessoftheoxidelayer,exposingtheunderlyingsmoothsiliconsurface.Thisworkpiecewaschosenbecauseitspatternedsurfacewouldallowforsimpliedlocationandtrackingofspecicsurfacefeaturesbeforeandafternishing.Additionally,ifthenishingprocesswastomachineasignicantamountofsurfacematerial,itwouldlikelyresultinmodicationoftheedgesofthepatternedfeatures;suchchangescouldpotentiallyhintatthecharacteristicsofthenishingmechanism.Thisexperimentwasaimedattrackingthesurfacemorphologyversusnishingtimeusinghigh-resolutionAFMmeasurements.Additionally,thechemicalcompositionofthesurfacewouldbeobservedbyFESEM,EDSandAFM-basedphasecontrastmicroscopybeforeandafterthenishingphases.Table 4-2 liststheexperimentalconditionsusedinthisexperiment.The0.5mdiamondabrasiveemployedinpreviousexperimentsis 102

PAGE 103

Figure4-5. Schematicoftheworkpiecesurfacepattern Table4-2. Flatworkpieceexperimentalconditions Workpiece550.5mmsilicon100nmthickSiO2,patterned(seeFigure 4-5 )AbrasiveslurryPolycrystallinediamond0.5mdiameter,1mLMagneticuidWater-based,anionicsurfactant1.8wt%Fe3O4,1mLPole-polegap22.5mmAlternatingcurrent22Hz,1AMagneticuxdensity40.3mTatcenterbetweenpoletipsFinishingtimeTwo1hrphases(2hrtotal) used.Thepole-polegap(22.5mm)islargerthanthatusedinthemirrorchipnishingexperimentsduetothediameteroftheatworkpieceholder.Thewiderpole-polegapreducedthemagneticuxdensityatthecenterbetweenthepoletipsto40.3mT.Thesaturationmagnetizationofthe1.8wt%Fe3O4water-basedmagneticuidusedtopreparetheMAFMis10mT.ItwasthereforeexpectedthattheMAFMwouldoperateinmagneticallysaturatedconditions,asitdidduringmirrorchipnishingexperiments.Twonishingphaseslasting1hreachwereexecuted;anewvolumeofMAFMwaspreparedforeachphase.Aftereachnishingphasethechipwassubmergedinaseriesofultrasonicbathsofwater-baseddetergentsolution,deionizedwater,and200proofethanol.Thesurfaceoftheworkpiecewasthenblastedwith99.999%CO2snowtoremoveanyparticulatesonthesurface;immediatelyafter,thesurfacewasdriedwithultra-highpuritynitrogen 103

PAGE 104

gas.Allnishingandcleaningprocedureswereperformedinaclass1000cleanroomorbetter.Sincetheoxidelayerisintegraltothesiliconchip,itwasnotdamagedbythecleaningproceduresasthegoldsurfacewasinSection 4.2.2 .AsinglepocketwasselectedforAFMscanningbeforeandaftereachnishingphase.Thepocketwasapproximatelycenteredintheworkpiecesurface.UsingtheAFM'sbuilt-inopticalmicroscopeandlarge(9090m)AFMscansofthesurfaces,thepocketwasfoundaftereachnishingphase.Figure 4-6 (A)showsthree-dimensionalsurfaceshapesofthetrackedpocketbeforeandaftereachnishingphase.Apartfromrandomsurfacedebris,nochangesareobservedwithnishingtimeinFigure 4-6 (A).Toverifywhetheranyslightchangesoccurredtotheperimeterofthepocket,suchasanelongationinthenishingdirection(coincidentwiththeindicatedscandirection),top-downviewsofthesurfacesofFigure 4-6 (A)weremanipulatedtoexposetheperimeterofthepocketedgeontopoftheoxidelayer.TheseimagesareshowninFigure 4-6 (B).Acloseexaminationoftheseimagessuggeststhatnochangestothepocketperimetergeometryoccurred.Toverifywhetherachangetotheheightdifferencebetweentheoxidesurfaceandtheexposedsiliconsurfaceoccured,multipleprolelinesacross3030mscansofthesurfaceweretaken.Theaverageheightdifference,h,betweentheoxidelayerandthesiliconsurfacewascalculated.ResultsfromthesemeasurementsareshowninFigure 4-7 .Thespeciedoxidelayerthicknesswas100nm.Theinitialsurfaceandthesurfaceafter1hrofnishingappearedtohavealarger-than-speciedhvaluewhereasafter2hrofnishing,thehvalueis100nm.Itisbelievedthatdebrisontheworkpiecesurfacecouldhavebeengraduallyremovedbythenishingprocessandcleaningprocedure,resultingintheobservedtrendinh.Toobservechangesinthesurfaceroughnessduetothenishingprocess,threeareasaroundthesingletrackedpocketwerechosenforanalysis.TheseregionsofinterestareidentiedinFigure 4-8 .Asshown,thetopspaceandtheleftspace 104

PAGE 105

Figure4-6. AFMimagesofthetrackedpocket Figure4-7. Changesinthemeasuredheightdifference,h,withnishingtime 105

PAGE 106

regionlieatoptheoxidelayersurface.Thecentralspaceregionisdirectlyonthesilicon.Beforeandaftereachnishingphase,a33mAFMscanwastakeninthethreeregionsofinterest.Ten3msamplinglengthsurfaceprolelineswereusedfromeachAFMscan;RzandRqvaluescalculatedfromtheseprolelineswereaveragedandplottedinFigure 4-9 Figure4-8. Schematicofregionschosenforobservation ThereappearstobeareductionofRzandonlyaslightreductionoftheinthetopspaceafter1hrofnishing.Nosignicantchangeappearstohaveoccurredintheleftspace.Thecentralspace,however,appearstohavereducedinbothRzandRqwithnishingtime.Itispossiblethatthecentralspace,whichiscomposedofsilicon,machinesataslightlyfasterratethanthesurroundingoxidelayersincesilicondioxide(quartz)isgenerallyslightlyharderthansilicon.Anattemptattrackingthechangeinsinglesurfaceprolesinthethreeregionsofinterestversusnishingtimewasmade;acomparisonofsurfaceprolesofapproximatelythesamelineonthemeasuredsurfacesisshowninFigure 4-10 .LikethecalculatedRzandRqvaluesofFigure 4-9 ,itisdifculttodeterminespecicdifferencesintheshapeofthesurfaceprolesofthetopspaceandleftspace.Thecentralspacehoweverexhibitsaratherclearreductionintheamplitudeoflow-wavelengthsurfaceproleelements.Itisbelievedthattheslightlylowerhardnessofthecentralspacesurfaceisthereasonwhymorematerialseemstohavebeenremovedordeformed.ItcouldbespeculatedthatperhapstheowofMAFMduringthenishingprocessexhibits 106

PAGE 107

Figure4-9. MeasuredvaluesofRzandRqwithnishingtime someadditionalturbulenceasitowsovertheedgeoftheoxidelayerandontothecentralspace,causinggreaternishingforces;howeveritwouldbedifculttoverifywhethersuchturbulenceoccursandactuallyaffectsthenishedsurface.Thechangestothesurfaceprolesinthecentralspacesuggestthatthesurfacewasunevenlymodied;itispossiblethenthatthesurfacewasmechanicallymodiedbyabrasiveaction.Inthedevelopmentofasurfacemodifyingprocess,itseffectonthesurfacematerialcompositionshouldbestudied.Foursurfaceanalysistechniqueshavebeenappliedtodiscoverchanges,ifany,tothesurfaceandsubsurfacematerialcompositionofthe 107

PAGE 108

Figure4-10. Comparisonofsurfaceprolesinthethreetrackedregionswithpolishingtime patternedworkpiece.Figure 4-11 showslowmagnicationFESEMandBSEimagesoftheworkpiecesurfacebeforenishingandafter2hrofnishing.NolargechangestothesecondaryelectronorbackscatteredelectronemissioncharacteristicsofthesurfaceareobservedinFigure 4-11 whichsuggestthatnolargeresidueorimpregnationofforeignmaterialshasoccurred.Figure 4-12 showshighermagnicationFESEMandBSEimagesoftheworkpiecesurface.TheFESEMimagesshowsmalldarkspotsscatteredrandomlyontheas-receivedsurface;thesearenotpresentonthesurfaceafter2hrofnishing.Besidesthedebris,nochangeisobservedtothegeometryofthepocketfeatures.TheBSEimagesofFigure 4-12 (B)exhibitadifferenceincontrastbetweenthesiliconandoxidelayersurfaces;however,thisislikelyduetotheconditionoftheFESEMatthattime.Thus,nodifferenceinthesubsurfaceatomiccompositionisvisibleatthismagnication.TheFESEMusedtoimagetheworkpiecesurfacewasalsoequippedwithanenergydispersiveX-rayspectroscopy(EDS)system.EDSmeasurementsweretakeninthecentralspaceandtheleftspacebeforenishingandafter2hrofnishing.The 108

PAGE 109

Figure4-11. Low-magnicationFESEMandBSEimagesoftheworkpiecesurface Figure4-12. High-magnicationFESEMandBSEimagesoftheworkpiecesurface 109

PAGE 110

acquiredspectrafromthesemeasurementsareshowninFigure 4-13 .Inbothareas,nochangeappearstohaveoccurredtothesubsurfaceatomiccomposition. Figure4-13. ComparisonofEDSspectra TheAFMwasoperatedinnon-contactmodeinwhichtheAFM'scantileverwasoscillated(atsomeappropriatecommandedfrequency)asitscanned;measurementsdoneinthiswaytendtobelesssensitivetosuddenchangesinsamplesurfaceheight.VariationsencounteredbytheoscillatingAFMtipinthesampletopographyandsurfacematerial'smechanicalandintermolecularpropertiesmightcauseaphaselaginthecantilevertiposcillations.Ifthephaselaginformationismapped,thedistributionofdifferentmaterialsonasurfacemaybeimaged.Figure 4-14 showsphasecontrastimagesattainedfromtheAFMmeasurementsperformedonthesurfaceinthethreeregionsofinterestversusnishingtimeareshown.Intheseimages,verydarkorlightcolorsindicatechangesinthephaseofthetiposcillation.AcloseexaminationofimagesofFigure 4-14 suggeststhatforeignsurfacematerialswereremovedbythenishingand 110

PAGE 111

cleaningprocessesappliedtotheworkpiecesurface,sincetheimagesofthesurfacesafter2hrofnishingappeartobemostlyfreeofforeignmaterials. Figure4-14. Phasecontrastimagesofthethreeregionsofinterestwithnishingtime Itishypothesizedthatthediamondabrasiveparticlesareresponsible,atleastinpart,fordetectedchangestotheworkpiecesurfaceafternishing.Tobetterunderstandtheeffectivenessofthediamondparticles;theexperimentwasrepeatedwith2mLmagneticuidonly.Duetoasupplierdifculty,aworkpiecewitha122nmoxidelayerthicknesswasused;thisdifferenceisnotexpectedtosignicantlychangethenishingcharacteristicsoftheprocesssincethepocket'slowaspectratio(50:1)ishardlychanged.ThesurfacewasobservedinthesamewayastheinitialworkpiecebyusingFESEM,BSE,andEDS.Theseimagesandmeasurementsarenotshown;however,no 111

PAGE 112

changestothesurfacegeometryoratomiccompositionwereobserved.Phasecontrastimagesshowedadecreaseinthepresenceofforeignmaterial;however,thischangewasnotasperceptibleasinFigure 4-14 .Figure 4-15 showsh,Rz,andRqvaluesmeasuredbyAFMonthesurfaceoftheworkpiecenishedbymagneticuidonly.Nodifferenceisobservedinhversusnishingtime.Theinitialsurfaceroughnessoftheworkpiece(0.2nmRq)isslightlyhigherthanthatofthepreviouspatternedworkpiece(0.15nmRq);however,theslightroughnessreductiontrendsobservedinFigure 4-9 arenotseeninFigure 4-15 (BD).Theseresultssuggestthatdiamondisnecessarytocreatetheobservedsurfacemodications.Theexperimentsdetailedinthischaptershowedpromisingresults,yettheknowledgegainedfromtheseinvestigationsdoesnotrevealthesurfacemodicationmechanism.Arequirementofanefforttodiscoverthesurfacemodicationmechanismofasurfacenishingprocessrequiresdirectobservationofsurfacemodications.Previousattemptsatcreatingthesehaveyieldedinconclusiveorminimalchangestotheoriginalsurface.Theprimaryreasonfortheseunclearresultsisbelievedtobethefactthattheat-workpiecenishingcongurationdoesnotrecreatethelevelsofMAFMowstressesexperiencedbymicroporesidewallsduringnishing.Thus,thenextsectionwilldetaileffortsandinsightsgainedthroughexperimentationwithaat-workpiecenishingsetupinwhichtheowofMAFMisconstrictedtoasmallchannel,whichiscomparableindimensionstoasinglemicropore. 4.3ObservationofSurfaceModicationsandFluidFlowthroughaSingleMicropore 4.3.1ExperimentalSetupFigure 4-16 showsaschematicofasetupdesignedtorestrictMAFMowthroughachannelofcomparabledimensionstoasinglemicroporeonamicroporeoptic.Figure 4-16 (A)depictsaatworkpiece.Twothinandwidespacersareplacedontheworkpiecesuchthatthedistancebetweenthemissimilartothelengthofamicropore.Acylinder 112

PAGE 113

Figure4-15. Resultsofexperimentrepeatedwithoutabrasive 113

PAGE 114

contactsthelm,leavingathingapbeneathitandtheworkpiece.AcontainerisbondedtotheworkpiecetomaintaintheMAFMnearthetargetsurface.AsshowninFigure 4-16 (B),whenanappropriateamountofMAFMisaddedtotheleftsideofthecylinderandamagneticeldisappliedontheoppositeside,theMAFMwillowthroughthegapbeneaththecylinder.Thegeometryandplacementofthecylinderisexpectedtoincreasethenormalandshearstressonthetargetsurfacesinceessentiallycreatesaconverging-divergingnozzle.Increaseduidowstressesonthetargetsurfaceareexpectedtoyieldgreatersurfacemodicationsthaninpreviousexperimentalefforts. Figure4-16. Schematicandcross-sectionofthemicroporereplicationsetup Figure 4-17 showsthematerialsusedtorealizetheconceptshowninFigure 4-16 .Figure 4-17 (A)showsa316Lstainlesssteelrod(6.3515.93mm).Itiswrappedwithtwostrips25.4m-thickpolyimidelmushagainsttheouteredgesofthecylinder;thesestripsactasthespacerofFigure 4-16 .Figure 4-17 (B)showsabottomviewoftheworkpiece,which,inthiscase,isaglassmicroscopeslide.Theglassworkpieceisadhered(siliconorepoxy)toasquarepolycarbonatetubewhichhasbeencuttoalengthof6.5mm.Figure 4-17 (C)showsaphotographofanotherlengthofpolycarbonatetubelledwithepoxyputtythathasbeenmoldedtotheshapeofthecylindercomponent.Thepurposeoftheepoxyputtyistopositionthecylinderinthecenterofthecontainerinarepeatablemanner.Duetopositioningerrorsduringthecuringprocess,thecylinderaxisisapproximately0.25mmoffcenter.Oncecured,thecylinderwrappedinlmisplacedinsidethecontainer,contactingtheworkpiece. 114

PAGE 115

Figure4-17. Photographsofthemicroporereplicationsetupcomponents(PhotographscourtesyofRaulRiveros) Thecylinderguideispositionedabovethecylinder,andtheentirestructureisclampedtogetherasshowninFigure 4-17 (D).MAFMisaddedpriortoclamping,andthecontactofthecylinderisveriedbyvisualinspectionasshowninFigure 4-17 (E).Anattemptwasmadetomeasurethecross-sectionaldimensionsofthereplicatedmicropore.SWLImeasurementsoftheproleonthesurfacealongtheaxisofthecylinderwrappedinlmrevealsthat,intheparticularsetupshowninFigure 4-18 ,thelengthofthemicroporeis1.75mmandthewidthofthemicroporeis25.4m.Thewidthiscontrolledbythethicknessofthelmandbyanycompressionofthelmwhichmayoccurduringclampingofthestructure.Tocheckwhetherthelmissignicantlycompressible,itsthicknesswasmeasuredwithamicrometer.Whenmeasuredappropriately,themicrometer'sreadoutindicatedathicknessof26m(resolutionis1m).Byintentionallycompressingthelmbya1/8turnofthemicrometerscrew,athicknessof25mwasreadout.Thisresultsuggeststhatthelmhasasufcientlyhigh 115

PAGE 116

compressivestrengthandwillnotyieldsignicantlywhenclampedinthemicroporereplicationstructure.ASWLImeasurementofthecylinder'sgroundsurfacetextureindicatesasurfaceroughnessof2.100.14mRz.Thelengthofthemicroporeis1.75mm;however,thiscanvarybyabout0.5mm,dependingonhowthelmstripsarecutandplaced.Thewidthofthemicroporecanbeestimatedbyconsideringthenominalthickness(25.4m),themanufacturer-speciedthicknesstoleranceof2.5m,andthesurfaceroughnessRzofthecylindersurface.Sincethelmrestsatoptheroughnessofthecylinder,thenominalwidthofthemicroporeshouldbeequaltothatofthelm.ThemicroporewidthtolerancewillequalthelmthicknesstoleranceandabestestimateofhalftheRzvariation.Thewidthofthemicroporecanbestatedas25.4+3.6)]TJ /F6 7.97 Tf 6.59 0 Td[(2.5m. Figure4-18. Measurementsofthereplicatedmicroporecross-sectionaldimensions ToobservethecharacteristicsoftheMAFMowthroughthemicroporeandoftheMAFMsurface,adigitalmicroscopewithavideoscreenresolutionof320240 116

PAGE 117

pixelswasplacedbeneaththepoletips.Itismountedonahigh-strengthtripodwithathree-axismanualnepositioningstageasshowninFigure 4-19 .Whenthemicroporereplicationsetupisclampedbetweenthepoletips,themicroscope,whichisconnectedtoacomputer,iscapableofdetecting(notresolving)abrasiveparticles.Afront-facingvideocameraisplacedsuchthatasideviewoftheMAFMsurfaceiswithinitsframe,allowingfortheobservationofinterestingvibrationswhichoccurontheMAFMsurface.Boththemicroscopeandthevideocamerarecordataframerateof30framespersecond(fps).Videosrecordedfrombothdevicesaresynchronizedbyaligningtheirrespectiveaudiotracks.Thismethodresultsinasynchronizationerroroflessthan1/30s. Figure4-19. AphotographshowingthemicroscopeandvideocameraplacementsforMAFMmotionobservation(PhotographcourtesyofRaulRiveros) 4.3.2EffectsofAbrasiveSizeonSurfaceRoughnessInitialtestingattemptswerecarriedoutusingglassmicroscopeslidesasaworkpiece.ThisallowedforobservationoftheMAFMmotionandrenementoftheexperimentalprocedures.Unfortunately,noclearmodicationoftheglassworkpiecesurfaceoccurred.Theworkpiecematerialwasswitchedtocold-rolledCDA260brasssheets(70%Cu,30%Zn)soldcommerciallywithagroundsurfacetexture.ThebrassmaterialissofterthanglassandshoulddeforemmoreeasilytoabrasiveparticlesintheMAFM.Figure 4-20 (A)isaphotographofthemicroporereplicationsetupinwhicha300 117

PAGE 118

m-thickbrasssheetisadheredtoaglasssubstrate.Thepolycarbonateisadhereddirectlytothebrasssheetwithepoxy.Sincebrassisopaque,thelmcontactareacannotbeobserved.Tocheckforadequatecontactbetweenthecylinderandthebrasssheet,alightwasshoneononesideofthecylinderasshowninFigure 4-20 .Thelightpatterntransmittedthroughthereplicatedmicroporeindicateswhetherfullcontactisachievedornot. Figure4-20. Photographsofthemicroporereplicationsetupusingabrasssheetastheworkpiecematerial(PhotographscourtesyofRaulRiveros) PreliminarynishingtrialsonbrassworkpiecescreatedmodicationstothebrasssurfaceshapethatweredetectabletotheSWLI,validatingthisexperimentalmethod.Thusanabrasive-sizeeffectstudywascarriedout.Theabrasivessizesusedinthisstudywereselectedbecauseoftheirknowneffectivenessinpriormicroporeopticnishingexperiments;thusthespecicabrasivesizeselectionisnotlinear.Asabaselinecomparison,anishingtrialwasrunwithoutabrasiveslurry,whereinthemissingvolumeofabrasiveslurrywasreplacedwithandequalvolumeofextramagneticuid.Theconditionsofthisabrasive-lesstrialareshowninTable 4-3 .Onevolumeof200Lofmagneticuidwasaddedtobothsidesofthecylinder.EarlyMAFMowobservationswithglassworkpiecessuggestedthatthealternatingfrequenciesusedinthemicroporenishingexperimentsweretoohightodevelopadequateow; 118

PAGE 119

Table4-3. Finishingconditionswithmagneticuidonly WorkpieceCDA260brass19.5300.304mmMagneticuidWater-based,anionicsurfactant3.6wt%Fe3O4400LAbrasiveslurryNoneSuppliedWaveformAlternatingsquarewave+25.00V)]TJ /F1 11.955 Tf 9.3 0 Td[(26.00V15speriodFinishingphasetime4hrNumberofnishingphases2Pole-polegap20.00mmMagneticuxdensityatmicropore145mT thisisthoughttooccurbecausethisexperimentalsetuprestrictsMAFMowtoasinglemicropore.Evenafrequencyof1Hzyieldedanabrasive-particle-reciprocationamplitudeoflessthan300m,whichisshorterthanthethicknessofatypicalmicroporeopticandconsideredinadequate.Thus,thepowersupplywasconguredtosupplyanalternatingsquarewavewithlong(30s)cycleperiod.TheowofMAFM,accordingtothemicroscopevideo,appearedtoslowsignicantlyafter8sofacoilactivation;therefore,theperiodwaslimitedto15s,suchthatquickandconsistedowoccursthroughoutthenishingtrial.Thecoilwiredtoreceivethenegativesideofthewaveformappearedtogenerateaweakermagneticeldthanitscounterpart;tocompensate,thenegativevoltagewasset1Vlowerthanthepositivevoltage.Apreliminarytrialonaseparatebrassworkpieceindicatedthata4hrnishingtimeisenoughtocreatesurfacemodicationsvisibletotheSWLI.Two4hrphaseswerecarriedoutinanattempttoobservetheeffectsofthenishingprocessversustime.Thepole-polegapwaslimitedto20mmbythewidthofthepolycarbonatecontainer.Afternishing,theworkpiecewasultrasonicallycleanedinwater-baseddetergentanddeionizedwater.ItwasthenblastedwithultrapurecarbondioxideCO2snowanddriedwithultrapurenitrogenN2gas.Figure 4-21 showsseveralresultsfromthis 119

PAGE 120

experiment.Figures 4-21 (AC)arereected-lightintensitymaps(attainedformSWLImeasurements)ofthesamespotinthetargetarea(areadirectlybeneaththecylinder).Intheseintensitymaps,darkregionsareareasinwhichlightiseitherabsorbedorreectedawayfromtheSWLI'sopticalaxis.Figures 4-21 (AC)donotindicatethatanychangehasoccurred.Figure 4-21 (D)showsensembleaveragepowerspectraldensity(PSD)curvescalculatedfromthreehundred100mlongprolelinestakenfromthe100100marea(planeremoved)SWLImeasurements(lateralresolutionx=0.276m)alongthenishingdirection,whichisperpendiculartothegroundsurfacetexturedirection.IfanyofthenishedPSDcurvesdepartgreatlyfromtheunnishedcurve,thisindicatesthattheamplitudeofsurfaceasperitieshasreducedorincreased[ 110 ].NosuchchangesareseeninFigure 4-21 (D). Figure4-21. SWLImeasurementsofthebrassworkpiecesurfacenishedbymagneticuidonly 120

PAGE 121

Typically,anishingprocesswillaffectacertainbandofspatialwavelengths;thus,tocomparethechangeinsurfaceroughnessbetweentheunnishedandnishedsurface,aband-passlterwithcutoffscorrespondingtothespatialfrequenciesofinterestisapplied.Whencompilingtheresultsofalltrialsinthissection,itwasdeterminedthataGaussiansplinelterwithalowcutoffc=5mandahighcutoffofs=1madequatelyencompassedthechanges.Figure 4-21 (E)showschangesintheaveragesurfaceroughnessRafromthenishingtrialwithoutabrasive.TheRaparameterwascalculatedandaveragedfromten100mprolelinesalongthenishingdirectionwithina100100mareafromeachmeasurement.Theblack(larger)errorsbarsarethestandarddeviationsofRa;theblue(smaller)errorbarsindicatethestandarddeviationofthevariationofthemeanRavalue(calculatedusingthesameten-prole-linemethod)fromtenmeasurementsofthesameareaonthistypeofworkpiece.AchangeinthemeanvalueofRathatissignicantlylargerthanthisstandarddeviation(blueerrorbars)indicatesthatsomedetectablechangeoccurredwhichmodiedthemeasuredsurfaceproles.Suchachangedoesnotnecessarilyimplythatthesurfacewassignicantlynished,however.Figure 4-21 (E)suggestthatlittletonochangeresultedfromnishingwithonlymagneticuid.Thenishingconditionsforthevarying-abrasivesizetrialsareshowninTable 4-4 .Inthesetrials,200Lofmagneticuidand200Lofabrasiveslurrywasused.Asmentionedearlier,theabrasiveswerechosenfortheyhavebeeneffectiveinpriormicroporeopticnishingstudies.AllotherparametersandproceduresareidenticaltothoseofTable 4-3 .Itisusefultonotethatthe50nmand6mdiamondslurriesarewater-based.The0.2mand0.5mdiamondslurriesarebasedinaproprietarychemicalmixturemiscibleinbothwaterandoil,hereafterreferredtoasuniversalbase.Figure 4-22 portraysvariousresultsfromthenishingtrialinwhichMAFMwaspreparedwith50nmwater-baseddiamondslurry.Uponcloseinspectionoftheintensitymaps,darkspotsofabout2mappearscatteredthroughoutthesurfaceapparently 121

PAGE 122

Table4-4. Brassworkpiecenishingconditions WorkpieceCDA260brass19.5300.304mmMagneticuidWater-based,anionicsurfactant3.6wt%Fe3O4200LAbrasiveslurry50nmwater-basedpolycrystallinediamondslurry0.2mpolycrystallinediamondslurry0.5mpolycrystallinediamondslurry6mmonocrystallinediamondslurry200LSuppliedMAFM400L,200LaddedtoeachsideofcylinderSuppliedWaveformAlternatingsquarewave+25.00V)]TJ /F1 11.955 Tf 9.3 0 Td[(26.00V15speriodFinishingphasetime4hrNumberofnishingphases2Pole-polegap20.00mmMagneticuxdensityatmicropore145mT concentratingontheupperregionsofthesurfacetexture,suchastheregiondelineatedbythedottedredboxinFigures 4-22 (AC).ThePSDcurveafter8hrofnishingexhibitsalargedeparturefromtheunnishedPSDcurve.ThisreductioninsurfaceasperityamplitudeisreectedinFigure 4-22 (E),whichshowsasteadyreductioninRawithnishingtime.Figure 4-23 portraysvariousresultsfromthenishingtrialinwhichMAFMwaspreparedwith0.2muniversalbasediamondslurry.TheintensitymapsofFigures 4-23 (AC)indicatethatdarkspotsappearscatteredthroughoutthesurfacewithnishingtime.ThiseffectisnotasclearasthatofFigures 4-22 (AC).ThePSDcurvesofFigure 4-23 (D)showonlyslightdeparturefromtheunnishedPSDcurve.AccordingtoFigure 4-23 (E),arelativelylargereductionoccurredaftertherst4hrphase,butthesecond4hrdidnotsucceedinasimilarway.Itwasnoticedthatthisabrasiveslurrywillsometimesformamuddyresidueatthetargetarea.CareistakentothoroughlymixtheMAFMbeforeitisaddedtothecontainer;however,particlesmaystillaggregateduring 122

PAGE 123

Figure4-22. SWLImeasurementsofthesurfacenishedwith50nmdiamondwater-baseddiamondslurry thenishingphases.Itispossiblethatamuddyresidueformedearlyoninthesecondnishingphaseandpreventednishingfromoccurringinthetargetarea.Figure 4-24 portraysvariousresultsfromthenishingtrialinwhichMAFMwaspreparedwith0.5muniversalbasediamondslurry.TheintensitymapsofFigures 4-24 (AC)donotshowclearevidenceofdarkspotsorachangeinreectivity.ThePSDcurvesofFigure 4-24 (D)appeartohaveonlyslightlyreducedinamplitude.AslightreductioninRaappearstohaveoccurredaftertherst4hrofnishing,whilealargerreductionoccurredafterthesecond4hrphase.Amuddyresiduewasalsofoundonthetargetsurfaceaftertherst4hrnishingphase.Itispossiblethatthismuddyresiduelimitedthenishingaction. 123

PAGE 124

Figure4-23. SWLImeasurementsofthesurfacenishedwith0.2mdiamonduniversalbasediamondslurry Figure 4-25 portraysvariousresultsfromthenishingtrialinwhichMAFMwaspreparedwith6mwater-baseddiamondslurry.TheintensitymapsofFigures 4-25 (AC)shownoclearlydiscernibledifferences.ThePSDcurvesofFigure 4-25 (D)exhibitnointerestingdeparturefromeachother.ThislackofchangeisreectedintheRavaluesplottedinFigure 4-25 (E).TheresultsofFigure 4-25 suggestthatthis6mdiamondwater-baseddiamondslurrywasineffectiveatnishingthetargetsurface.ThechangesinRaforallnishingphasesaresummarizedinFigure 4-26 .Figure 4-26 (A)plotsthechangesinRacausedbytherstandsecond4hrnishingphases.Itappearsthatonlythe50nmwater-baseddiamondandthe0.5muniversalbasediamondslurryachievedconsistentreductionsinRa.ThetotalchangesinRaareplottedinFigure 4-26 (B).Themosteffectivediamondslurrysizeappearstohavebeen 124

PAGE 125

Figure4-24. SWLImeasurementsofthesurfacenishedwith0.5mdiamonduniversalbasediamondslurry 50nm.The0.2mand0.5mabrasiveswerealsoeffectiveatsmoothingthetargetsurface.Figure 4-26 (B)suggeststhatbothmagneticuidand6mwater-baseddiamondslurrywereineffectiveatnishingthetargetsurfaceinthisexperiment.Itispossiblethatabrasivesizesbetween0.5mand6mwouldbeeffective;however,thisisfuturework. 4.3.3High-ResolutionImagingofSurfaceModicationsTheinvestigationsoftheprevioussectionindicatedthatMAFMpreparedwith50nmwater-baseddiamondslurryiseffectiveatcreatingsurfacedeformationslargeenoughtobedetectedbytheSWLI.Thus,thisslurrywasselectedforanishingexperimentwherethebrassworkpiecehadbeenpreviouslysmoothedbyastandard-pad-polishing 125

PAGE 126

Figure4-25. SWLImeasurementsofthesurfacenishedwith6mdiamondwater-baseddiamondslurry Figure4-26. MeasuredchangesinRaversusmeanabrasiveparticlediameter 126

PAGE 127

Table4-5. Smoothbrassworkpiecenishingconditions WorkpieceCDA260brass19.5300.304mmMagneticuidWater-based,anionicsurfactant3.6wt%Fe3O4200LAbrasiveslurry50nmwater-baseddiamondslurry200LSuppliedMAFM400L,200LaddedtoeachsideofcylinderSuppliedWaveformAlternatingsquarewave+25.00V)]TJ /F1 11.955 Tf 9.3 0 Td[(26.00V15speriodFinishingphasetime4hrPole-polegap20.00mmMagneticuxdensityatmicropore145mT process.Usinghigh-resolutionAFMimaging,theeffectsofthenishingprocessonthesmoothbrassworkpieceusingthemicroporereplicationsetupareanalyzed.Table 4-5 detailsthenishingconditionsusedinthisexperiment.Incontrasttothenishingtrialsoftheprevioussection,onlyone4hrphasewasapplied.Thesmoothbrassworkpiecewascleanedusingthesamecleaningproceduresappliedintheprevioussectionbeforeandafterthenishingphase.Themicroporereplicationsetupwasassembledwiththesmoothbrassworkpiece,andthenishingtrialwasexecuted.After4hrofnishing,thesetupwasunclampedandtheepoxyputtycylinderguideremoved.UponremovalofthecylinderandmuchoftheMAFM,itwasnoticedthattheremainingMAFMmixturedelineatedthelmcontactregions.Adiamondscribewasusedtoscratchwavylinesalongthelmcontactregions.Aroughly-squareboxwasalsoscratched,enclosingthetargetarea.ThesescribemarksareusedtoreferencethelocationofAFMmeasurements.Figure 4-27 showsthelabelingscheme.Figure 4-27 (A)showsamicroscopeimageattainedfromaprevioustrialusingaglassworkpiece.VisibleinFigure 4-27 (A)aresettledandaggregatedabrasiveparticles,thelmcontactregionsandthespacebetweenthelmcontact 127

PAGE 128

regionsknownasthetargetarea.Figure 4-27 (B)isaphotographofthenishedsmoothbrassworkpiecemountedontheAFMstagebeforescanning.VisibleinFigure 4-27 (B)arethescribemarks,whichareschematicallydepictedinFigure 4-27 (C). Figure4-27. Detailsoftheregionsofinterestonthenishedsmooth-brassworkpiece(PhotographcourtesyofRaulRiveros) ThecleanedunnishedworkpiecesurfacewasscannedbyAFMpriortonishinginanareathatwasroughlyinthecenteroftheworkpiece.Whenthemicroporereplicationsetupwasassembled,carewastakentopositionthecylinderapproximatelyabovethisregion.TheconsistencyoftheunnishedsurfacetexturewasconrmedbySWLImeasurementsofthesurface.Thus,theAFMmeasurementsoftheunnishedsurfacearebelievedtoberepresentativeoftheoverallsurfacetexture.Afternishingandcleaning,twoareaswerechosenforanalysis;theseareindicatedinFigure 4-27 (C)astargetareaand"outsidebox."asseeninFigure 4-27 (A)thetargetarealiesdirectlyunderneaththecylinderandisexposedtothehighestvelocityowasgatheredfrommicroscopevideoobservations.Theoutsideboxarealies1.5mmawayfromthetargetareaandisonlyexposedtoslightmovementsofthesettledabrasiveparticles.InFigure 4-27 (A)thecoiltotheleftoftheimageisactivatedandstriationsareseeninthesettledabrasives.Duringalternation,thisimagesomewhatappearsasifitishorizontallyipping.AllAFMscanswereperformedalongthenishingdirection,which 128

PAGE 129

isperpendiculartothecylindercontact(linecontact).SupersharpAFMcantilevertipswereusedwithaspeciedtipradiusof2nm.Threescansizesweremeasuredfromeachregion:90m,10m,and1m.ThesesurfaceheightimagesfromthesemeasurementsareshowninFigure 4-28 .TheunnishedsurfacetextureofFigure 4-28 (A)ischaracterizedbyrandomlinearscratchesofvaryingdepth.The1mscanofFigure 4-28 (Aiii)showsscratchesandhigh-frequencytexturecreatedbythepadpolishingprocess.ThenishedsurfacesofFigure 4-28 (Bi)and 4-28 (Ci)indicatethatthenishedsurfacehaspatchesof5nmlowerinheight.Thesearethoughttobecausedbythebrass'grainstructure.Nocolordifferencesintheseregionswereseeninopticalmicrographsofthesurface,andAFMphaseimagesdonotindicatealargedifferenceinmechanicalpropertiesbetweentheupperandlowerregionsofthistexture.The10mand1mscansweretakeninsidethesedeeperregions.TheonceclearlydenedscratchmarksseeninFigure 4-28 (Aii)appeartohaveagrainytextureinFigures 4-28 (Bii)and 4-28 (Cii).ThescratchesinFigure 4-28 (Cii)exhibitamorepronouncedgrainytexturethanthoseofFigure 4-28 (B-ii).Figure 4-28 (B-iii)revealsthepresenceofanewroughertexture,thoughthesurface'soriginalscratchesarestillvisible.Inthetargetarea,Figure 4-28 (Ciii)showsahighlymodiedtexturecharacterizedbyconvexprojectionsranginginwidthfromroughly50nmto80nm.ThesurfacemodicationsobservedinFigures 4-28 (B)and 4-28 (C)arebelievedtohavebeencausedbythenishingprocess.BothultrasoniccleaningandCO2snowcleaningaremechanicalprocessesandareknowntocausedamagetosoftorlooselyboundmaterials.Surfacedamagebytheseprocesseswasaconcernonthissoftbrassmaterial,butnoevidenceofdamagewasobserved.ThelackofsimilaritybetweenFigures 4-28 (Biii)and 4-28 (Ciii)rulesoutdamagebyultrasoniccleaningsinceultrasoniccleaning,appliedinarigorousmanner,woulddamagetheentire 129

PAGE 130

Figure4-28. Heightmapsoftheunnishedandnishedsmooth-brasssurfacemeasuredbyAFM surface,producingsimilarfeaturesthroughout.CO2snowcleaninghasbeenshowntocausedamagetothingoldlayers[ 111 ].AneffortwasmadeduringtheCO2snowcleaningofthisbrasssampletoblasttheentiresurfaceoftheworkpieceequally,soastonotdwellinoneareafortoolong.Thus,iftheCO2snowcleaningprocedurehadsignicantlymodiedthesurface,similardamagewouldbeobservedinallthree1mmeasurements;thisisnotthecase,however. 130

PAGE 131

Figures 4-29 (AC)arethree-dimensionalimagesofthe1mscansoftheunnished,outsidebox,andtargetareasurfaces.TherawAFMheightdatawasprocessedbyrstapplyingazero-orderline-wisecorrectionandthenapplyingarst-ordertwo-dimensionalpolynomialt(planet).PSDcurvescalculatedfromthesethreemeasurementsandSWLImeasurementsofthesameregionsareplottedinFigure 4-30 .WhiletheSWLI-measuredPSDcurvessuggestthatthenishedsurfacesarerougher,boththeoutsideboxandtargetareaAFM-measuredPSDcurvesexhibithigheramplitudesbetween3m)]TJ /F6 7.97 Tf 6.58 0 Td[(1and30m)]TJ /F6 7.97 Tf 6.59 0 Td[(1;thisrangecorrespondstothewidthoftheconvexfeaturesobservedonthenishedsurfaces.TheareaaveragesurfaceroughnessSawascalculatedusingallpointsfromeachmeasurementinsteadofindividualprolelines;thecalculatedSavaluesareshowninFigure 4-31 .TheoutsideboxhasthehighestSavalue,thisispossiblyexplainedbythefactthattheoutsideboxregionresemblestheunnishedsurfacewhilesuperimposingapartialversionofthenishingprocesseffects.Thetargetarea'sSavalueissimilartothatoftheunnishedsurface,thoughverydifferentinsurfacetexture.Thenishingprocessappearstohaveheavilymodiedthetargetareasurface. Figure4-29. Three-dimensionalAFM-measuredimagesoftheunnishedandnishedsurfacesandtheircorrespondingPSDcurvesandSavalues 4.3.4ObservationofMAFMMotion 4.3.4.1MicroporeowobservationsetupTheowofMAFMusingaglassworkpiecematerialwasoftendisturbedbythepresenceoflargesettledaggregatedparticlespartiallyblockingthemicropore.This 131

PAGE 132

Figure4-30. ComparisonofPSDcurvesoftheunnishedandnishedsurfacesfromAFMandSWLImeasurements Figure4-31. AFM-measuredSavaluesoftheunnishedandnishedsurfaces wouldnotnormallyoccurinthenishingofmicroporeopticssincelargeparticlestypicallysettleatthebottomofthenishingcontainer,farawayfromanymicropores.TopreventthisissueandtoattempttovisualizethetypeofMAFMowcharacteristicsoccurringinthenishingofmicroporeoptics,adifferentmicroporereplicationsetupwasdesignedandbuilt.AschematicofthissetupisshowninFigure 4-32 (A).Amicroscopeslideisusedasabasematerial.Sectionedglassisusedtoformtheloweranduppermicroporesidewalls.Theloweranduppermicroporesidewallsareseparatedbytwostripsofpolyimidelm(spacers)whichcontrolthewidthofthemicropore.A 132

PAGE 133

polycarbonatecontainerisplacedaroundthemicroporestructuretoholdMAFMasshowninFigure 4-32 (B).Thelowermicroporesidewallglassisbondedtothebaseviaanopticalqualityadhesive.Allothercomponentsarebondedwithepoxy;therefore,thisstructurecannotbedisassembledandnondestructiveroughnessmeasurementsofthemicroporesidewallsurfacesarenotfeasible.AnopticalmicrographofthemicroscopeviewintothemicroporeareaisshowninFigure 4-32 (C).VisibleinFigure 4-32 (C)aretheedgesofthetwostripsofpolyimidelmandtheundersideoftheuppersidewallwhichasagroundsurfacetexture.Thelowermicroporesidewallhasasmooth(0.3nmRq)texture. Figure4-32. Microporereplicationsetupforowvisualization 4.3.4.2VisualizationofowdrivenbyanalternatingmagneticeldUsingthemicroscopeandfrontalcameracongurationshowninFigure 4-19 ,thedynamicmotionoftheMAFMowthroughthemicroporeandsurfacemovementsoftheMAFMwereobservedunderanalternatingmagneticeldatvariousfrequencies.TheexactconditionsusedareshowninTable 4-6 .Foreachfrequency,thesupplied 133

PAGE 134

Table4-6. Alternatingcurrentowvisualizationconditions MagneticuidWater-based,anionicsurfactant3.6wt%Fe3O4200LAbrasiveslurry50nmwater-baseddiamondslurry0.5mdiamondslurry200LSuppliedMAFM400LSuppliedWaveformAlternatingcurrent1A1Hz,5Hz,10Hz,15Hz,20Hz,25Hz,30Hz,35Hz,40Hz,45Hz,50HzPole-polegap20.00mm voltagewassetsuchthatthemaximumcurrentwas1A.TwoseparateobservationswereperformedusingMAFMpreparedwith50nmwater-baseddiamondslurryandanotherwith0.5muniversalbasediamondslurry.ThemotionoftheMAFMmixtureswasrecordedandthevideosweresynchronizedforside-by-sideplayback.Sincelargevideoscannotbeincludedintothisdocument,themotionandnotableoccurrencesaredetailedinthefollowinglist.Forreference,selectedframesfromthecorrespondingvideosareshownintheappendix(Figures A-1 A-12 ). ZeroField(0Hz,0A) 50nmdiamondslurry Microscope:Afewrandomlyscatteredimperfectionsarevisibleatorneartheglassporesurface. FrontalCamera:AnequalamountofMAFMresidesonbothsidesoftheuppersectionoftheglassporestructure.AmeniscuscurvesthesurfaceoftheMAFM. 0.5mdiamondslurry Microscope:Small(1m)groupsofdiamondparticlesappeartobeuniformlydispersedthroughouttheuidinthemicropore. FrontalCamera:Thismixturehasasmallermeniscuseffectthanseeninthe50nmdiamondMAFM,suggestingthatthismixturehasloweradhesiveforce. 134

PAGE 135

1Hz 50nmdiamondslurry Microscope:Looselydenedhorizontalstripesofabrasiveslurryareformedontheactivationoftheeld.Theyarealignedinthedirectionofmagneticuxandmovetowardstheactivecoil.Theresponseoftheuidappearstobelinearwiththealternatingcurrentwaveform,andtheowappearslaminar.Observationofthestripesofabrasiveslurrysuggestthatindividualdiamondparticlesmovefastenoughtoenterandexitthemicroporeinasinglepass(1/2cycle). FrontalCamera:TheMAFMsurfaceonthesideclosertotheactivemagneticpoletiltstowardsit.TheMAFMsurfaceawayfromtheactivemagneticpoleexhibitsauidlevel1mmlowerthaninthezeroeldcase.ThereappearstobealinearresponseoftheMAFMsurfaceshape. 0.5mdiamondslurry Microscope:Theevenlydispersedparticlesappeartoreciprocatelinearlywiththealternatingwaveformanddonotformstripes.Individualparticlesappeartoreciprocateadistanceofapproximately400m. FrontalCamera:Theuidbehavessimilarlytothe50nmdiamondMAFM,however,theMAFMsurfaceawayfromtheactivemagneticpoleshowsslighttilting. 5Hz 50nmdiamondslurry Microscope:TheMAFMformsmoreclearlydenedstripesofanestimated5minwidth.Theabrasiveparticlesappeartoreciprocatelinearlywiththealternatingwaveform.Theiramplituderoughlyisestimatedtobe900m. FrontalCamera:ThebehavioroftheMAFMissimilartothatofthe1Hzcase;thoughslighttiltingoftheMAFMsurfaceawayfromtheactivemagneticpoleoccurs. 0.5mdiamondslurry Microscope:TheMAFMformsstripesofanestimated10minwidth.Theabrasiveparticlesappeartomovelinearlywiththealternatingwaveform.Theiramplitudeisroughlyestimatedtobe100m. 135

PAGE 136

FrontalCamera:ThebehavioroftheMAFMissimilartothatofthe1Hzcase. 10Hz 50nmdiamondslurry Microscope:Theabrasivestripesarethinnerthaninthe5Hzcase.Theirthicknessissimilartotheresolutionofthecamera(2.5m).Theirmovementappearstorespondlinearlywiththealternatingwaveform.Itisdifculttoestimatetheamplitudeofreciprocationsincethestartandendpointsoftheabrasivestripesarenotdiscernible. FrontalCamera:ThemagneticeldalternationisfasterthanthebulkrelaxationoftheMAFMatthispoint.TheshapeoftheMAFMsurfaceischaracterizedbyameniscuslikecurvecontactingtheglassmicroporereplicationstructureandasteepslopecontactingthepolycarbonatecontainerwalladjacenttothepoletip.VibrationsareseenontheMAFMsurface. 0.5mdiamondslurry Microscope:Thestripesofabrasivearethinner(5m)andreciprocatelinearlywiththealternatingwaveform.Someabrasiveparticlesappeartohavesettledontheporesidewall;theseparticlesalsoreciprocate.Theamplitudeofthereciprocationisapproximately100m. FrontalCamera:TheMAFMsurfaceissimilartothatofthe50nmdiamondMAFMcaseatthisfrequency,butthesurfacesarenotassteeplycurved. 15Hz 50nmdiamondslurry Microscope:SettledabrasivestripesarepresentthoughtheyarequicklydispersedbytheMAFMalternation.Theabrasiveparticlesappeartoreciprocatelinearlywiththealternatingwaveform. FrontalCamera:TheMAFMsurfaceissimilartothatofthe10Hzcase.Asecond-orderharmonicisclearlyvisibleonthevibratingMAFMsurface. 0.5mdiamondslurry Microscope:Theabrasiveparticlesstripesareshorterinlengthandreciprocatewithanamplitudeof50m. 136

PAGE 137

FrontalCamera:Secondandthird-orderharmonicsarevisibleonthesurfaceoftheMAFM. 20Hz 50nmdiamondslurry Microscope:Themicroscope'svideoframerateis30fps.Thestripesofabrasivearebarelyvisibleatthisfrequency,andanymotionobservedwouldbealiasedsince20HzisabovehalftheNyquistfrequency(15Hz). FrontalCamera:Thecamera'svideoframerateis30fps.Anymotionobservedislikelyaliased;howeverthesurfaceappearstohaveathird-orderharmonicvibration. 0.5mdiamondslurry Microscope:Asstatedabove,theobservedmotionwillbealiased.Sinceoneofthecoilscreatesaslightlystrongermagneticeld,thereciprocatingabrasiveparticlestrendtowardsthestrongermagnet. FrontalCamera:Asstatedabove,theobservedmotionwillbealiased.Giventhealiasing,theMAFMsurfaceappearstohaveafourth-orderharmonicvibration. 25Hz 50nmdiamondslurry Microscope:Stripesofabrasiveareclearlydenedandroughlyestimatedtobe<5mwide.RandomareasofsettledabrasivereciprocateanddispersewiththeMAFMmotion.Motionoftheabrasivesisstroboscopic. FrontalCamera:TheMAFMsurfacevibrationsappearslowduetoastroboscopiceffect. 0.5mdiamondslurry Microscope:Themotionissimilartothatofthe20Hzcase. FrontalCamera:TheMAFMsurfacevibrationsappearslowduetoastroboscopiceffect. 30Hz 50nmdiamondslurry 137

PAGE 138

Microscope:Stripesofabrasiveasimilarinsizetothoseofthe25Hzcase.Somesettledabrasiveareastrendtowardstheleftcoil,whileotherareastrendtowardstherightcoil.Theareaspassdirectlynexttoeachotheronoccasion. FrontalCamera:TheMAFMappearstostandperfectlystillduetoastroboscopiceffect. 0.5mdiamondslurry Microscope:Thevibrationoftheabrasiveparticlesappearstohavestoppedduetoastroboscopiceffect.Theonlymotionvisibleisatrendtowardsthestrongcoil. FrontalCamera:TheMAFMappearstostandperfectlystillduetoastroboscopiceffect.However,aslighlyhigher(0.3mm)uidlevelisseenontheMAFMsidenearthestrongercoil(seenontherightsideofFigure A-8 (Bii)). 35Hz 50nmdiamondslurry Microscope:Stripesofsettledabrasiveappear,thoughthenestripesseeninthelowerfrequenciesarepresentaswell. FrontalCamera:Themeniscusattheglassporestructureisalmostnotvisible.Someslow-movingvibrationsarevisible,thoughtheiramplitudeappearstobelessthan0.25mm. 0.5mdiamondslurry Microscope:Therightwards-trendingmotionoftheabrasiveparticlesisnotseen.Instead,acrossingmotionofsettledabrasive,whichissimilartothatofthe50nmMAFMat30Hz,isobserved.Theabrasiveparticlesappeartohavesomeamplitudeofreciprocation;however,thestroboscopiceffectpreventsacondentmeasurement. FrontalCamera:Themeniscusattheglassporestructureisalmostnotvisible.Someslow-movingvibrationsarevisible,thoughtheiramplitudeappearstobelessthan0.25mm. 40Hz 50nmdiamondslurry Microscope:Settledabrasivesstripesarethinandfaintyetmovinginseeminglyrandomdirections.Non-settledabrasivestripesarevisible. 138

PAGE 139

FrontalCamera:Fluidappearsnearlystill.FaintvibrationsareindicatedbyanoscillatingbrightnessattheMAFMsurface. 0.5mdiamondslurry Microscope:Stripesofsettledabrasives(10m)areseen.Theseexhibitacrossingmotioninwhichasinglestripewillshiftdirectlypastanadjacentstripemovingintheoppositehorizontaldirection. FrontalCamera:TheMAFMsurfaceappearstobealmostentirelystill. 45Hz 50nmdiamondslurry Microscope:Faintsettledabrasivestripesappeartomovetowardstheleft.Non-settledabrasivestripesarevisible. FrontalCamera:TheMAFMsurfaceappearstobeentirelystill. 0.5mdiamondslurry Microscope:Themotionisamorepronouncedversionofthemotionobservedinthe40Hzcase. FrontalCamera:TheMAFMsurfaceappearstobeentirelystill. 50Hz 50nmdiamondslurry Microscope:Theobservedabrasivemotionissimilartothe45Hzcase. FrontalCamera:TheMAFMsurfaceappearstobeentirelystill. 0.5mdiamondslurry Microscope:Theobservedabrasivemotionissimilartothe45Hzcase. FrontalCamera:TheMAFMsurfaceappearstobeentirelystill.TheseobservationsweretakenwithMAFMowrestrictedtoasinglemicropore.Inmicroporeopticnishing,hundredsofmicroporesareetchedsidebyside,increasingthepotentialowratepercycle.Thus,itisthoughttheamplitudeoftheowreciprocationislesslimitedthaninthissingle-microporecaseduetolessresistancetoow;therefore,nishingwithhigherreciprocationfrequenciesispossiblewithactualmicroporeoptics. 139

PAGE 140

ThehighermodesofvibrationsseenontheMAFMsurfaceviathefrontalcameraarecausedbytheMAFMsinertia.TheseobservationsclearlyindicatethatMAFMpreparedwith50nmdiamondabrasiveslurrysimplyhasagreaternumberofabrasiveparticlesavailableforcuttingthandoesMAFMpreparedwith0.5muniversalbasediamondslurry.TheMAFMpreparedwith50nmdiamondabrasiveslurryappearstohavelessresistancetoow,whichexposesthemicroporesidewallstohighervelocityabrasiveimpactsthanMAFMpreparedwith0.5muniversalbasediamondslurry.Thesearepotentialreasonsforwhichthe50nmdiamondslurryoutperformedthe0.5mdiamondslurryintheexperimentsofSection 4.3.2 .Overall,theobservationsdetailedinthissectionindicatethatthereciprocationamplitudeofabrasiveparticlesinMAFMthroughamicroporereduceswithincreasingalternationfrequency.Atfrequenciesabove5Hz,thereciprocationamplitudeappearstobesmallerthanthedepthofthemicropore,whichmaylimittheprocess'machiningcapabilitybylimitingthenumberofabrasiveparticlesavailableforsurfacenishing.Inadditiontotheappliedmagneticeld,theabrasiveparticlereciprocationamplitudeisalsolikelydependentontheMAFMowresistancethroughthepore,whichdependsontheporegeometryanduidproperties. 4.3.4.3VisualizationofowdrivenbyswitchingdirectcurrentInthecourseofdeterminingsuitablesustainedMAFMowconditionsforthenishingtrialsofSection 4.3.2 ,videoobservationsofowinthatsetupindicatedaneedforlowalternationfrequencies;therefore,aswitchingDCcurrentwasusedtocreatetheswitchingmagneticeld.ThissuppliedcurrentcreatedahighspeedandlaminarMAFMow,whichisthoughttohaveyieldedastrongersurfacenishingeffectthanthatcreatedbyalternatingfrequenciesof1Hzorabove.ToseeifasimilartypeofowwoulddevelopinthenishingsetupofFigure 4-32 ,observationssimilartothoseoftheprevioussectionweremadeofthedynamicmotionoftheMAFMowthroughthemicroporeandsurfacemovementsoftheMAFMunderaswitchingmagneticeld 140

PAGE 141

Table4-7. Switchingdirectcurrentowvisualizationconditions MagneticuidWater-based,anionicsurfactant3.6wt%Fe3O4200LAbrasiveslurry50nmwater-baseddiamondslurry0.5mdiamondslurry200LSuppliedMAFM400LSuppliedWaveformAlternatingsquarewave+25.00V)]TJ /F1 11.955 Tf 9.3 0 Td[(26.00V15speriodPole-polegap20.00mmMagneticuxdensityatmicropore145mT drivenbytheswitchingDCcurrent.Table 4-7 detailstheconditionsusedfortheseobservations.TwoseparateobservationswereperformedusingMAFMpreparedwith50nmwater-baseddiamondslurryandanotherwith0.5muniversalbasediamondslurry.Figure 4-33 showstwoframesofmicroscopevideostakenatlowandhighmagnicationsandtheircorrespondingfrontalcameravideoframesoftheowofMAFMpreparedwith50nmwater-baseddiamondslurry.Theabrasiveparticlestripesformwhenacoilisactivated,andahighvelocityowdevelops.Theowappearstobestrictlylaminar;itisdifculttoconrmtheexistenceofaboundarylayerfromthemicroscopeimagessinceallvisibleabrasiveparticlestripesappeartomoveatasimilarrate.ThefrontalimageclearlyshowsadropinMAFMlevelof1mm(inthesideoppositeoftheactivecoil)whichoccursduringa7.5shalf-cycle.This1mmdroproughlyrelatestoapproximately119L;whichresultsinavolumetricowrateofq50nm=15.9L/s.Assumingainviscidow,theparticlevelocityisdescribedbyv=q=(wl)wherevisthevelocity,wistheporewidthof25.4mandlisthelengthofthemicroporeof2.25mm.Theparticlevelocityisthenestimatedtobev50nm=277.5mm/sor16.6m/min. 141

PAGE 142

Figure4-33. StillvideoframesofMAFMpreparedwith50nmwater-baseddiamondslurry Figure 4-33 showstwoframesofmicroscopevideostakenatlowandhighmagnicationsandtheircorrespondingfrontalcameravideoframesoftheowofMAFMpreparedwith0.5muniversalbasediamondslurry.Abrasiveparticlestripesformwhenacoilisactivatedandahighvelocitylaminarowdevelops.ThefrontalimageclearlyshowsadropinMAFMlevelof0.6mm(inthesideoppositeoftheactivecoil)whichoccursduringa7.5shalf-cycle.Thisdroprelatestoapproximately71L;whichresultsinanestimatedvolumetricowrateofq0)]TJ /F6 7.97 Tf 6.58 0 Td[(0.5m=9.5L/sandan 142

PAGE 143

estimatedparticlevelocityofv0)]TJ /F6 7.97 Tf 6.59 0 Td[(0.5m=166.5mm/sor10m/min.Frame-by-frameobservationofthevideosunfortunatelydoesnotpermitthetrackingofsingleparticles(tocomparetheirvelocity)intheMAFMbecausetheyappearonasingleframeandaregonebythenextframe(atthecalculatedspeed,theabrasiveparticlestravel5.5mmper1/30sframe).TheseobservationsoftheMAFMowthroughamicroporeestablishedthefeasibilityofthismethodtoidentifytheMAFMowregimeandtoquantifycertainprocessparameterssuchasMAFMowrate.Futureeffortscouldemployhigherresolutionandhigherframeratevideoequipment. 4.4DeliberationsontheSurfaceModicationMechanism 4.4.1NanoscaleStructuresinMagneticFluidsMagneticuidsconsistofasuspensionofnanoscalemagneticnanoparticlesinaliquid.Magneticnanoparticlesinmagneticuidsaretypicallysphereswithdiametersrangingfrom4to50nm,dependingonhowtheyaresynthesized;commerciallyavailablemagneticuidscontain10nmmagneticnanoparticlesmadebychemicalprecipitation.Magneticnanoparticlesarecommonlycomposedofcobaltandironcompoundssuchasmaghemiteandmagnetite.Topreventaggregationofmagneticnanoparticles,theyarecoatedina2nm-thicksurfactant(repulsive)layer.Suchsmallnanoparticlesaretypicallycrystalsexhibitingasinglemagneticdomain.Thealignmentoftheirelectronspinsformamagneticdipolemoment.Whenwelldispersedinaliquidcarrier,magneticparticlesarefreetorotateandtranslate.Thestructureandbehaviormagneticnanoparticlesinazeroeld(noexternalmagneticeldapplied)isaffectedbymultiplefactorsincludingtheamountofthermalenergy,vanderWaalsattraction,thethicknessandelectro-chemicalpropertiesofthesurfactant,andtheparticlesizeanddistribution[ 112 113 ].Ithasbeenobservedthatdipoleinteractionbetweenmagneticnanoparticlesisminimalforparticlessizesof16nmorless[ 95 114 ].Themagneticdipolemomentsofparticleslargerthan16nmisusuallystrongenoughtoovercomerepulsionfromthesurfactantlayers,resultingintheformationofweaklybound 143

PAGE 144

Figure4-34. StillvideoframesofMAFMpreparedwith0.5muniversalbasediamondslurry structuresconsistingofmultiplemagneticnanoparticles.Thesestructuresmaybeintheformofchains,rings,bent-chainloops,andothers[ 95 114 ].Undertheinuenceofastrong(>0.1T)externalmagneticeld,themagneticdipolesaligninthedirectionofuxandtheparticlesattract,formingminimum-energyparticlecongurations.Largestructures,bestdescribedasbrouschains,whichconsistofnanoparticleclusters,chains,andux-closureringsinalong(>1000nm)massthatis500nmwide[ 95 115 ].Theexistenceofux-closureringswas 144

PAGE 145

predictedin1963byJacobsandBean[ 116 ]andrstobservedin2002byPhilipseetal.[ 117 ].Flux-closureringsformwhenbothendsofachainofparticlesmeet,closingthemagneticcircuit.Theringstructureischiralandproducesnonetmagneticmoment,minimizingitsenergy.Flux-closureringshavebeenobservedinvariousimagingstudieswithmagneticnanoparticles[ 115 118 119 ]. 4.4.2PossibleInteractionbetweentheMagneticNanoparticleStructureandAbrasiveParticlesAfewstudiesdiscusstheinteractionbetweenlarge(1m)nonmagneticparticleswithmagneticuids[ 120 121 ].Thesestudies,however,treatmagneticuidsasacontinuum.Nostudiesareknowntotheauthorinwhichthespecicnanoscalearrangementofmagneticnanoparticlesandnonmagneticparticlesisdirectlyconsidered.Thus,basedontheavailableliteratureofthenanoscalestructureofmagnetizedmagneticuids,atheoryregardingapossiblearrangementofabrasiveparticlesandmagneticparticlesinanexternallyappliedmagneticeldispresentedinthissection.MAFMisamixtureofmagneticuidandabrasiveslurry.Inazeroeld,theMAFMmicroscopicallyisthoughttobeasuspensioncontainingabrasiveparticlesandsurfactantcoatedmagneticnanoparticles.Throughoutthisresearch,10nmFe3O4particleshavebeenused,whicharesmallenoughtoremainevenlydispersedinazeroeld.Iftheabrasiveparticlesaremuchlargerinsizetothemagneticnanoparticle(>500nm),themagneticnanoparticleswouldformbrouschainstructuresaroundtheabrasiveparticleswhenanexternaleldisapplied.Magneticallyinducedmovementofthelargeabrasiveparticleswouldbepossible,asinstudieswithlargenonmagneticparticles[ 120 ].Inthatcase,apossiblenishingmechanismforthenishingprocessofthisresearchcouldbescratchingduetolaminarowofabrasivessuchascuttingmarkscreatedbyoatpolishing[ 122 ];however,nosuchmarkshaveeverbeenobservedinthisresearch.Additionally,nishingstudiesusingMAFMpreparedwith 145

PAGE 146

largeabrasives(e.g.,6mwater-baseddiamond)yieldednoeasilydiscerniblesurfacemodications(Figure 4-25 );thoughapreliminaryexperimentdidresultinanX-rayreectanceimprovement(Section 3.3.1 ).Apossibleabrasiveparticlemagneticnanoparticleinteractionscenarioinvolvesthegraspingofsuspendedabrasiveparticlesduringtheformationofnanoparticlestructuresupontheactivationoftheexternalmagneticeld.AmagneticuidnanostructuresuchasthatimagedbyWuetal.[ 115 ]iscomposedofhollowring-likestructureswhichresembleux-closurerings.Ifabrasiveparticlesareevenlydispersedwiththemagneticparticles,theseabrasiveparticlescouldoccupythehollowstructuresofthebrousnanoparticlechain.Clearly,thesizeoftheabrasiveparticlesrelativetothediameteroftheringstructuresisofgreatimportanceinthiscase.Asmentionedabove,underanappliedeld,magneticparticlesformstructuresinenergeticallyfavorablecongurations.UsinganapproachpublishedbyClarkeandPateyin1994[ 123 ],Lavenderetal.calculatedthepotentialenergyoflinearandringcongurationsoftheoreticalpointdipolesseparatedbyadistancewithadipolemagnitudeofversusthenumberofdipolesnintheconguration[ 124 ].Theyfoundthatring-typedipoleparticlecongurationsofn4havealowerpotentialenergiesthanlinearchainsofdipoles.Thedifferenceisminorandasymptoticalhowever,andforn=1,thedifferenceconvergestozero.Thesizeofaringstructureisdeterminedbythenumberofparticlescomprisingit.Assumingthedipolarmagneticnanoparticlesaggregatetoformaregularpolygonattheircenters,theinnerdiameterofaringofnparticlesofknownradiuscanberoughlyestimatedbycalculatingthepolygon'sapothem2.InthisresearchtheFe3O4nanoparticlediameteris10nm,andthesurfactantlayerthicknessisestimatedtobe3nm,yieldingatotaldiameterof16nmperparticle.Figure 4-35 plotsseveralroughlyestimatedinnerringdiametersforringsofvaryingnumbersofdipolarparticles.Additionally,Figure 4-35 plotstheenergydifferencebetweenthelinearchainandring 146

PAGE 147

congurationsfor4n20dipolarparticlesfromthevaluespublishedbyLavenderetal.[ 124 ].ThedottedlineinFigure 4-35 indicatesavisualapproximationofanasymptotethateventuallyconvergestozero. Figure4-35. Approximatevaluesfortheinnerdiameterofdipolarparticleringsandthedifferenceinpotentialenergybetweenlinearchainsandringcongurationsofdipoles Ringdiameterisnottypicallydirectlymeasuredbyresearcherswhohaveimagedsuchfeatures;however,carefulobservationoftheirpublishedimagesofthesestructuresrevealsthatthesizeoftheseringstructuresroughlyrangesbetween20and300nm[ 114 115 117 ].Thisrangeofringdiametershappenstocorrespondtotherangeofabrasiveparticlediameterswhichyieldedthelargestchangesinsurfaceroughness(Figure 4-26 ).ThelargestchangeinsurfaceroughnesswashadbyusingMAFMpreparedwith50nmdiamondabrasiveslurry,asseeninFigure 4-26 (B).Figure 4-35 showsthatringsof50nmdiameterhavenearlythelargestdifferenceinpotentialenergywithlinearchainsofthesamenumberofparticles(n=10).Ithasbeenreportedthathigherconcentrationofparticlesyieldsshorterdipolechains[ 115 ];thoughthisrelationshipisempiricalanddoesnotappeartohavebeenquantied.Perhapstherelativelyhighconcentrationsofmagnetiteparticlesusedinthisresearchlimitthelengthofchainstructurestosuchalengththatringsarethemoreenergeticallyfavorable 147

PAGE 148

particlestructure.Conrmationoftheoftheexistenceofthisabrasiveparticlemagneticnanoparticleinteractionwouldlikelyrequiredirectimagingviacryogenictransmissionelectronmicroscopyacomplexprocedure;however,suchobservationswouldprovideinvaluableinsightsintothefundamentalworkingsofthisnishingprocess. 4.4.3IndentationasaSurfaceModicationMechanismIftheabrasiveparticlemagneticnanoparticleinteractiontheoryis,infact,theprimaryabrasiveparticleactuationmechanism,thenishingprocesssurfacemodicationmechanismwouldbeclariedsomeextent.Inthiscase,theactivationofthemagneticeldinducesmotionofthemagneticparticlesandthebrouschainformationinwhichabrasiveparticlesaregraspedbyringstructures.Thenowmovingbrouschainsowathighspeeds(seeSection 4.3.4.3 )throughthemicroporestructures.Exposedabrasiveparticleedgesgraspedbychainsowingadjacenttothemicroporesidewallsmayencountersurfaceasperitiesonthesidewalls.Contactbetweenhighspeedabrasiveparticlesandmicroporesidewallsurfaceasperitieswouldresultinarelativelylargeimpulseforce.Thisimpulseforcewouldlikelybefargreaterthantherupturetensionofthemagneticparticleringstructure,whichisestimatedtobe<20pN[ 121 ];thus,contactbetweenanabrasiveparticlegraspedbyamagneticnanoparticlestructureandasurfaceasperitywouldresultininstantdislodgementofthesurfaceasperityfromthenanoparticlestructure.Theabrasiveparticlewouldthereforeonlyhavecontactedthesurfaceasperityonce;iftheabrasiveparticlevelocityisgreatenough,theabrasiveparticlewouldindentthesurface.Figure 4-36 (A)isanexampleoftheappearanceofashallownanoindentationcharacterizedbyarelativelylargevolumeofpile-uparoundtheindentduetoplasticowofsiliconmaterial.Figure 4-36 (B)isathree-dimensionalAFMimageofasmoothSi(100)thathasbeenindentedwithapolycrystallinediamondcube-cornernanoindenterthathasa32nmradiustip.Thetipapproachedthesiliconsurfaceat100nm/suntiltheinstrumentdetectedcontactatwhichitcontinuedtoforcethetipintothesiliconsurface 148

PAGE 149

ata200N/srate.Oncetheinstrumentmeasuredaforceof50N,itretractedthetip,leavinganindentationwitharesidualdepthof3nm.Theseconditionscreatedananoindentationwithpiled-upmaterialsimilarinsizetothefeaturesobservedinFigures 4-28 (Ciii)and 4-29 (C);acomparisonisshowninFigure 4-36 (C).ItisthoughtthatthenishedsurfacetextureshowninFigures 4-28 (Ciii)and 4-29 (C)wasgeneratedbymanyrandomhigh-speedindentationsthat,duetotheirlowpenetrationdepthandhighplasticityofthemetalworkpiecematerial,exhibitlargeamountsofpileup.ThefeaturesonthesurfaceofFigures 4-28 (Ciii)and 4-29 (C)havecomparableheight,width,anddepthtotheindentshowninFigure 4-36 Figure4-36. AFMimageofsingle3nm-deepnanoindentationinsilicon(100)(DatausedwithpermissionfromJaredN.Hann) Theforcesrequiredtomakesuchshallowindentfeaturesonthebrassmaterialshouldbelessthanthe50NusedtocreatetheindentofFigure 4-36 onsilicon(VickershardnessHV:9500MPa),sincethebrassmaterialissoft;though,atthissmallscale,thesurfaceismostlikelycomposedofagrainofcopper(HV:369MPa). 149

PAGE 150

Theclosestforcevaluesinliteratureforsuchlowindentationdepthsknowntotheauthorisa3nmindentinsinglecrystalaluminum(HV:167MPa)createdbya15Npeakload[ 125 ].Consideringtheabove,theforcesinvolvedinthecreationoftheseindent-likefeaturesmayliebetween15Nand50N.Ananoindentationstudyonasimilarsmoothbrassworkpiecetoattainrelevantload-depthplotscombinedwithAFMimagingoftheseindentswouldassistinthecharacterizationofthenanoscaledeformationmechanismsofthenishingprocessonthisparticularworkpiecematerial. 150

PAGE 151

CHAPTER5CONCLUSIONS 5.1ConcludingStatementsThedevelopmentofX-rayopticsiscloselytiedwithgrowthintheeldofX-rayastronomy.MEMSmicroporeX-rayopticswerepresentedasalightweightandlowcostalternativeexistingX-rayopticsystems;however,manufacturingchallengescurrentlylimittheirfullrealization.Specically,thesurfacequalityofmicroporesidewallsneedsimprovement.TheconceptandbackgroundofMAFwaspresentedalongwithexamplesofMAF'sapplicabilitytooptics,MEMSdevices,andinternalsurfacenishing.AnMAFprocesswasthenconceivedtonishthemicroporesidewallsurfacesofMEMSmicroporeX-rayoptics.Twonishingmachineswerebuilttorealizetheconceivedprocessontwodifferenttypesofworkpieces,MEMSmicroporemirrorchipsandfull-sizedMEMSmicroporeX-rayoptics.Todate,severalmirrorchipnishingexperimentshavebeenperformed.Additionally,carefullyexecutednishingexperimentshavebeencarriedouttoobservethenishingprocess'characteristicsurfacemodicationsandtoclarifythecorrespondingsurfacemodicationmechanism.Theresultsofthisresearchmaybesummarizedasfollows. 1. Analternatingmagneticeld-assistednishingprocesswasconceivedtonishthesidewallsoftheMEMSmicroporeX-rayoptics,andmachineswerebuilttorealizetheprocess.ThenishingefciencyofthisprocessappearstodependontheamountofdynamicMAFMmotion,whichiscontrolledbythestrength,frequency,andtransientbehavioroftheexternallyappliedmagneticeld. 2. Amachinewasbuiltwhichiscapableofrealizingthenishingprocessonmirrorchips.Anothermachinewasdesignedandbuilttonishfull-sized4MEMSmicroporeX-rayoptics,thoughithasnotbeentestedyet. 3. Severalexperimentshavedemonstratedthenishingprocess'abilitytoreducethesurfaceroughnessofmirrorchips;aroughnessof0.2nmRqwasachievedonthemicroporesidewallssurfaceofahydrogenannealedDRIE-fabricatedsiliconmirrorchip. 151

PAGE 152

4. X-rayreectancetestssuggestthattheX-rayreectionperformanceofmicroporesidewallsofnishedmirrorchipsisimprovedbytheapplicationofthenishingprocess. 5. Thenishingprocessappearstoremovematerialfromthetargetsurfacesatsub-nanometerincrements.Onatworkpiecenishingexperiments,nishedsurfaceprolesshowonlyslightsurfacemodications.Highresolutionimagesoffeaturesbeforeandafternishingonahydrogenannealedmirrorchipshowedtheremovalof1msizedsurfacefeaturesin10hrofnishing. 6. Varioussurfaceproleobservationssuggestthattheformandwavinessofnishedsurfacesisnotaffectedbythenishingprocess.Theform-followinguidnatureandlownishingforcesappliedbytheMAFMduringnishingislikelyresponsibleforthischaracteristic. 7. Non-uniformremovalofsurfaceasperitieswasobservedinvariousmirrorchipnishingexperimentsandinaatworkpiecenishingexperiment;specically,relativelylowwavelengthsurfaceasperitiesareremovedbythenishingprocess.Thisselectiveremovalofsurfacefeaturesischaracteristicofprocessesinwhichmechanicalmaterialremovaloccurs. 8. Observationsofthesurfaceandsubsurfacecompositionindicatethatthenishingprocessdoesnotleavesurfaceresiduesorimpregnatedmaterials. 9. ExperimentsinwhichMAFMowwasrestrictedtoasinglemicroporeindicatedthatdiamondabrasivesof50nm,0.2m,and0.5mwereeffectiveatcausingsurfaceroughnessreductionsonatbrassworkpieces. 10. VideoobservationsofMAFMowinsideasinglemicroporerevealthattheparticlevelocitiesinamicroporecanrangebetween150and300mm/sinanon-varyingmagneticeld.SimilarobservationsofMAFMowactuatedbyanalternatingmagneticeldsuggestthatparticlereciprocationamplitudesarehighlydependentonalternatingfrequency.Itappearsthatanupperlimitonfrequencyexistsbeyondwhichnishingisnotexpectedtooccur;thislimitappearstodependontheworkpiecesoverallowresistance(e.g.,uidproperties,geometryandnumberofmicropores). 11. High-resolutionAFMimagingofapreviously-smoothedbrassworkpiecenishedbyMAFMpreparedwith50nmdiamondabrasivesuggeststhattheprocessgeneratedasurfacetexturethoughttohavebeencreatedbyrandomrepeatedindentation.Consideringthenatureofthenanoscalebehaviorofmagneticuid,atheoryisproposedonalikelyabrasiveparticleactuationmechanism,whichinvolvesthegraspingofsuspendedabrasiveparticlesduringtheformationofnanoparticlestructuresupontheactivationoftheexternalmagneticeld.Thismechanism,iftrue,wouldresultinthecreationofindentsontheworkpiece 152

PAGE 153

surface.Resultsfromsingle-microporenishingexperimentsonbrassworkpiecessupportthistheory. 5.2FutureWorkTheworkpresentedinthisdissertationrepresentsafoundationofanewnishingprocessandrelevantcharacterizationmethods.Machineswerebuilt,andtheprocessfeasibilitywasdemonstrated.Thereareseveralavenuestowardswhichresearcheffortscouldbeaimed.Althoughthefull-sizedopticnishingmachinewasbuilt,nishingoffull-sizedopticshasnotyetbeenattempted.Thediscoveryandoptimizationofparameterswithwhichthesefull-sizedopticsshouldbenishedmustbeexecuted.Experimentsinwhichasinglemicroporewasreplicatedyieldedpromisingresults.Furtherresearchinthisareawouldinvolvethenishingofsoftsingle-crystalmetalworkpiecessuchthatthemorphologicalchangestotheworkpiecesurfacearefreeofinuencefromanalloysgrainstructure.Additionally,experimentsinwhichvariablessuchastheMAFMchemicalconstituentsandtheabrasiveslurrychemistryarecarefullycontrolledwouldyieldmorepreciseresults.Researchexistsintheliteraturewherethenanoscalestructureofnanoparticlesinmagneticuidareimagedviacryogenictransmissionelectronmicroscopy;similarobservationsofthestructureofmagnetizedMAFMwouldcertainlyproveuseful.Additionally,higher-framerateobservationsofMAFMowandsurfaceshapewouldallowforprecisecharacterizationofthedynamicmotionofabrasiveparticlesandadeeperunderstandingofthesurfacemodicationmechanism. 153

PAGE 154

APPENDIX:MAFMFLOWUNDERANALTERNATINGMAGNETICFIELD FigureA-1. MAFMthroughamicroporeinazeroeld. 154

PAGE 155

FigureA-2. MAFMthroughamicroporeinanalternatingmagneticeldat1Hz. FigureA-3. MAFMthroughamicroporeinanalternatingmagneticeldat5Hz. 155

PAGE 156

FigureA-4. MAFMthroughamicroporeinanalternatingmagneticeldat10Hz. FigureA-5. MAFMthroughamicroporeinanalternatingmagneticeldat15Hz. 156

PAGE 157

FigureA-6. MAFMthroughamicroporeinanalternatingmagneticeldat20Hz. FigureA-7. MAFMthroughamicroporeinanalternatingmagneticeldat25Hz. 157

PAGE 158

FigureA-8. MAFMthroughamicroporeinanalternatingmagneticeldat30Hz. FigureA-9. MAFMthroughamicroporeinanalternatingmagneticeldat35Hz. 158

PAGE 159

FigureA-10. MAFMthroughamicroporeinanalternatingmagneticeldat50Hz. FigureA-11. MAFMthroughamicroporeinanalternatingmagneticeldat45Hz. 159

PAGE 160

FigureA-12. MAFMthroughamicroporeinanalternatingmagneticeldat50Hz. 160

PAGE 161

REFERENCES [1] Rntgen,W.C.,1896,OnaNewKindofRays,Nature,58(1369),pp.274. [2] Seward,F.,andCharles,P.,2010,ExploringtheX-rayUniverse,CambridgeUniversityPress,NewYork,USA. [3] Brandon,D.,andKaplan,W.D.,2008,MicrostructuralCharacterizationofMaterials,JohnWiley&Sons,Ltd.,WestSussex,England. [4] Elder,F.,Gurewitsch,A.,Langmuir,R.,andPollock,H.,1947,RadiationfromElectronsinaSynchrotron,PhysicalReview,71(11),pp.829. [5] Friedrich,W.,Knipping,P.,andLaue,M.,1913,InterferenzerscheinungenbeiRntgenstrahlen,AnnalenderPhysik,346(10),pp.971. [6] Compton,A.,1923,TheTotalReexionofX-rays,PhilosophicalMagazine,45(270),pp.1121. [7] Friedman,H.,1962,SolarObservationsObtainedFromVerticalSounding,ReportsonProgressinPhysics,25,pp.164. [8] Giacconi,R.,Gursky,H.,Paolini,F.R.,andRossi,B.B.,1962,EvidenceforX-RaysFromSourcesOutsidetheSolarSystem,PhysicalReviewLetters,9,pp.439. [9] Archer,F.S.,1851,OntheUseofCollodioninPhotography,TheChemist:AMonthlyJournalofChemicalPhilosophy,andofChemistryAppliedtotheArts,Manufactures,Agriculture,andMedicineandRecordofPharmacy,2,pp.257. [10] Tipler,P.A.,andMosca,G.,1995,PhysicsforScientistsandEngineers,Volume2,WorthPublishers,NewYork,USA. [11] Spiller,E.,2000,X-RayOptics,Proceedingsofthe49thAnnualDenverX-rayConferenceAdvancesinX-rayAnalysis,42,pp.297. [12] Bass,M.,2010,HandbookofOpticsVolumeV,McGraw-Hill,NewYork,USA. [13] Murata,K.,andTanaka,H.,2010,Surface-WettingEffectsontheLiquidLiquidTransitionofaSingle-ComponentMolecularLiquid,NatureCommunications,1,pp.1. [14] Bass,M.,1995,HandbookofOpticsVolumeI,McGraw-Hill,NewYork,USA. [15] Snigirev,A.,Kohn,V.,Snigireva,I.,andLengeler,B.,1996,ACompoundRefractiveLensforFocusingHigh-energyX-rays,Nature,384(6604),pp.49. [16] Bragg,W.,1913,TheDiffractionofShortElectromagneticWavesbyaCrystal,ProceedingsoftheCambridgePhilosophicalSociety,17,pp.43. 161

PAGE 162

[17] Kirkpatrick,P.,andBaez,A.,1948,FormationofOpticalImagesbyX-rays,JournaloftheOpticalSocietyofAmerica,38(9),pp.766. [18] Abbe,E.,1879,berdieBedingungendesAplanatismusderLinsensysteme,SitzungsberichteerJenaischeGesellschaftfrMedicinundNaturwissenschaft,129,pp.129. [19] Wolter,H.,1952,SpiegelsystemeStreifendenEinfallsalsAbbildendeOptikenfrRntgenstrahlen,AnnalenderPhysik,445(1-2),pp.94. [20] Mangus,J.,andUnderwood,J.,1969,OpticalDesignofaGlancingIncidenceX-rayTelescope,AppliedOptics,8(1),pp.95. [21] Saha,T.,1987,GeneralSurfaceEquationsforGlancingIncidenceTelescopes,AppliedOptics,26(4),pp.658. [22] Giacconi,R.,Branduardi,G.,Briel,U.,Epstein,A.,Fabricant,D.,Feigelson,E.,Forman,W.,Gorenstein,P.,Grindlay,J.,Gursky,H.,etal.,1979,TheEinstein(HEAO2)X-rayObservatory,TheAstrophysicalJournal,230,pp.540. [23] Aschenbach,B.,1988,Design,Construction,andPerformanceoftheROSATHigh-resolutionX-rayMirrorAssembly,AppliedOptics,27(8),pp.1404. [24] Weisskopf,M.,2003,TheChandraX-RayObservatory:AnOverview,AdvancesinSpaceResearch,32(10),pp.2005. [25] Wynn,J.,2000,Conguration,Assembly,andTestoftheChandraX-rayTelescope,OpticalFabricationandTesting,OpticalSocietyofAmerica. [26] Aschenbach,B.,Briel,U.,Haberl,F.,Bruninger,H.,Burkert,W.,Oppitz,A.,Gondoin,P.,andLumb,D.,2000,ImagingPerformanceoftheXMM-NewtonX-rayTelescopes,ProceedingsofSPIE,4012,p.731. [27] Mitsuda,K.,etal.,2007,TheX-rayObservatorySuzaku,AstronomicalSocietyofJapan,59(I),p.1. [28] Zhang,W.,2009,ManufactureofMirrorGlassSubstratesfortheNuSTARMission,ProceedingsofSPIE,7437,p.74370N. [29] Bavdaz,M.,Peacock,A.,Beijersbergen,M.,andParmar,A.,2002,DevelopmentofX-rayOpticsatESA,ProceedingsofSPIE,4496,p.94. [30] Fraser,G.W.,etal.,2010,TheMercuryImagingX-raySpectrometer(MIXS)onBepiColombo,PlanetaryandSpaceScience,58(1-2),pp.79. [31] Beijersbergen,M.,Kraft,S.,Gunther,R.,Mieremet,A.,Collon,M.,Bavdaz,M.,Lumb,D.,andPeacock,A.,2004,SiliconPoreOptics:NovelLightweightHigh-resolutionX-rayOpticsDevelopedforXEUS,ProceedingsofSPIE,5488,p.868. 162

PAGE 163

[32] Tucker,W.,andTucker,K.,2001,RevealingtheUniverse:TheMakingoftheChandraX-rayObservatory,HarvardUniversityPress. [33] Petre,R.,andSerlemitsos,P.,1985,ConicalImagingMirrorsforHigh-speedX-rayTelescopes,AppliedOptics,24(12),pp.1833. [34] Schmidt,W.,1975,AProposedX-rayFocusingDevicewithWideFieldofViewforUseinX-rayAstronomy,NuclearInstrumentsandMethods,127(2),pp.285. [35] Angel,J.,1979,LobsterEyesasX-rayTelescopes,TheAstrophysicalJournal,233,pp.364. [36] Wilkins,S.,Stevenson,A.,Nugent,K.,Chapman,H.,andSteenstrup,S.,1989,OntheConcentration,Focusing,andCollimationofX-raysandNeutronsUsingMicrochannelPlatesandCongurationsofHoles,ReviewofScienticInstruments,60(6),pp.1026. [37] Peele,A.,1999,InvestigationofEtchedSiliconWafersforLobster-eyeOptics,ReviewofScienticInstruments,70,p.1268. [38] Ezoe,Y.,Koshiishi,M.,Mita,M.,Mitsuda,K.,Hoshino,A.,Ishisaki,Y.,Yang,Z.,Takano,T.,andMaeda,R.,2006,MicroporeX-rayOpticsUsingAnisotropicWetEtchingof(110)SiliconWafers,AppliedOptics,45(35),pp.8932. [39] Ezoe,Y.,Koshiishi,M.,Mita,M.,Mitsuda,K.,Hoshino,A.,Ishisaki,Y.,Takano,T.,andMaeda,R.,2007,SiliconMicro-poreX-rayOptics,NuclearInstrumentsandMethodsinPhysicsResearchSectionA:Accelerators,Spectrometers,DetectorsandAssociatedEquipment,579(2),pp.817. [40] Ezoe,Y.,Koshiishi,M.,Mita,M.,Maeda,Y.,Mitsuda,K.,Osawa,T.,Suzuki,M.,Hoshino,A.,Ishisaki,Y.,Takano,T.,etal.,2008,MicromachinedX-rayCollectorforSpaceAstronomy,SensorsandActuatorsA:Physical,145,pp.201. [41] Mitsuishi,I.,Ezoe,Y.,Koshiishi,M.,Mita,M.,Maeda,Y.,Yamasaki,N.,Mitsuda,K.,Shirata,T.,Hayashi,T.,Takano,T.,etal.,2010,EvaluationoftheSoftX-rayReectivityofMicroporeOpticsUsingAnisotropicWetEtchingofSiliconWafers,AppliedOptics,49(6),pp.1007. [42] Mitsuishi,I.,Ezoe,Y.,Takagi,U.,Mita,M.,Riveros,R.E.,Yamaguchi,H.,Kato,F.,Sugiyama,S.,Fujiwara,K.,Morishita,K.,etal.,2009,NovelUltra-lightweightandHigh-resolutionMEMSX-rayOptics,ProceedingsofSPIE,7360,pp.73600C1C9. [43] Ezoe,Y.,Mitsuishi,I.,Takagi,U.,Koshiishi,M.,Mitsuda,K.,Yamasaki,N.,Ohashi,T.,Kato,F.,Sugiyama,S.,Riveros,R.,etal.,2010,UltraLight-weightandHigh-resolutionX-rayMirrorsUsingDRIEandX-rayLIGATechniquesforSpaceX-rayTelescopes,MicrosystemTechnologies,16(8),pp.1633. 163

PAGE 164

[44] Jansen,H.,Gardeniers,H.,Boer,M.,Elwenspoek,M.,andFluitman,J.,1996,ASurveyontheReactiveIonEtchingofSiliconinMicrotechnology,JournalofMicromechanicsandMicroengineering,6,p.14. [45] Laermer,F.,andSchilp,A.,2003,MethodofAnisotropicEtchingofSilicon,Mar.11USPatent6,531,068. [46] Ishikawa,Y.,Ikeda,N.,Kenmochi,M.,andIchinokawa,T.,1985,UHV-SEMObservationsofCleaningProcessandStepFormationonSilicon(111)SurfacesbyAnnealing,SurfaceScience,159(1),pp.256. [47] Kuribayashi,H.,Hiruta,R.,Shimizu,R.,Sudoh,K.,andIwasaki,H.,2003,ShapeTransformationofSiliconTrenchesDuringHydrogenAnnealing,JournalofVacuumScience&TechnologyA:Vacuum,Surfaces,andFilms,21,p.1279. [48] Kitahara,K.,andUeda,O.,1994,ObservationofAtomicStructurebyScanningTunnelingMicroscopyofVicinalSi(100)SurfaceAnnealedinHydrogenGas,JapaneseJournalofAppliedPhysics,33(11B),pp.L1571L1573. [49] Kuribayashi,H.,Gotoh,M.,Hiruta,R.,Shimizu,R.,Sudoh,K.,andIwasaki,H.,2006,ObservationofSi(100)SurfacesAnnealedinHydrogenGasAmbientbyScanningTunnelingMicroscopy,AppliedSurfaceScience,252(15),pp.5275. [50] Sato,N.,andYonehara,T.,1994,HydrogenAnnealedSilicon-on-insulator,AppliedPhysicsLetters,65(15),pp.1924. [51] Gao,F.,Wang,Y.,Cao,G.,Jia,X.,andZhang,F.,2006,ReductionofSidewallRoughnessinSilicon-on-insulatorRibWaveguides,AppliedSurfaceScience,252(14),pp.5071. [52] Riveros,R.E.,Yamaguchi,H.,Boggs,T.,Mitsuishi,I.,Mitsuda,K.,Takagi,U.,Ezoe,Y.,Ishizu,K.,andMoriyama,T.,2010,MagneticFieldAssistedFinishingofSiliconMEMSMicro-PoreX-RayOptics,ASMEConferenceProceedings,1,ASME,pp.87. [53] Nakajima,K.,Fujiwara,K.,Pan,W.,andOkuda,H.,2004,ShapedSilicon-crystalWafersObtainedbyPlasticDeformationandtheirApplicationtoSilicon-crystalLenses,NatureMaterials,4(1),pp.47. [54] Ezoe,Y.,Shirata,T.,Ohashi,T.,Ishida,M.,Mitsuda,K.,Fujiwara,K.,Morishita,K.,andNakajima,K.,2009,DevelopmentofHigh-resolutionandLight-weightX-rayOpticswithDeformedSiliconWafers,ProceedingsofSPIE,7360,p.73600B. [55] Ezoe,Y.,Shirata,T.,Mitsuishi,I.,Ishida,M.,Mitsuda,K.,Morishita,K.,andNakajima,K.,2009,ShapedSiliconWafersObtainedbyHotPlasticDeformation:PerformanceEvaluationforFutureAstronomicalX-rayTelescopes,AppliedOptics,48(19),pp.3830. 164

PAGE 165

[56] Mika,M.,Pina,L.,Landova,M.,Sveda,L.,Havlikova,R.,Marsikova,V.,Hudec,R.,andInneman,A.,2011,LightweightMirrorsforSpaceX-RayTelescopes,WorldAcademyofScience,EngineeringandTechnology,58,pp.319. [57] Mitsuishi,I.,Ezoe,Y.,Takagi,U.,Hayashi,T.,Sato,T.,Morishita,K.,Nakajima,K.,Yamasaki,N.,andMitsuda,K.,2009,OpticalImageAnalysisoftheNovelUltra-lightweightandHigh-resolutionMEMSX-rayOptics,InternationalConfer-enceonOpticalMEMSandNanophotonics,IEEE/LEOS,pp.123. [58] Mitsuishi,I.,Ezoe,Y.,Ishizu,K.,Moriyama,T.,Maeda,Y.,Hayashi,T.,Sato,T.,Mita,M.,Yamasaki,N.,Mitsuda,K.,etal.,2010,X-rayImagingTestforaSingle-stageMEMSX-rayOpticalSystem,InternationalConferenceonOpticalMEMSandNanophotonics,IEEE,pp.151. [59] Ezoe,Y.,Moriyama,T.,Ogawa,T.,Kakiuchi,T.,Mitsuishi,I.,Mitsuda,K.,Aoki,T.,Morishita,K.,andNakajima,K.,2012,LargeApertureFocusingofX-rayswithMicroPoreOpticsUsingDryEtchingofSiliconWafers,OpticsLetters,37(5),pp.779. [60] Spears,D.,andSmith,H.,1972,High-resolutionPatternReplicationUsingSoftX-rays,ElectronicsLetters,8(4),pp.102. [61] Becker,E.,Ehrfeld,W.,Mnchmeyer,D.,Betz,H.,Heuberger,A.,Pongratz,S.,Glashauser,W.,Michel,H.,andSiemens,R.,1982,ProductionofSeparation-nozzleSystemsforUraniumEnrichmentbyaCombinationofX-rayLithographyandGalvanoplastics,Naturwissenschaften,69(11),pp.520. [62] Horade,M.,andSugiyama,S.,2009,StudyonOptimizationofExposureEnergyDistributionforFabricationofArbitrary3-DMicrostructurebyShapedBeamusingSynchrotronRadiation,InternationalSymposiumonMicro-NanoMechatronicsandHumanScience,IEEE,pp.559. [63] Becker,E.,Ehrfeld,W.,Hagmann,P.,Maner,A.,andMnchmeyer,D.,1986,FabricationofMicrostructureswithHighAspectRatiosandGreatStructuralHeightsbySynchrotronRadiationLithography,Galvanoforming,andPlasticMoulding(LIGAprocess),MicroelectronicEngineering,4(1),pp.35. [64] Horade,M.,Sugiyama,S.,Ezoe,Y.,Mitsuishi,I.,Ishizu,K.,Moriyama,T.,Mitsuda,K.,Kinuta,S.,Yamanashi,T.,andIchinosawa,Y.,2010,StudyonFabricationofNiMirrorChipforSpaceX-rayTelescopesUtilizingLIGAProcess,Inter-nationalSymposiumonMicro-NanoMechatronicsandHumanScience,IEEE,pp.263. [65] Peele,A.,Irving,T.,Nugent,K.,Mancini,D.,Christenson,T.,Petre,R.,Brumby,S.,andPriedhorsky,W.,2001,LIGAforLobster:FirstObservationofLobster-eyeFocusingfromLithographicallyProducedOptics,ReviewofScienticInstruments,72,p.1843. 165

PAGE 166

[66] Riveros,R.,Yamaguchi,H.,Mitsuishi,I.,Takagi,U.,Ezoe,Y.,Kato,F.,Sugiyama,S.,Yamasaki,N.,andMitsuda,K.,2010,DevelopmentofanAlternatingMagnetic-eldAssistedFinishingProcessforMicroelectromechanicalSystemsMicroporeX-rayOptics,AppliedOptics,49(18),pp.3511. [67] Zhang,W.,2011,IntelligentEnergyFieldManufacturing,CRCPress. [68] Coats,H.,1940,MethodandApparatusforPolishingContainers,Apr.2USPatent2,196,058. [69] Baron,Y.M.,1975,TechnologyofAbrasiveMachininginaMagneticField,Masino-strojenije,Leningrad,Russia. [70] Ruben,H.,1987,AdvancesinSurfaceTreatments,PergamonPress,NewYork,USA. [71] Shinmura,T.,andYamaguchi,H.,1995,StudyonaNewInternalFinishingProcessbytheApplicationofMagneticAbrasiveMachining:InternalFinishingofStainlessSteelTubeandCleanGasBomb,JSMEInternationalJournalSeriesC,Dynamics,Control,Robotics,DesignandManufacturing,38(4),pp.798. [72] Shinmura,T.,andAizawa,T.,1989,StudyonInternalFinishingofaNon-ferromagneticTubingbyMagneticAbrasiveMachiningProcess,BulletinoftheJapanSocietyofPrecisionEngineering,23(1),pp.37. [73] Yamaguchi,H.,andShinmura,T.,2000,StudyofanInternalMagneticAbrasiveFinishingUsingaPoleRotationSystem:DiscussionoftheCharacteristicAbrasiveBehavior,PrecisionEngineering,24(3),pp.237. [74] Yamaguchi,H.,andShinmura,T.,2004,InternalFinishingProcessforAluminaCeramicComponentsbyaMagneticFieldAssistedFinishingProcess,PrecisionEngineering,28(2),pp.135. [75] Yamaguchi,H.,Shinmura,T.,andKobayashi,A.,2001,DevelopmentofanInternalMagneticAbrasiveFinishingProcessforNonferromagneticComplexShapedTubes,JSMEInternationalJournalSeriesC,44(1),pp.275. [76] Shinmura,T.,Takazawa,K.,Hatano,E.,Matsunaga,M.,andMatsuo,T.,1990,StudyonMagneticAbrasiveFinishing,CIRPAnnals-ManufacturingTechnology,39(1),pp.325. [77] Fox,M.,Agrawal,K.,Shinmura,T.,andKomanduri,R.,1994,MagneticAbrasiveFinishingofRollers,CIRPAnnals-ManufacturingTechnology,43(1),pp.181. [78] Chang,G.,Yan,B.,andHsu,R.,2002,StudyonCylindricalMagneticAbrasiveFinishingUsingUnbondedMagneticAbrasives,InternationalJournalofMachineToolsandManufacture,42(5),pp.575. 166

PAGE 167

[79] Kim,J.,andChoi,M.,1997,StudyonMagneticPolishingofFree-formSurfaces,InternationalJournalofMachineToolsandManufacture,37(8),pp.1179. [80] Kim,J.,andNoh,I.,2007,MagneticPolishingofThreeDimensionalDieandMoldSurfaces,TheInternationalJournalofAdvancedManufacturingTechnology,33(1),pp.18. [81] Marinescu,I.,2006,HandbookofLappingandPolishing73CRC. [82] Hanada,K.,Yamaguchi,H.,andZhou,H.,2008,NewSphericalMagneticAbrasiveswithCarriedDiamondParticlesforInternalFinishingofCapillaryTubes,DiamondandRelatedMaterials,17(7-10),pp.1434. [83] Yamaguchi,H.,andHanada,K.,2008,DevelopmentofSphericalMagneticAbrasiveMadebyPlasmaSpray,JournalofManufacturingScienceandEngineering,130,p.031107. [84] Gille,G.,Bredthauer,J.,Gries,B.,Mende,B.,andHeinrich,W.,2000,AdvancedandNewGradesofWCandBinderPowderTheirPropertiesandApplication,InternationalJournalofRefractoryMetalsandHardMaterials,18(2),pp.87. [85] Svoboda,J.,2004,MagneticTechniquesfortheTreatmentofMaterials,KluwerAcademicPublishers,TheNetherlands. [86] Kordonski,W.,Prokhorov,I.,Gorodkin,G.,Jacobs,S.,Puchebner,B.,andPietrowski,D.,1993,MagnetorheologicalFinishing,OpticsandPhotonicsNews,4(12),pp.16. [87] Kordonski,W.,andGolini,D.,1999,FundamentalsofMagnetorheologicalFluidUtilizationinHighPrecisionFinishing,JournalofIntelligentMaterialSystemsandStructures,10(9),p.683. [88] Tricard,M.,DeMarco,M.,andGolini,D.,2002,PolishingHighAspectRatioSubstratesforOptics,Telecommunications,andMicroelectronicsApplicationsUsingMagnetorheologicalFinishing(MRF),OpticalFabricationandTesting,OpticalSocietyofAmerica. [89] Dumas,P.,Fleig,J.,Forbes,G.,Golini,D.,Kordonski,W.,Murphy,P.,Shorey,A.,andTricard,M.,2004,FlexiblePolishingandMetrologySolutionsforFree-formOptics,ProceedingsofASPE. [90] Yamaguchi,H.,Yumoto,K.,Shinmura,T.,andOkazaki,T.,2009,StudyofFinishingofWafersbyMagneticField-AssistedFinishing,JournalofAdvancedMechanicalDesign,Systems,andManufacturing,3(1),pp.35. [91] Kim,W.,Lee,S.,andMin,B.,2004,SurfaceFinishingandEvaluationofThree-dimensionalSiliconMicrochannelUsingMagnetorheologicalFluid,JournalofManufacturingScienceandEngineering,126,p.772. 167

PAGE 168

[92] Yamaguchi,H.,Shinmura,T.,andIkeda,R.,2007,StudyofInternalFinishingofAusteniticStainlessSteelCapillaryTubesbyMagneticAbrasiveFinishing,JournalofManufacturingScienceandEngineering,129,p.885. [93] Papell,S.,1965,LowViscosityMagneticFluidObtainedbytheColloidalSuspensionofMagneticParticles,Nov.2USPatent3,215,572. [94] Sudo,S.,Nakagawa,A.,Shimada,K.,andNishiyama,H.,2005,ShapeResponseofFunctionalFluidDropsinAlternatingMagneticFields,JournalofMagnetismandMagneticMaterials,289,pp.321. [95] Butter,K.,Bomans,P.,Frederik,P.,Vroege,G.,andPhilipse,A.,2003,DirectObservationofDipolarChainsinIronFerrouidsbyCryogenicElectronMicroscopy,NatureMaterials,2(2),pp.88. [96] Imanaka,O.,Kurobe,T.,andMatsushima,K.,1981,AStudyonFieldAssistedFinishingUsingMagneticFluid,ProceedingsofTheJSPESpringConference,774,p.777. [97] Kurobe,T.,Imanaka,O.,andTachibana,S.,1983,MagneticField-assistedFineFinishing,BulletinoftheJapanSocietyofPrecisionEngineering,17(1),pp.49. [98] Kurobe,T.,andImanaka,O.,1984,MagneticField-assistedFineFinishing,PrecisionEngineering,6(3),pp.119. [99] Suzuki,H.,Kodera,S.,Hara,S.,Matsunaga,H.,andKurobe,T.,1989,MagneticField-assistedPolishingApplicationtoaCurvedSurface,PrecisionEngineering,11(4),pp.197. [100] Tani,Y.,Kawata,K.,andNakayama,K.,1984,DevelopmentofHigh-efcientFineFinishingProcessUsingMagneticFluid,CIRPAnnals-ManufacturingTechnology,33(1),pp.217. [101] Umehara,N.,andKato,K.,1988,AStudyonMagneticFluidGrinding,TransactionsoftheJapanSocietyofMechanicalEngineers,54,pp.1599. [102] Childs,T.,Jones,D.,Mahmood,S.,Zhang,B.,Kato,K.,andUmehara,N.,1994,MagneticFluidGrindingMechanics,Wear,175(1-2),pp.189. [103] Umehara,N.,andKalpakjian,S.,1994,MagneticFluidGrinding-ANewTechniqueforFinishingAdvancedCeramics,CIRPAnnals-ManufacturingTechnology,43(1),pp.185. [104] Umehara,N.,Kobayashi,T.,andKato,K.,1995,InternalPolishingofTubewithMagneticFluidGrindingPart1,FundamentalPolishingPropertieswithTaper-typeTools,JournalofMagnetismandMagneticMaterials,149(1),pp.185. 168

PAGE 169

[105] Umehara,N.,andKomanduri,R.,1996,MagneticFluidGrindingofHIP-Si3N4Rollers,Wear,192(1-2),pp.85. [106] Yamaguchi,H.,Shinmura,T.,andTakenaga,M.,2003,DevelopmentofaNewPrecisionInternalMachiningProcessUsinganAlternatingMagneticField,PrecisionEngineering,27(1),pp.51. [107] Takashi,S.,Yongbo,W.,Weimin,L.,andKunio,S.,2009,StudyonMagneticCompoundFluid(MCF)PolishingProcessUsingFluctuatingMagneticField,TransactionsoftheJapanSocietyofMechanicalEngineers,75(753),pp.1007. [108] deBoer,D.,1994,InuenceoftheRoughnessProleontheSpecularReectivityofXraysandNeutrons,PhysicalReviewB,49(9),pp.5817. [109] Yamaguchi,H.,Riveros,R.,Mitsuishi,I.,Takagi,U.,Ezoe,Y.,Yamasaki,N.,Mitsuda,K.,andHashimoto,F.,2010,MagneticField-assistedFinishingforMicroporeX-rayFocusingMirrorsFabricatedbyDeepReactiveIonEtching,CIRPAnnals-ManufacturingTechnology,59(1),pp.351. [110] Elson,J.,andBennett,J.,1995,CalculationofthePowerSpectralDensityFromSurfaceProleData,AppliedOptics,34(1),pp.201. [111] Dligatch,S.,Gross,M.,Nettereld,R.,Pereira,N.,Platt,B.,andNemati,B.,2005,GoldCoatingsforCube-cornerRetro-reectors,JPLTechnicalReportServer1992. [112] Kumar,C.,2009,MagneticNanomaterials,JohnWiley&Sons,Ltd. [113] Holm,C.,andWeis,J.,2005,TheStructureofFerrouids:AStatusReport,CurrentOpinioninColloid&InterfaceScience,10(3),pp.133. [114] Klokkenburg,M.,Vonk,C.,Claesson,E.,Meeldijk,J.,Ern,B.,andPhilipse,A.,2004,DirectImagingofZero-eldDipolarStructuresinColloidalDispersionsofSyntheticMagnetite,JournaloftheAmericanChemicalSociety,126(51),pp.16706. [115] Wu,J.,Aslam,M.,andDravid,V.,2008,ImagingofMagneticColloidsUndertheInuenceofMagneticFieldbyCryogenicTransmissionElectronMicroscopy,AppliedPhysicsLetters,93,pp.082505. [116] Jacobs,I.,andBean,C.,1963,Magnetism,EditorsGT. [117] Philipse,A.,andMaas,D.,2002,MagneticColloidsFromMagnetotacticBacteria:ChainFormationandColloidalStability,Langmuir,18(25),pp.9977. [118] Tripp,S.,Dunin-Borkowski,R.,andWei,A.,2003,FluxClosureinSelf-AssembledCobaltNanoparticleRings,AngewandteChemieInternationalEdition,42(45),pp.5591. 169

PAGE 170

[119] Kasama,T.,Dunin-Borkowski,R.,Scheinfein,M.,Tripp,S.,Liu,J.,andWei,A.,2008,ReversalofFluxClosureStatesinCobaltNanoparticleRingswithCoaxialMagneticPulses,AdvancedMaterials,20(22),pp.4248. [120] Davies,P.,Popplewell,J.,Martin,G.,Bradbury,A.,andChantrell,R.,1986,MonteCarloSimulationsoftheStructureofMagneticFluidComposites,JournalofPhysicsD:AppliedPhysics,19,p.469. [121] Furst,E.,andGast,A.,1999,MicromechanicsofDipolarChainsUsingOpticalTweezers,PhysicalReviewLetters,82(20),pp.4130. [122] Soares,S.,Baselt,D.,Black,J.,Jungling,K.,andStowell,W.,1994,Float-polishingProcessandAnalysisofFloat-polishedQuartz,AppliedOptics,33(1),pp.89. [123] Clarke,A.,andPatey,G.,1994,GroundStateCongurationsofModelMolecularClusters,TheJournalofChemicalPhysics,100,p.2213. [124] Lavender,H.,Iyer,K.,andSinger,S.,1994,GlobalOrientationalOrderinModelPolarClusters,TheJournalofchemicalphysics,101,p.7856. [125] Kulkarni,A.,andBhushan,B.,1996,Nano/PicoindentationMeasurementsonSingle-crystalAluminumUsingModiedAtomicForceMicroscopy,MaterialsLetters,29(4-6),pp.221. 170

PAGE 171

BIOGRAPHICALSKETCH RalRiveroswasbornin1985intheVenezuelatoIvanandAngelicaRiveros.HeandhisfamilyimmigratedtotheUnitedStatesin1992andhehaslivedinFloridaeversince.HejoinedtheUniversityofFloridain2005andgraduatedwithaBachelorofScienceinmechanicalengineeringin2007.Asanundergraduate,RalconductedresearchattheMachineToolResearchCenter(MTRC).HestartedhisgraduateworkundertheguidanceofDr.HitomiYamaguchiGreensletinJanuaryof2008andgraduatedwithaMasterofScienceinMay2009.HethenpursuedaDoctoraldegreeandgraduatedinAugustof2012.Ralhopestocontinuehiscareerasaresearcher. 171