University of Southern California Dissertations and Theses

Laboratory investigations of the near surface plasma field and charging at the lunar terminator

(USC Thesis Other)

Laboratory investigations of the near surface plasma field and charging at the lunar terminator

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content LABORATORYINVESTIGATIONSOFTHENEARSURFACEPLASMAFIELDAND CHARGINGATTHELUNARTERMINATOR by JohnL.Polansky ADissertationPresentedtothe FACULTYOFTHEUSCGRADUATESCHOOL UNIVERSITYOFSOUTHERNCALIFORNIA InPartialFulfillmentofthe RequirementsfortheDegree DOCTOROFPHILOSOPHY (ASTRONAUTICALENGINEERING) August2013 Copyright 2013 JohnL.Polansky Dedication ...to my family and most dear ii Acknowledgments Thepathtakentothisdissertationhasbeenimmenselyrewarding. Specifically,thepastfive years have significantly broadened the scope of my knowledge base, intellect, problem- solving ability, and life in general. I enthusiastically await applying both what I have learned and the tools gained in doing so to the world beyond the University of Southern California. I owe thanks and gratitude to numerous individuals and institutions that have providedassistance,someofwhichinevitablyareforgottenandnotmentionedhere. First of all, to my advisor, Dr. Joseph Wang, thank you for your unwavering academic support and guidance throughout the past four years. You have instilled in me a strong plasma physics foundation, and have also enormously enhanced my scientific research capability through your courses and our work, papers, proposals, and countless discus- sions. I cannot thank you enough. It has been an honor to assist in the establishment of yourexperimentalspace-plasmaphysicslaboratoryhereatUSC. To Dr. Dan Erwin, thank you for the initial offer to join the Astronautical Engineering Department and your continued support ever since. To Dr. Keith Goodfellow, thank you for the many ion source and plasma diagnostics discussions. The details you provided were immensely valuable. To Dr. Mike Gruntman and Dr. Joseph Kunc, thank you for continued inspiration and advice. This department and my role within it could not have existedwithoutyou. ToDr. PhilMuntzandDr. WernerD¨ appen,thankyouforyouruseful feedbackandtimespentreviewingmywork. ToDr. MenguChooftheKyushuInstituteof iii Technology,thankyouforthesixmonthsofplasmaphysicstraininginyourlaboratory,as wellastheupcomingopportunitytojoinyourresearchgroup. IadditionallythankDellCuason,MarriettaPenoliar,AnaOlivares,andallothersinthe department who have impacted and enhanced my work and life on a daily basis. To my former lab partner Ning Ding, special thanks for all that you taught me. You are a circuity wizard and a very talented engineer. To my former lab partner Doug Codron, thank you for your perspective and creativity. The research lab is a far less interesting place now that you both have earned your Ph.D.’s and entered the work place. To Sam Barbour, Ouliang Chang, Kevin Chou, Daoru Han, Will Yu and all current and future Astronautical EngineeringPh.D.students,mayyoucarryonourdepartment’sresearchwithvigor. Finally,Iamprofoundlyindebtedtothoseclosesttome. Tomyparents,unquestionably thehardestworkingindividualsthatIhaveeverencountered,Ithankyouforyoursacrifice and love. The joy that I experience when I come home to the farm is unspeakable. To my siblings,Ithankyouforeverymomentthatwehavespentandwillspendtogether. Youare rolemodelstotheworld,andIamexceptionallyproudofeachofyou. AndtoAmyZaoshi Yuan,Icouldnothavedonethiswithoutyou. JohnL.Polansky May13,2013 This work is supported in part by NASA Lunar Advanced Science and Exploration ResearchgrantNNX11AH21G,NationalScienceFoundationgrant0909364,andtheUSC ViterbiSchoolofEngineering. iv TableofContents Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii ListofFigures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix ListofTables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv ListofSymbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi Chapter1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 MotivationandObjectives . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 PlasmaWakeExpansion . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 LunarPhysicalCharacteristics . . . . . . . . . . . . . . . . . . . . . . . . 4 1.4 LunarPlasmaEnvironment . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.4.1 Earth’sMagnetotailandPlasmaSheet . . . . . . . . . . . . . . . . 6 1.4.2 InfluenceoftheSun . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4.3 TheLunarWake . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.4.4 Additionalfluxsources . . . . . . . . . . . . . . . . . . . . . . . . 8 1.5 LunarRegolithSurface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.6 LunarSurfaceChargingandDustCharging . . . . . . . . . . . . . . . . . 9 1.7 DissertationOutlineandApproach . . . . . . . . . . . . . . . . . . . . . . 13 Chapter2: DevelopmentandCharacterizationofaMesothermalPlasmaSource . . 15 2.1 VacuumFacility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.2 ElectronBombardmentGriddedIonSource . . . . . . . . . . . . . . . . . 15 2.3 PlasmaDiagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.3.1 EmissiveProbe . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.3.2 FaradayProbe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3.3 LangmuirProbe . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.3.4 RetardingPotentialAnalyzer . . . . . . . . . . . . . . . . . . . . . 23 2.3.5 RetardingPotentialAnalyzerIonFocusingEffect . . . . . . . . . . 25 v 2.3.6 DataAcquisitionandTraversingSystem . . . . . . . . . . . . . . . 29 2.4 IonBeamandBackgroundPlasmaCharacterization . . . . . . . . . . . . . 32 2.4.1 ExperimentalSetup . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2.4.2 PlumeMeasurements . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.4.3 PlumeExpansionvs. PlumePotential . . . . . . . . . . . . . . . . 36 2.4.4 EffectsofPlumeCharge-ExchangePlasma . . . . . . . . . . . . . 37 2.4.5 EffectsofFacilityBackgroundPlasma . . . . . . . . . . . . . . . . 44 2.5 SummaryandConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Chapter3: CharacterizationofSurfaceChargingMeasurements . . . . . . . . . . 49 3.1 SurfacePotentialChargingTheoryandSheathConsiderations . . . . . . . 49 3.2 SurfacePotentialDetectionTechniques. . . . . . . . . . . . . . . . . . . . 51 3.2.1 EmissiveProbeSurfacePotentialMeasurements. . . . . . . . . . . 51 3.2.2 EmbeddedWireSurfacePotentialMeasurements . . . . . . . . . . 53 3.2.3 Non-contactingTrekProbeSurfacePotentialMeasurements . . . . 54 3.2.4 DetectionTechniqueConclusions . . . . . . . . . . . . . . . . . . 57 3.3 ConductingPlateAngleofAttackVariation . . . . . . . . . . . . . . . . . 57 3.4 ConductingPlateMaterialVariation . . . . . . . . . . . . . . . . . . . . . 60 3.5 DielectricMaterialProperties . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.5.1 AluminaSilicate . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3.5.2 Mykroy/Mycalex R ⃝ . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3.5.3 JSC-1A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3.6 MainBeamLowAngleofAttackDielectricSurfaceChargingResults . . . 67 3.6.1 Aluminumvs. Aluminum . . . . . . . . . . . . . . . . . . . . . . . 68 3.6.2 AluminaSilicatevs. Mykroy/Mycalex R ⃝ . . . . . . . . . . . . . . 69 3.6.3 AluminaSilicatevs. JSC-1A . . . . . . . . . . . . . . . . . . . . . 70 3.6.4 AluminaSilicateSolidvs. AluminaSilicateDust . . . . . . . . . . 71 3.7 SummaryandConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Chapter4: ExperimentalandNumericalInvestigationsoftheNearSurfaceLunar PlasmaField . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.1 PlasmaDensityandLengthScaling. . . . . . . . . . . . . . . . . . . . . . 74 4.2 LunarSimulatorFacilityDevelopment . . . . . . . . . . . . . . . . . . . . 75 4.2.1 SimulatedLunarObjectSelectionandDimensions . . . . . . . . . 76 4.2.2 IonSpeciesCharacterizationbyRPA . . . . . . . . . . . . . . . . . 78 4.3 PlasmaSpeciesControlwithCEXPlateandFinalExperimentalSetup . . . 84 4.3.1 PlasmaSpeciesControlTestResults . . . . . . . . . . . . . . . . . 86 4.4 ExperimentalPlasmaWakeExpansion . . . . . . . . . . . . . . . . . . . . 90 4.4.1 ExperimentalWakeExpansionResults . . . . . . . . . . . . . . . . 90 4.4.2 AnalyticalWakeExpansionComparison . . . . . . . . . . . . . . . 93 4.4.3 NumericalWakeExpansionComparison . . . . . . . . . . . . . . . 93 4.5 NumericalPlasmaWakePICCodeDevelopmentandValidation . . . . . . 95 4.5.1 DomainSetup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 vi 4.5.2 ParticleInjection . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 4.5.3 NodeCharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 4.5.4 SORPotentialandElectricFieldSolver . . . . . . . . . . . . . . . 98 4.5.5 ParticleMover . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 4.5.6 ChargeDeposition . . . . . . . . . . . . . . . . . . . . . . . . . . 99 4.5.7 ApplyBoundaryConditions . . . . . . . . . . . . . . . . . . . . . 100 4.5.8 ParticleStatistics . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.5.9 PICcodeValidation . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.5.10 LunarSimulatorFacilityPICcodeApplication . . . . . . . . . . . 102 4.6 SummaryandConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Chapter5: ExperimentalInvestigationsofLunarSurfaceCharging . . . . . . . . . 106 5.1 LengthScalingandExperimentalConsiderations . . . . . . . . . . . . . . 106 5.1.1 LengthScaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 5.1.2 BackgroundPressureImplications . . . . . . . . . . . . . . . . . . 107 5.1.3 Argonplasmavs. HydrogenplasmaImplications . . . . . . . . . . 108 5.1.4 LunarSimulantImplications . . . . . . . . . . . . . . . . . . . . . 109 5.2 ExperimentalSetupandExperimentalFluxSpecies . . . . . . . . . . . . . 110 5.3 DielectricSurfaceCharging: AluminaSilicatePlatevs. Mykroy/Mycalex R ⃝ Plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 5.4 Dielectric Surface Charging: Alumina Silicate Plate vs. Alumina Silicate Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5.5 DielectricSurfaceCharging: AluminaSilicatePlatevs. JSC-1ADust . . . 120 5.6 IonChargeDepositionMechanisms: Ion-inducedSecondaryEmissionYield andIonReflection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 5.6.1 Aluminasilicatesolidvs. aluminasilicatedust . . . . . . . . . . . 125 5.6.2 Aluminasilicatesolidvs. Mykroy/Mycalex R ⃝ . . . . . . . . . . . . 128 5.6.3 Aluminasilicatesolidvs. JSC-1Adust . . . . . . . . . . . . . . . . 128 5.7 ChargeDecay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 5.8 AluminaSilicateSurfaceDustChargetoSingleIsolatedDustChargeCom- parison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 5.9 SummaryandConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Chapter6: Conclusions,ImplicationsandFutureResearch . . . . . . . . . . . . . 134 6.1 ConclusionsandImplications . . . . . . . . . . . . . . . . . . . . . . . . . 134 6.1.1 PlasmaWakeExpansion: VacuumChambervs. inSpace . . . . . . 135 6.1.2 SurfaceCharging: PlanarDielectricvs. DustDielectric . . . . . . . 136 6.1.3 GrainCharge: DustySurfaceGrainsvs. IsolatedGrain . . . . . . . 138 6.2 FuturePlasmaResearch . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 6.2.1 RetardingPotentialAnalyzerIonFocusing . . . . . . . . . . . . . 140 6.2.2 CEXandFacilityEffects . . . . . . . . . . . . . . . . . . . . . . . 140 6.2.3 ElectronEnergyDistribution . . . . . . . . . . . . . . . . . . . . . 140 6.3 FutureUVResearch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 vii 6.4 FutureSurfaceChargingResearch . . . . . . . . . . . . . . . . . . . . . . 141 6.4.1 Kinetic Energy Ion-Induced Secondary Electron Emission and Ion AbsorptionRates . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 6.4.2 ChargeDecay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 6.4.3 DustLayerGrainCharge . . . . . . . . . . . . . . . . . . . . . . . 142 6.4.4 ChargedDustMotion . . . . . . . . . . . . . . . . . . . . . . . . . 142 6.4.5 LocalTopography . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 6.5 FutureNumericalResearch . . . . . . . . . . . . . . . . . . . . . . . . . . 142 ReferenceList . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 viii ListofFigures 1.1 LunarorbitaboutEarthduringApollo17mission . . . . . . . . . . . . . . 6 1.2 LunarorbitthroughEarth’smagnetosphere(nottoscale) . . . . . . . . . . 11 1.3 Schematicofthelunarchargingenvironment(nottoscale)[68] . . . . . . . 11 1.4 Chargingofthelunarregolithsurface . . . . . . . . . . . . . . . . . . . . 11 2.1 Primaryresearchchamber . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.2 Custom-built4cmelectronbombardmentgriddedionsource . . . . . . . . 17 2.3 Emissiveprobe: photographandelectricalschematic. . . . . . . . . . . . . 19 2.4 Faradayprobeelectricalschematicandphotograph . . . . . . . . . . . . . 20 2.5 Langmuirprobeelectricalschematicandphotograph . . . . . . . . . . . . 22 2.6 ExampleLangmuirprobeI-Vcurve . . . . . . . . . . . . . . . . . . . . . 23 2.7 RPAschematicandphotographs . . . . . . . . . . . . . . . . . . . . . . . 24 2.8 RPAmainbeamexperimentaldata . . . . . . . . . . . . . . . . . . . . . . 25 2.9 RPAI-Vtraces8inchesdownstreamof6mAbeam . . . . . . . . . . . . . 28 2.10 RPAionfocusingeffectexperimentalsetup . . . . . . . . . . . . . . . . . 29 2.11 RPA grid and collector currents for varying alpha, 8 inches downstream of 6mAbeam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2.12 QuantitativeanalysisofRPAI-V5 alphacase . . . . . . . . . . . . . . . . 31 2.13 Traversingsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2.14 Probesuite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.15 DAQandcontrolsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 ix 2.16 Experimentalsetup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.17 Plasmaplumenormalized2Dcontours,I =10mA . . . . . . . . . . . . . 35 2.18 Radial potential profiles at 3 downstream locations (left); Comparison of potentialprofilestoanalyticalexpansionsolution(right) . . . . . . . . . . . 38 2.19 Radial ion density as a function of _ m (left); Radial electron density as a functionof _ m(right) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 2.20 Faradayprobemountedonrotatingarmforhalf-angledivergenceexperiment 42 2.21 Half-anglecurrentdensitysweep,10mA,4.1sccm . . . . . . . . . . . . . 42 2.22 RPAI-Vcurvesfor1and10sccmoperatingconditions . . . . . . . . . . . 43 2.23 CEXiondensitymodelcomparison . . . . . . . . . . . . . . . . . . . . . 44 2.24 Radial ion density vs. number of graphite panels (left); Radial electron densityvs. numberofgraphitepanels(right) . . . . . . . . . . . . . . . . . 45 2.25 61cmx61cmgraphitetargetatchamberrear . . . . . . . . . . . . . . . . 46 3.1 Plasmasheathprofiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.2 Photoelectronsheathprofiles . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.3 Plasmapotentialaboveabiasedaluminumplate . . . . . . . . . . . . . . . 52 3.4 Emissiveprobeanddatalogexperimentalsetup(nottoscale) . . . . . . . . 52 3.5 Floatingconductingplatesgeometryandemissiveprobe . . . . . . . . . . 53 3.6 Conductingplatefloatingpotentialsandemissiveprobeerror . . . . . . . . 53 3.7 Embeddedwiresschematic,geometry,andchargingresults . . . . . . . . . 54 3.8 TrekESVMnon-contactingsurfacepotentialmeasurementmethod . . . . . 55 3.9 Floating dielectric plates, embedded wires, and Trek probe experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 3.10 Trek probe measured dielectric floating potentials (left); Embedded wires measureddielectricfloatingpotentials(right) . . . . . . . . . . . . . . . . 56 3.11 Flatplatechargingexperimentalsetup(nottoscale) . . . . . . . . . . . . . 58 3.12 Flatplatefloatingpotentialasafunctionofangleofattack . . . . . . . . . 59 x 3.13 Analyticalsurfacechargingmodelasafunctionofalpha . . . . . . . . . . 60 3.14 Conductingplatematerialvariationexperimentalsetup(nottoscale) . . . . 61 3.15 Conductingplatematerials: Stainlesssteel(SS),aluminum(Al),copper(Cu) 61 3.16 Conductingplatefloatingpotentialsat0 inmainbeam(left);Conducting platefloatingpotentialsat90 outsidemainbeam(right); . . . . . . . . . 62 3.17 ConductingmaterialAugeremissionproperties . . . . . . . . . . . . . . . 63 3.18 Dielectricmaterialproperties . . . . . . . . . . . . . . . . . . . . . . . . . 64 3.19 Crushingandsievingaluminasilicate . . . . . . . . . . . . . . . . . . . . 66 3.20 Mainbeamdielectricmaterialchargingexperimentalsetup . . . . . . . . . 68 3.21 5 ′′ x5 ′′ Watlowpolyimideheaterformoistureoutgassing . . . . . . . . . . 68 3.22 Aluminumconductingstripssetupandmainbeamchargingresults . . . . . 69 3.23 AluminasilicateandMykroy/Mycalex R ⃝ mainbeamcharging . . . . . . . 70 3.24 AluminasilicateandJSC-1Amainbeamcharging . . . . . . . . . . . . . . 71 3.25 Aluminasilicatesolidandaluminasilicatedustmainbeamcharging . . . . 72 4.1 Experimentalplasmadensityandcharacteristiclengthscaling . . . . . . . . 76 4.2 Initiallunarsimulatorfacilityexperimentalsetup . . . . . . . . . . . . . . 77 4.3 Cubessimulatinglunarobjectsinlunarsimulatorfacilitydevelopment . . . 77 4.4 Initiallunarsimulatorfacilityplasmapotentialϕ p . . . . . . . . . . . . . . 79 4.5 Initiallunarsimulatorfacilityaxialiondensityn i . . . . . . . . . . . . . . 80 4.6 Initiallunarsimulatorfacilitytotalelectrondensityn e . . . . . . . . . . . . 81 4.7 Initiallunarsimulatorfacilityspacechargen i -n e . . . . . . . . . . . . . . 82 4.8 RPAlocationswithrespecttosimulatedlunarobject . . . . . . . . . . . . . 83 4.9 RPAmainbeamaxialflowhighvoltagescan . . . . . . . . . . . . . . . . . 83 4.10 1DJ i radialandtotaln e 0.75 ′′ abovedielectricsurface . . . . . . . . . . . 83 4.11 Shadowedregionradialflowscans . . . . . . . . . . . . . . . . . . . . . . 85 4.12 LunarsimulatorfacilityandCEXplatesketch . . . . . . . . . . . . . . . . 87 xi 4.13 Finallunarsimulatorfacilityexperimentalsetup . . . . . . . . . . . . . . . 87 4.14 1DJ i radialCEXionfluxandtotaln e 0.75 ′′ abovedielectricsurface . . . . 88 4.15 ExperimentalelectronkineticenergyincreasebyCEXplatebias . . . . . . 89 4.16 Finallunarsimulatorfacilityplasmapotentialϕ p andspacechargen i -n e . 91 4.17 Finallunarsimulatorfacilityaxialiondensityn i andtotalelectrondensityn e 92 4.18 Analyticalandnumericalsimulationionexpansionintheshadowedregion . 94 4.19 Generalnumericalrecipe . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 4.20 Numericaldomainsetupandmesothermalplasmaflow . . . . . . . . . . . 96 4.21 Nodechargeallocationscheme . . . . . . . . . . . . . . . . . . . . . . . . 97 4.22 SORalgorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 4.23 Plategeometryanddomainboundaryconditions . . . . . . . . . . . . . . . 100 4.24 Simulationvalidationtestsetup . . . . . . . . . . . . . . . . . . . . . . . . 102 4.25 Simulationtestcasesinputs . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.26 Simulationtestcaseenergyhistory . . . . . . . . . . . . . . . . . . . . . . 103 4.27 Numericalandanalyticalexpansionpotentialsolutionsoverabiasedplate . 103 4.28 Numericalflowdomain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 5.1 Lunartopographicobstructiontosolarwind . . . . . . . . . . . . . . . . . 108 5.2 Lunarsimulatorfacilitydielectricchargingexperimentalsetup . . . . . . . 111 5.3 Potentialejectionenergies . . . . . . . . . . . . . . . . . . . . . . . . . . 113 5.4 Experimentalfluxsourcestodielectricsolidanddustsurfaces . . . . . . . . 114 5.5 AluminasilicateandMykroy/Mycalex R ⃝ platesforlunarsimulatorfacility . 115 5.6 Alumina silicate and Mykroy/Mycalex R ⃝ surface potentials as a function ofCEXplatesetting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 5.7 Aluminasilicateplateandaluminasilicatedustinlunarsimulatorfacility . 118 5.8 Aluminasilicateplateandaluminasilicatedustsurfacepotentialsasafunc- tionofCEXplatesetting . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 xii 5.9 Alumina silicate plate and alumina silicate dust mirrored charging in the lunarsimulatorfacility . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 5.10 AluminasilicateplateandJSC-1Adustinlunarsimulatorfacility . . . . . . 121 5.11 Alumina silicate plate and JSC-1A dust surface potentials as a function of CEXplatesetting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 5.12 Schematicshowingmultiple-electronemissionbyafastionatgrazinginci- denceonasurface[6] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 5.13 Experimentalnormalincidenceonplanarsurfaceandgrazingincidenceon dustsurface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 5.14 SEMimagesofplanaranddustaluminasilicateexperimentalsurfaces . . . 128 5.15 Measureddielectricsurfacepotentialsasafunctionoftime . . . . . . . . . 130 5.16 Averagedielectricsurfacepotentialdifferenceasafunctionoftime . . . . . 131 xiii ListofTables 1.1 PhysicalcomparisonoftheMoonandEarth[29] . . . . . . . . . . . . . . . 5 1.2 Typicallunarambientplasmaproperties[24] . . . . . . . . . . . . . . . . . 7 2.1 Massflowratevariationoperatingconditions . . . . . . . . . . . . . . . . 35 2.2 Measuredandnormalizedplasmaparametersatthrusterexit . . . . . . . . 36 2.3 Plasmasourcecurrentemissionoperatingconditions . . . . . . . . . . . . 36 2.4 NSTARto4cmgriddedionsourcecomparison. . . . . . . . . . . . . . . . 43 5.1 Nominalplasmabeamsettings . . . . . . . . . . . . . . . . . . . . . . . . 115 5.2 CEXplatesettingsandlunarsimulatorradialplasmaflux . . . . . . . . . . 124 5.3 Tabulatedlunarsimulatorfacilitychargingresultsandcalculated si . . . . 125 5.4 Dustysurfacegrainchargetoisolatedgrainchargecomparison . . . . . . . 133 xiv ListofSymbols v te electronthermalvelocity . . . . . . . . . . . . . . . . . . . . . . . 3 v beam ionbeamvelocity . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 v ti ionthermalvelocity . . . . . . . . . . . . . . . . . . . . . . . . . 3 R e radiusofEarth . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 DC DCelectricalconductivity . . . . . . . . . . . . . . . . . . . . . . 8 ϕ p plasmapotential . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 n e electronnumberdensity . . . . . . . . . . . . . . . . . . . . . . . 17 n i ionnumberdensity . . . . . . . . . . . . . . . . . . . . . . . . . . 17 T e electrontemperature . . . . . . . . . . . . . . . . . . . . . . . . . 17 T i iontemperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 mfp;plasma plasmameanfreepath . . . . . . . . . . . . . . . . . . . . . . . . 18 L source ionsourcecharacteristiclength . . . . . . . . . . . . . . . . . . . . 18 k Boltzmannconstant . . . . . . . . . . . . . . . . . . . . . . . . . . 18 m e electronmass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 e elementarycharge . . . . . . . . . . . . . . . . . . . . . . . . . . 18 m i ionmass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 eAr electrontoargoncollisionalcross-section . . . . . . . . . . . . . . 18 n neu;exit neutraldensityatsourceexit . . . . . . . . . . . . . . . . . . . . . 18 ϕ o beamvoltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 J i ioncurrentdensity . . . . . . . . . . . . . . . . . . . . . . . . . . 20 xv v i velocityofsourceions . . . . . . . . . . . . . . . . . . . . . . . . 20 d Debyelength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 ϵ o permittivityoffreespace . . . . . . . . . . . . . . . . . . . . . . . 21 L o characteristiclength . . . . . . . . . . . . . . . . . . . . . . . . . 22 I esat electronsaturationcurrent . . . . . . . . . . . . . . . . . . . . . . 23 A probe Langmuirprobesurfacearea . . . . . . . . . . . . . . . . . . . . . 23 dI dV RPAI-Vcurvederivative . . . . . . . . . . . . . . . . . . . . . . . 25 A c RPAcollectorarea . . . . . . . . . . . . . . . . . . . . . . . . . . 25 q i ioncharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 V ionr ionretardingvoltage . . . . . . . . . . . . . . . . . . . . . . . . . 25 ISUM sumofRPAcollectorandwallcurrents . . . . . . . . . . . . . . . 27 I c RPAcollectorcurrent . . . . . . . . . . . . . . . . . . . . . . . . . 27 %open RPAgridtransparency . . . . . . . . . . . . . . . . . . . . . . . . 27 n i -n e spacecharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 R b beamradius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 I beamcurrent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 P c vacuumchamberpressure . . . . . . . . . . . . . . . . . . . . . . 35 v i;avg averagevelocityofsourceions . . . . . . . . . . . . . . . . . . . . 36 v te;avg averageelectronthermalvelocity . . . . . . . . . . . . . . . . . . 36 V fil dischargefilamentvoltage . . . . . . . . . . . . . . . . . . . . . . 36 I fil dischargefilamentcurrent . . . . . . . . . . . . . . . . . . . . . . 36 n io plasmadensityatsourceexit . . . . . . . . . . . . . . . . . . . . . 36 M o Machnumber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 v io ionvelocityatsourceexit . . . . . . . . . . . . . . . . . . . . . . 36 C s ionacousticvelocity . . . . . . . . . . . . . . . . . . . . . . . . . 36 radialexpansionangle . . . . . . . . . . . . . . . . . . . . . . . . 36 n neu;exit neutraldensityatsourceexit . . . . . . . . . . . . . . . . . . . . . 37 xvi ionizationrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 c n speedofneutrals . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 A orifice totalexitgridopenarea . . . . . . . . . . . . . . . . . . . . . . . . 37 cex;exit exitplanecharge-exchangemeanfreepath . . . . . . . . . . . . . 37 cex charge-exchangecrosssection . . . . . . . . . . . . . . . . . . . . 37 L chamber chambercharacteristiclength . . . . . . . . . . . . . . . . . . . . 37 _ m argonmassflowrate . . . . . . . . . . . . . . . . . . . . . . . . . 37 J OML orbit-motionlimitedionfluxdensity . . . . . . . . . . . . . . . . . 39 v cex charge-exchangeionvelocity . . . . . . . . . . . . . . . . . . . . . 39 Q i orbit-motionlimitedtheoryparameter . . . . . . . . . . . . . . . . 39 ϕ FP Faradayprobenegativebias . . . . . . . . . . . . . . . . . . . . . 39 div 95%half-anglebeamdivergence . . . . . . . . . . . . . . . . . . . 39 r sweep half-angledivergencesweepradius . . . . . . . . . . . . . . . . . 40 n neu localneutraldensity . . . . . . . . . . . . . . . . . . . . . . . . . 40 _ n cex charge-exchangeproductionrate . . . . . . . . . . . . . . . . . . . 40 A halfsphere CEXexitsurfacearea . . . . . . . . . . . . . . . . . . . . . . . . 40 n cex charge-exchangeiondensity . . . . . . . . . . . . . . . . . . . . . 40 r volume radialdistancefromsourcecenterlineoutwards . . . . . . . . . . . 42 cex;avg chamberaveragedcharge-exchangemeanfreepath . . . . . . . . . 47 n neu;avg chamberaveragedneutraldensity . . . . . . . . . . . . . . . . . . 47 T neu neutralargontemperature . . . . . . . . . . . . . . . . . . . . . . 47 ϕ surf surfacefloatingpotential . . . . . . . . . . . . . . . . . . . . . . . 49 I k varioussurfacecurrentsources . . . . . . . . . . . . . . . . . . . . 49 I e plasmaelectroncurrent . . . . . . . . . . . . . . . . . . . . . . . . 49 I i plasmaioncurrent . . . . . . . . . . . . . . . . . . . . . . . . . . 49 I se electron-inducedsecondaryelectroncurrent . . . . . . . . . . . . . 49 I si ion-inducedsecondaryelectroncurrent . . . . . . . . . . . . . . . 49 xvii I bse backscatteredelectroncurrent . . . . . . . . . . . . . . . . . . . . 49 I ph photoelectroncurrent . . . . . . . . . . . . . . . . . . . . . . . . . 49 ϕ s Trekprobesamplingsurfacepotential . . . . . . . . . . . . . . . . 54 V Trekprobevibratingelectrodepotential . . . . . . . . . . . . . . . 54 ! Trekprobeupperelectrodevibratingfrequency . . . . . . . . . . . 54 d 1 Trekprobeupperelectrodevibratingamplitude . . . . . . . . . . . 54 d Trekprobetimedependentgap . . . . . . . . . . . . . . . . . . . . 54 d 0 Trekprobetimedependentgapreference . . . . . . . . . . . . . . 54 angleofattack . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 e analyticalelectronflux . . . . . . . . . . . . . . . . . . . . . . . . 58 v de electrondriftvelocity . . . . . . . . . . . . . . . . . . . . . . . . . 58 i analyticalionflux . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 si analyticalion-inducedsecondaryelectronflux . . . . . . . . . . . 58 si ion-inducedsecondaryelectronemissionyield . . . . . . . . . . . 58 crit criticalangleofattack . . . . . . . . . . . . . . . . . . . . . . . . 59 T esec secondaryelectrontemperature . . . . . . . . . . . . . . . . . . . 59 Auger Augeremissionyield . . . . . . . . . . . . . . . . . . . . . . . . . 63 J ph photoemissioncurrentflux . . . . . . . . . . . . . . . . . . . . . . 75 dmoon lunarDebyelength . . . . . . . . . . . . . . . . . . . . . . . . . . 75 dchamber lunarsimulatorfacilityinletDebyelength . . . . . . . . . . . . . . 75 ϕ edge sheathedgepotential . . . . . . . . . . . . . . . . . . . . . . . . . 84 ^ ϕ edge normalizedsheathedgepotential . . . . . . . . . . . . . . . . . . . 84 n edge sheathedgedensity . . . . . . . . . . . . . . . . . . . . . . . . . . 86 ^ n edge normalizedsheathedgedensity . . . . . . . . . . . . . . . . . . . 86 edge sheathedgeDebyelength . . . . . . . . . . . . . . . . . . . . . . 86 V cexplate CEXplatebias . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 M psh ionpre-sheathMachnumber . . . . . . . . . . . . . . . . . . . . . 86 xviii ^ d sh normalizedCEXplatesheaththickness . . . . . . . . . . . . . . . 86 d sh CEXplatesheaththickness . . . . . . . . . . . . . . . . . . . . . . 86 ^ v te normalizedelectronthermalvelocity . . . . . . . . . . . . . . . . . 96 ^ v beam normalizedionbeamvelocity . . . . . . . . . . . . . . . . . . . . 96 ^ v ti normalizedionthermalvelocity . . . . . . . . . . . . . . . . . . . 96 totalchargedensity . . . . . . . . . . . . . . . . . . . . . . . . . . 98 A coefficientmatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 ⇀ ϕ potentialfield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 ⇀ B nodechargematrix . . . . . . . . . . . . . . . . . . . . . . . . . . 98 ! SOR SORover-relaxationparameter . . . . . . . . . . . . . . . . . . . 98 ⇀ E electricfield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 ⇀ a particleaccelerationvector . . . . . . . . . . . . . . . . . . . . . . 99 ⇀ v f particlefinalvelocity . . . . . . . . . . . . . . . . . . . . . . . . . 99 ⇀ v i particleinitialvelocity . . . . . . . . . . . . . . . . . . . . . . . . 99 dt timestep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 ⇀ x f particlefinalposition . . . . . . . . . . . . . . . . . . . . . . . . . 99 ⇀ x i particleinitialposition . . . . . . . . . . . . . . . . . . . . . . . . 99 analyticalexpansionsolutionrestrictionparameter . . . . . . . . . 101 ^ ϕ plate normalizedsimulationplatepotential . . . . . . . . . . . . . . . . 101 ^ L plate normalizedsimulationplatelength . . . . . . . . . . . . . . . . . . 101 o absoluteexpansionangle . . . . . . . . . . . . . . . . . . . . . . . 101 KE argon argonkineticenergy . . . . . . . . . . . . . . . . . . . . . . . . . 108 m argon argonmass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 v argon argonvelocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 KE proton protonkineticenergy . . . . . . . . . . . . . . . . . . . . . . . . . 108 m proton protonmass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 v proton protonvelocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 xix wavelength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 E energyleveldifference . . . . . . . . . . . . . . . . . . . . . . . . 110 E k Augeremissionelectronenergy . . . . . . . . . . . . . . . . . . . 112 E i ionizationpotential . . . . . . . . . . . . . . . . . . . . . . . . . . 112 E g bandgapenergy . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 electronaffinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 J e electronflux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 ϕ AlSilplate aluminasilicateplatefloatingpotential . . . . . . . . . . . . . . . 123 ϕ measured measuredfloatingpotentialofagivendielectricsurface . . . . . . . 123 ϕ AlSildust aluminasilicatedustfloatingpotential . . . . . . . . . . . . . . . . 125 ϕ MMplate Mykroy/Mycalex R ⃝ platefloatingpotential . . . . . . . . . . . . . 125 ϕ JSC1A JSC-1Adustfloatingpotential . . . . . . . . . . . . . . . . . . . . 125 siAlSildust aluminasilicatedustion-inducedsecondaryemissionyield . . . . . 125 siMM Mykroy/Mycalex R ⃝ plateion-inducedsecondaryemissionyield . . 125 siJSC1A JSC-1Adustion-inducedsecondaryemissionyield . . . . . . . . . 125 r d dustgrainradius . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 C d freespacecapacitance . . . . . . . . . . . . . . . . . . . . . . . . 131 Q dexp experimentalaluminasilicatedustgraincharge . . . . . . . . . . . 132 surfacechargedensity . . . . . . . . . . . . . . . . . . . . . . . . 132 ϵ rd aluminasilicatedielectricconstant . . . . . . . . . . . . . . . . . . 132 d layer dustlayerthickness . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Q diso calculatedisolatedaluminasilicatedustgraincharge . . . . . . . . 132 xx Abstract The objective of this dissertation is to study the near surface plasma environment and sur- face charging environment at the lunar terminator through experimental, analytical and numericalinvestigations. Specifically,thisdissertationinvestigates: 1. theplasmawakeexpansionprocess, 2. regolithsurfacecharging,and 3. regolithgraincharging. Inthefirstarea,thepotentialanddensityoftheplumeemittedfromagriddedionsource isinvestigated. Acapabilityforefficienttwo-dimensionalmeasurementsoftheplumepro- file is developed. The effects of plume expansion, plume charge-exchange plasma, and facilitybackgroundplasmaonplumecharacteristicsarequantified. Itisfoundthatthepro- pellant charge-exchange plasma is the primary factor that terminates the plume expansion process, and thus largely controls the magnitude of the plume potential with respect to the ambient. The plasma plume wake expansion process is then controlled to experimentally simulate low angle of attack lunar plasma flow and surface charging. The control method isvalidatedbyanalyticalandnumericalmodels. Inthesecondarea,conductors,planardielectrics,anddustdielectricsurfacesarecharged under mesothermal plasma flux. Charging experiments are conducted in both the main plasma beam and the simulated lunar plasma environment. The effects of high energy ion xxi flux, low energy ion flux, and low electron flux are quantified as a function of angle of attack, material properties, and surface properties. It is found that a dielectric dust surface willchargetoasignificantlymorepositivesurfacepotentialthanaplanardielectricsurface of the chemical same composition and thickness, when the ion and electron flux is of the same order. Further, this effect is found to be driven by ion-induced secondary electron emissionandionchargedepositionmechanisms. Thehighion-drivenpotentialofdustsur- faces compared to planar surfaces may be contributing to lunar terminator region charged dustmotionandtransport. In the third area, dusty surface layer individual grain charge is compared to the charge of a single, isolated dust grain at the same potential. Dusty surface layer potentials are recorded experimentally, and individual grain charge is derived using a capacitance sys- tem model. The results are compared to isolated grain charge that is calculated using the free space capacitance model and the experimental surface potential. It is shown that the charge of dust grains comprising a dusty surface layer is one to two orders of magnitude lower than the charge of an isolated dust grain. This demonstrates the packed dust grain capacitive effect on dusty surface layers such as the lunar regolith, and will lead to signif- icantly different predictions of electrostatic levitation and dust transport dynamics on the lunarsurface. xxii CHAPTER 1: INTRODUCTION Between 1966 and 1976 the first field explorations of the Moon were conducted through theApolloprogramandbyvariousautomatedprobeslaunchedbytheU.S.andU.S.S.R.In addition to achieving the first, and currently only, human exploration to another planetary body, these missions also greatly enhanced technology, inspired new generations of sci- entists and engineers, and significantly broadened understanding of the Moon’s origin and composition. While the “Space Race” of that era does not exist today, there is no question thathumankindwillcontinuetoprobetheboundaryoftheimpossible,alongitsmovement intotheuniverse. 1.1 MotivationandObjectives TheMoonisdirectlyexposedtothesolarradiationandvariousspaceplasmaenvironments whichinteractwiththelunarsurfaceandanyobjectsonthesurface. Adirectconsequence of such interactions is surface charging. Lunar surface charging is relevant to almost all aspectsofhumanandroboticlunarexplorationactivitiesaswellasmanynaturalprocesses onthelunarsurface. Probetheoryhasprovidedanoverallscopeoflunarsurfacecharging. Data from the Apollo surface missions, and more recently Lunar Prospector, has yielded furtherunderstandingofthelunarelectrostaticsurfacepotentialanditsresponsetoincident currents. However, charging in the terminator region is the most complex and is still not wellunderstood. Theterminatorregionisconsideredtobeboththesweepinglight-shadow transition line and the lunar polar regions. As the mean inclination of the lunar equator to the ecliptic is only 1.54 , some high altitude polar surfaces may be perennially sunlit, and some low altitude polar surfaces may be perennially dark. Accordingly, the polar regions areattractivesurfacemissionlocationsduetoavailablesolarpowerandcold-trappedlunar waterice,amongotherreasons. 1 In the terminator region surface potential is sensitively influenced by both solar illu- mination and plasma flow. The combined effects from localized shadow generated by low sun elevation angles and localized wake generated by plasma flow over the rugged terrain can generate strongly differentially charged surfaces. Few models currently exist that can accuratelyresolvethecombinedeffectsofplasmaflowandsolarilluminationoverrealistic lunarterminatortopographies. This dissertation focuses on the near surface plasma field and charging at the lunar terminator. Laboratory investigations are carried out to simulate localized lunar surface chargingduetolowangleofattackplasmaflow. Theresearchseekstoanswerthefollowing longstandingquestions: How different is the plasma wake expansion process in the vacuum chamber com- paredtoinspace? How different is the charging process of the lunar regolith surface compared to that ofaplanardielectricsurface? How different is the charge of dust grains as part of a regolith surface compared to thatofasingle,isolateddustgraininplasma? Answerstothesequestionsarecriticalstepsalongthewaytodevelopingacomprehen- sive lunar terminator surface potential model that resolves the combined plasma flow and solar illumination effects. The model could then be used to enhance lunar surface mission planning and design, predict lunar dust motion and transport, and support lunar resource prospectingandprocessing,amongotherapplications. 1.2 PlasmaWakeExpansion Thepredictionoftheelectricpotentialintheplasmaplumeemittedfromaplasmathruster oraplasmasourceisafundamentalprobleminthestudyofspacecraft-plasmainteractions 2 andelectricpropulsion. Itisalsoacorollarytoplasmaexpansionoverobjectsinspace,such as plasma flow over lunar surface topography or equipment. The plasma plume emitted by an ion thruster type plasma source is composed primarily of a “propellant” plasma (high-energy beam ions and neutralizing thermal electrons), un-ionized neutrals, and low- energy charge-exchange (CEX) ions generated by collisions between the beam ions and neutrals in the plume. Here, beam ions and neutralizing electrons shall be referred to as propellant plasma, and the CEX plasma generated by propellant ions and un-ionized neutrals as the plume CEX plasma. All other plasma populations generated due to the presence of the vacuum chamber shall be referred to as the facility background plasma. The characteristic of the propellant plasma emitted is that of a collisionless, mesothermal plasma flow with v te >> v beam >> v ti , where v te , v beam , and v ti are the electron thermal velocity,beamvelocity,andionthermalvelocity. Plasmaplumesemittedbyothertypesof plasmathrustersoftenhaveasimilarcompositionandflowcharacteristics. The plume potential with respect to the ambient is determined by the net charge inside the plume, which in turn is controlled by the ion beam neutralization near the source and the subsequent plasma expansion from source to ambient. The primary plasma expan- sion process is carried out by the propellant plasma. The factors that influence the plasma expansion include the electron-ion coupling in the plume, the electron kinetics, the back- groundplasma,andtheboundarycondition[79]. Astheexpansionwillbeterminatedwhen the density is reduced to that of the background, the plume CEX plasma and the facility backgroundplasmaplayanimportantroleincontrollingtheprimaryexpansionprocess. The ion thruster potential has been measured extensively by many vacuum chamber tests of ion thrusters. In contrast, the Deep Space 1 mission provides the only detailed in-flight measurements [76]. As few detailed studies have been carried out to understand quantitativelytheprocessesandfactorsthatcontroltheplumepotentialwithrespecttothe ambient, it is not well understood how one should extrapolate ground measured plasma plume potential to in-flight conditions. In modeling and simulation studies, the plume 3 potential is taken as an input parameter taken from measurements in almost all existing plume models (e.g., see [77, 78] and references therein). Wang et al. [79] recently devel- oped a full particle kinetic Particle-in-Cell (PIC) simulation of the near thruster plasma plume. This fully kinetic model resolves the electron-ion coupling process and electron kinetics, and solves self-consistently the plume potential. However, due to computational constraints, the study in [79] considers a simplified plasma plume including only the pro- pellantplasma. In this dissertation, an experimental investigation on plume potential and density from an ion thruster type electron bombardment plasma source is presented. The objective is to understand processes and factors that influence the plume potential and density, including the plume expansion process and effects from the plume CEX plasma and facility back- ground plasma. The plasma plume wake expansion process is then controlled to experi- mentallysimulatelowangleofattacklunarplasmaflowandsurfacecharging. 1.3 LunarPhysicalCharacteristics The Moon is Earth’s nearest neighbor in space, with a mean semi-major axis of only 384,400km(60R e ). TheMoonisalsoremarkablymassivecomparedtoEarthitself,as their mass ratio is far larger than similar ratios of other natural satellites to their orbiting bodies. Further, the Moon’s equatorial radius is 1738 km ( .27 R e ), yielding a surface areaof37.9610 6 km 2 ,whichis1/4ofEarth’slandtotal. Thesefactorscombinetocreate the tidal resonance that locks the lunar “nearside” toward Earth, and keeps the “farside” completely hidden from Earth’s view. A summarized Earth to Moon physical property comparisonisgiveninTable1.1. The lunar atmosphere is approximately 14 orders of magnitude less than Earth’s, with an undisturbed gas concentration of only about 210 5 molecules/cm 3 during the lunar night, which falls to nearly 110 4 molecules/cm 3 during the lunar day. While localized 4 Table1.1: PhysicalcomparisonoftheMoonandEarth[29] Property Moon Earth Mass 7.35310 22 kg 5.97610 24 kg Meandensity 3.342g/cm 3 5.517g/cm 3 Equatorialradius 1738.14km 6378.14km Surfacearea(land) 37.9610 6 km 2 149.810 6 km 2 Equatorialgravity 1.62m/s 2 9.81m/s 2 Equatorialescapevelocity 2.38km/s 11.2km/s Siderealrotationtime 27.322days 23.9345hr Meanobliquity 6 41’ 23 28’ Meansurfacetemperature 107 Cday;-153 Cnight 22 C Temperatureextremes -233 Cto123 -89 Cto58 C Atmosphere 110 4 molecules/cm 3 day 2.510 19 molecules/cm 3 (STP) 210 5 molecules/cm 3 night MagneticField 0(smallpaleofield) 24-56A/m weakcrustalmagneticfieldshavebeenidentified,theyarethoughttobeonlyremnantsofa globaldipolefieldthatnolongerexists[42]. ThesetwopropertiesleavetheMoonexposed to any surrounding or oncoming plasmas, electromagnetic radiation, and charged particle flux. 1.4 LunarPlasmaEnvironment As the Moon orbits the Earth it encounters three distinct regions: Interplanetary space and thesolarwind,theEarth’smagnetosheath,andtheEarth’smagnetotail(Fig.1.2). However, at 60 R e the magnetosheath plasma properties are similar to that of the solar wind [43], while in the magnetotail conditions can vary greatly due to plasma sheet formation. This means that there are three different general plasma environments for the Moon: the solar wind, magnetotail, and plasma sheet, when encountered. In addition to the surrounding environment,localsurfacechargingalsodependsonthedirectionofthephotonandplasma fluxes,alongwiththephysicalpropertiesofthelunarsurface. 5 1.4.1 Earth’sMagnetotailandPlasmaSheet The majority of the lunar 27.32 day period is spent in interplanetary space or the magne- tosheath, but for approximately 6 days the Moon passes through Earth’s magnetotail. In Fig. 1.1 empirical models of the Earth’s bow shock (BS) and magnetopause (MP) bound- ariesaredisplayed[68]. Figure1.1: LunarorbitaboutEarthduringApollo17mission In this region either the tenuous plasma of the magnetotail lobes or the turbulent and energetic plasma sheet is encountered. Flow speed in these regions is generally assumed to be0 due to highly thermal conditions. Ambient plasma property comparisons to the solar wind are given in Table 1.2. It is important to note that the Moon’s position in the geomagnetic tail can vary as much as 5 R e with respect to the ecliptic plane, and that plasmasheetpropertiesarehighlydependentonmagneticfieldstability. 1.4.2 InfluenceoftheSun Carriedalonginterplanetarymagneticfieldlines,bothlowenergysolarwindparticlesand highenergyplasmaburstsinteractwiththelunarsurface. Thesolarwindisrelativelysteady 6 plasma emission composed primarily of protons (H + ) and electrons with energies on the orderoftensofeV.Itsstreamingvelocitygenerallyvariesbetween400and600km/s. Table1.2: Typicallunarambientplasmaproperties[24] Property Magnetotail PlasmaSheet SolarWind LunarWake e density 0.001-0.5cm 3 0.01-1cm 3 0.5-10cm 3 0.001-0.1cm 3 e temp <100eV 100eVto2keV 5-30eV 50-150eV Surfpoten -150to0V -1000to0V <20V -200to0V High energy bursts are sporadic, but potentially on the order of MeV’s, and originate from solar energetic particle (SEP) events such as coronal mass ejections, solar flares, or solar holes. Solar electromagnetic radiation flux also interacts with and affects the lunar surface, specifically at ultra-violet (UV), extreme ultra-violet (EUV), and X-ray wave- lengths. In the dissertation, an argon ion source is used to generate a mesothermal plamsa that simulatesaveragesolarwindplasmaconditions. 1.4.3 TheLunarWake Solarwindmagneticfieldspassthroughtherelativelynon-conductivelunarsurfaceessen- tially unimpeded, with only a slight inductive interaction with the more electrically con- ducting interior [22]. Thus, as the solar wind flows past the Moon the plasma is absorbed and neutralized, leading to formation of a plasma cavity and wake structure behind the lunar obstacle, first observed in situ by the WIND spacecraft in 1994 [52]. For any given temperature,electronthermalvelocityismuchgreaterthanionthermalvelocityduetothe much smaller electron mass. Therefore, plasma begins to expand into the resulting void fromtheflanks,ledbythermallyenergeticelectrons[68]. Ambientconditionsaregivenin Table 1.2. No such wake forms in the magnetotail or plasma sheet, due to sufficiently low plasmaflowspeed. 7 1.4.4 Additionalfluxsources Additional flux sources include galactic cosmic rays, interplanetary meteoroids, interplan- etary dust and hydrogen, and localized debris. However, in the context of lunar surface chargingthesesourcesarenegligible. 1.5 LunarRegolithSurface Created mainly by the continuous impact of small and large meteoroids over billions of years, an unconsolidated and fragmented rock layer known as the lunar regolith covers the Moon’s surface. It is the physical boundary between the solid Moon and external fluxes, and in most areas is 4 to 20 m thick [29]. The exposed uppermost portion is com- monly termed the lunar soil, and consists mainly of sub-centimeter particles. Currently analyzed soils have grain sizes between about 40 m and 800 m, and grain size aver- ages between 60 and 80 m, although the actual lunar soil is likely composed of grains with mean sizes between 45 and 100m [29]. Individual soil particles are primarily com- posed of agglutinates (aggregates of smaller soil particles bonded together by vesicular, flow-bandedglasscreatedbymeltingduringmicro-meteoroidimpacts),mineralfragments, pristine crystalline rock fragments, breccia fragments, and glasses of various kinds. All chemicalelementsthatmakeuptheEartharealsofoundinthelunarregolith,althoughthe relativeabundancesanddistributionsdiffergreatlybetweenthetwoplanets. Lunarsurfacephysicalpropertiesincludegeotechnicalpropertiessuchasspecificgrav- ity, relative density, compressibility, shear strenght, etc., as well as electrical and elec- tromagnetic properties. In terms of lunar surface charging, electrical properties are most significant, as they dictate material response to incident currents and describe electrical currentflow. Becausetheincidentcurrentsandanycurrentpropagatedthroughthesurface by potential gradients are DC (direct current), DC electrical conductivity, DC , is used to 8 measure how easily this flow occurs [29]. High electrical conductivity means that electri- cal current is easily passed through a material and charge accumulation does not readily occur, while low electrical conductivity implies low current flow, which allows a material to remain charged. Due to DC values generally on the order of those of glass (10 13 to 10 10 mho/m),lunarsurfacematerialisoftendescribedasdielectric. Thislowconductivity coupled with low loss tangents (0.001 to 0.1), the inherent dissipation of electromagnetic energy,allowsthesurfacetoreadilychargeandremainelectricallychargedforlongperiods oftime. However, DC hasdirectexponentialdependenceontemperature,andincreasesby a factor of10 1 under IR (infrared) irradiation, and10 6 under UV irradiation, the latter of which is comparable to an 800 C temperature increase [54, 29]. These factors indicate that electrical charge may flow across the moving lunar terminator boundary, providing a potential electrostatic charging mechanism responsible for dust levitation and transport. Additionally,moistureisknowntosignificantlyincreaseelectricalconductivity,whichmay affectsurfacecharginginpolar,shadowedcraterscontainingwaterorwaterice. Thelunar soil dielectric constant is dominantly controlled by bulk density (material mass within a given volume), and typically varies from 2 to 11 (unitless). A high dielectric constant increasesmaterialcapacitancebyallowingmorechargetobestoredatagivenvoltage. In this dissertation, planar dielectrics and dielectric dust surfaces are used to simulate thelunarregolith. Thematerialsusedarealuminasilicate,Mykroy/Mycalex R ⃝ glassmica, andJSC-1Alunarregolithsimulant. 1.6 LunarSurfaceChargingandDustCharging During an orbit, the Moon passes through the interplanetary environment dominated by thesolarwindandthroughtheterrestrialmagnetosphere’smagnetosheath,boundarylayer, lobe, plasma sheet, low latitude boundary layer, central plasma sheet, and plasma sheet 9 boundary layer (Fig. 1.2). Lacking both a significant global atmosphere and magneto- sphere, the lunar regolith surface is directly exposed to these space plasma environments and solar radiation, and is electrically charged by the ambient plasma and photoelectron andsecondaryelectronemission(Fig.1.3). Manystudieshavebeenperformedtodeterminethelocalplasmaenvironmentandlunar surface potential (see, for example, [59, 43, 20, 22, 26, 23, 24, 81, 17, 58] and references therein). Early studies suggested that the potential of the sunlit surface is typically a few tens of volts positive due to photoelectron emission [59, 43]. Night time potentials on the orderofahundredvoltsnegativearetobeexpectedwherethehotelectronfluxinthelunar plasma wake dominates the charging process [20]. More recent results using records from the Lunar Prospector spacecraft in low lunar orbit have confirmed the moderate positive surface potentials in sunlight and have shown that negative surface potentials on the order ofafewhundredvoltsmaybepresentinthelunarwake[22,26,24]. Duringsolarenergetic particle events, the surface potentials in the night time regions may even exceed a few kilovolts[23,25]. Bothsolarilluminationandplasmaflowsubstantiallyinfluencelunarsurfacecharging. It is expected that the influence of solar illumination and plasma flow is especially strong and complicated near the lunar terminator where the transition from sunlight-driven pos- itive surface potential to the plasma-charged negative surface potential occurs (Fig. 1.3). The plasma flow near the lunar surface is typically mesothermal (the directed plasma flow speedislargerthanionthermalspeedbutlessthanelectronthermalspeed). Thecombined effects from localized shadow generated by low sun elevation angle and localized wake regions generated by plasma flow over the rugged terrain can generate strongly differen- tiallychargedsurfaces[81]. A topic closely related to the plasma interaction and surface charging problem is the charging and transport of charged dust grains. Dust clouds suspended above the lunar sur- facewerefirstobservedashorizonglowbySurveyorspacecraft[61]andlaterbytheApollo 10 Figure1.2: LunarorbitthroughEarth’smagnetosphere(nottoscale) Figure1.3: Schematicofthelunarchargingenvironment(nottoscale)[68] Figure1.4: Chargingofthelunarregolithsurface 11 astronauts. Unexpected dust “storms” were also observed by the Lunar Ejecta and Mete- orite(LEAM)experimentdeployedbytheApollo17astronauts[8]. Thedustenvironment near the lunar surface is determined by the dynamic processes that levitate, launch, and transportdustgrains. Whilemicro-meteoroidimpactsand/ordisturbancesbyhumanactiv- itiescancontributetodustliftofffromthelunarsurface,electrostaticlevitationisgenerally acceptedastheprimarymechanism(see,forexample[1,34,65,66,67,69,70,46,81]and references therein). The local surface potential, electric field, plasma sheath, and the ini- tial dust charge largely determine whether electrostatic levitation would occur and what trajectoriesthedustgrainswouldfollow. Many studies have been carried out on the charging of regolith dust grains (see, for example [32, 85, 86, 73], and references therein). The focus of most ground experiments hasbeenonmeasuringthecharge-to-massratioofasingle,isolateddustgrain. Simulation studies involving dust charging have also only considered the charging model of a single, isolated dust grain. In a typical dust charging model, the dust capacitance is assumed to be that of an isolated sphere and plasma collection is calculated using Orbital Motion Limited(OML)theory(see,forexample,[50,40],andreferencestherein). Thedustcharge obtainedfromsingledustbasedmeasurementsandcalculationsisvalidonlyforthe“dust- in-plasma”situation,wheretheinter-dustdistanceismuchlargerthanthesheaththickness of each individual dust particle. For a “dust-in-plasma”, the dust charge is determined by the capacitance between a dust particle and its sheath boundary and plasma collection withineachindividualsheath,anddustparticlechargingisnotaffectedbytheneighboring dustparticles. Thisassumptioncannotbeappliedtoa“dustyplasma”wheretheinter-dust distanceismuchlessthanthesheaththickness. Italsocannotbeappliedtoadustysurface system,suchasthelunarregolith. ThelunarregolithsurfacechargingproblemmaybeillustratedinFig.1.4. Theproblem shown in Fig. 1.4 is different from the charging of a single, isolated dust grain or a solid surface. Onthedustlayersurface,theinter-dustdistanceisalmostzerobutthedustgrains 12 do not form a solid surface. The sheaths of each individual dust particle overlap to form onesinglesheathoverthesurface. Hence,thechargeofdustparticlesonaregolithsurface is strongly affected by that of the neighboring dust particles. One of the implications is that the difference in dust charging would lead to significantly different predictions for electrostaticlevitationofchargeddustanddusttransportdynamicsonthelunarsurface,as shown in a recent study by Wang et al. [82]. It is noted that NASA’s Lunar Atmosphere and Dust Environment Explorer (LADEE) designed to study the Moon’s thin exosphere andthelunardustenvironmentisscheduledtobelaunchedinAugust2013. In this dissertation, an experimental lunar simulator facility is established to simulate lunar mesothermal plasma flow over lunar surface objects at low angles of attack. Planar dielectricsanddielectricdustsurfacesusedtosimulatethelunarsurfacearechargedinthe lunar simulator facility. The objective is to understand how dust grains as part of a dust surface are charged under mesothermal plasma flux and in comparison to planar surfaces andisolateddustgrains. 1.7 DissertationOutlineandApproach Theremainingcontentofthisdissertationisgiveninthefollowingoutline: InChapter2theexperimentalmesothermalplasmaenvironmentischaracterized. In Chapter 3 conducting plates are charged in the main beam at varying angles of attackanddielectricmaterialsarechargedinthemainbeamatalowangleofattack. In Chapter 4 the lunar simulator facility is developed experimentally and validated withananalyticalmodelandnumericalcode. In Chapter 5 a systemic study of dielectric surface charging in the lunar simulator facilityisperformed. 13 InChapter6conclusionsandfutureworkaregiven. Compared to previous work, this dissertation includes a well-characterized mesother- mal plasma charging environment, develops a lunar simulator facility to simulate lunar plasma flow and surface charging at low angles of attack, employs planar dielectric and dielectric dust surfaces of the same chemical composition, and considers ion-induced sec- ondaryelectronemissionandionchargedepositionmechanisms. 14 CHAPTER 2: DEVELOPMENT AND CHARACTERIZATION OF A MESOTHERMAL PLASMA SOURCE Inordertoexperimentallysimulateandinvestigatethenearsurfaceplasmafieldandcharg- ingatthelunarterminator,itwasfirstnecessarytoestablishandcharacterizeamesothermal plasmaenvironment. Thischapterdetailsthevacuumfacility,plasmasource,anddiagnos- tictoolsusedtodoso,andpresentsafullcharacterizationoftheprimaryplasmabeamand background plasma. Additionally, the results obtained may be applied to predict plasma plumepotentialsduringin-flightthrusteroperation. 2.1 VacuumFacility Measuring3ft(.915m)indiameterand4ft(1.22m)inlength,acylindrical,stainlesssteel vacuum chamber is used to house all experimental apparatuses. An Alcatel mechanical pump is used for roughing, and high vacuum is reached through the use of a CVI TM500 cyropump. Floorvacuumpressureis110 7 Torr(1.310 5 Pa),and210 6 Torr(2.67 10 4 Pa)with4.0sccmofargongasflowingintothechamber. Thechamberisequipped withone8inchConflatport,sixISO-200ports,andelevenKF-40ports. 2.2 ElectronBombardmentGriddedIonSource For the vacuum facility dimensions and pumping capability described, a broad ion beam is suitable (a broad ion beam is typically several centimeters or more in diameter). The beam diameter is also much larger than the Debye length, which is the typical distance an 15 Figure2.1: Primaryresearchchamber electricfieldcanpenetrateintoaplasma. Tokeepabroadionbeamnearground(chamber) potential, it must be current neutralized (in a gridded source, when neutralizer emission equals ion beam current). Accordingly, a 4 cm diameter gridded ion source with a hot- filament neutralizer has been designed and fabricated (discussed in detail by Ding et al. and Ding [14, 13]). Figure 2.2 (a) and (b) show the ion source configuration and electrical circuit. To produce an ion beam, argon gas flows through the back of the ionization cham- ber and is ionized by thermal electrons emitted from the hot tungsten filament surface. In the region between the ring magnets and the back magnet, a magnetic field confines the electrons along magnetic field lines to enhance collisions with and ionization of the neu- tral argon gas. The ionization chamber is typically biased at 1100 V above ground. The anode cup, at 50 V higher than the ionization chamber, absorbs low energy electrons after collisions to maintain continuous ionization. The ion optics focus and accelerate the ions. A hot-filament neutralizer was located immediately downstream of the source exit plane. The enclosure is grounded to screen the accelerated beam from the high voltage internal components and prevent potential perturbations, and the neutralizer emits electron current to reduce space charge effects and arcing. Figure 2.2 (c) and (d) display the assembled 16 ion source with the screen grid, acceleration grid, neutralizer, and a grounded steel mesh enclosurefortestinginthevacuumchamber,andtheinteriorofthevacuumchamber. (a) Configuration (b) Electricalcircuit (c) Assembledsource (d) Vacuumchamberinterior Figure2.2: Custom-built4cmelectronbombardmentgriddedionsource 2.3 PlasmaDiagnostics In order to obtain plasma field parameters and accurately characterize the plasma environ- ment, an electrostatic Langmuir probe, two nude Faraday probes (radial and axial orienta- tions with respect to the ion beam), an emissive probe, and a retarding potential analyzer (RPA) were utilized. These probes yield plasma potentialϕ p , electron densityn e , ion den- sityn i , electron temperatureT e , and ion temperatureT i . The characteristics of the emitted 17 ion beam are that of a collisionless ( mfp;plasma >> L source ), mesothermal (v te >> v beam >> v ti ) plasma flow, where mfp;plasma , v te , v beam , and v ti are the plasma mean free path, electronthermalvelocity,beamvelocity,andionthermalvelocity,asdefinedinEqns.2.1- 2.4 mfp;plasma = 1 eAr n neu;exit = 16:4 cm >> L source (2.1) v te = √ kT e m e = 568 km=s (2.2) v beam = √ 2e(ϕ o ϕ p ) m i = 73 km=s (2.3) v ti = √ kT i m i = 2 km=s (2.4) where L source = 2 cm is the ion source characteristic length , eAr = 7.18 10 20 m 2 is the collisional cross-section between 4 eV electrons and argon [28], n neu;exit is the maximum neutral density at source exit (see Table 2.1), and ϕ o = 1100 V is the nominal beamvoltage. Note that magnetic field effects are negligible. The ion source discharge chamber samarium-cobalt magnets only confine discharge chamber electrons. From Ampere’s law, the induced magnetic field strength is on the order of 0.001 Gauss, while Earth’s surface magnetic field strength is 0.5 Gauss. This results in an electron gyroradius < L source , andaniongyroradius<<L source . 18 2.3.1 EmissiveProbe An emissive probe is an electrostatic probe with an exposed electrode that is electrically heated to the point of strong thermionic electron emission, resulting in a direct measure- ment of local plasma potential by recording the probe floating potential. Space charge effects are negligible in the absence of both a significant magnetic field and large density gradients (error has been quantified for plasma densities of 110 9 m 3 to 110 18 m 3 [36]). Given these two conditions and provided that emitted probe current is sufficiently large (much greater than incident electron current), a properly designed and fabricated emissive probe under the electron bombardment gridded ion source plasma densities will read the true plasma potential10%. Figure 2.3 displays the emissive probe constructed and utilized herein. The probe radius of curvature is 2 mm, and the tungsten filament diameter is 0.013 inches (0.33 mm). The probe is heated by an AC transformer (current controlled by a rheostat) and the transformer floating potential (probe floating potential) is then recorded to yield local plasma potentialϕ p . This process is not sensitive to plasma flowbecauseitdependsdirectlyonplasmapotentialratherthanelectronkineticenergyand it is less sensitive to probe surface contamination as the heated filament provides electron emission[3]. Figure2.3: Emissiveprobe: photographandelectricalschematic. 19 2.3.2 FaradayProbe A Faraday probe determines ion current density be measuring ion current flux to a fixed area collector. The design herein is a stainless steel nude Faraday probe, which is directly exposed to plasma flow. The stainless steel collecting surface directly faces the oncoming beam,allowingforcollectionofaxiallyflowingions. Theguardringisconcentricwiththe collectingsurface,withagapof1.0mmbetweenthetwo. Thepurposeoftheguardringis Guard ring Collecting surface Ceramic insulator V 120 kΩ -40V Figure2.4: Faradayprobeelectricalschematicandphotograph tocreateauniformsheathoverthecollectingsurfacebyminimizingedgeeffects. Thegap is chosen to be sufficiently small with respect to the Debye length to ensure overlap of the collectorandguardringsheathstoensureaflat,uniformsheathoverthecollector[4]. The collecting surface outer diameter is 5.0 mm, and the guard ring outer diameter is 8.0 mm. Boththecollectingsurfaceandguardringaredesignedtobebiasedtoanidenticalnegative potential, -40 V below the facility ground, as shown in Fig. 2.4, to prevent electrons from reachingthesurface. Measuredioncurrentatanypointinthefieldisrecordedbymeasuring the voltage drop across a resistor, and the value is divided by the collecting surface area to yieldioncurrentdensity,J i . J i = en i v i (2.5) 20 The conservation of energy, given by Eqn. 2.6, whereϕ o is the beam accelerating volt- age,isusedincombinationwithEqn.2.5andEqn.2.7toobtainn i ,givenbyEqn.2.8. 1 2 m i v i 2 = e(ϕ o ϕ p ) (2.6) v i = √ 2e(ϕ o ϕ p ) m i (2.7) n i = J i ev i (2.8) Under1100eVargonionbombardment,stainlesssteelnudeFaradayprobeerrorispri- marilyintroducedbyAugersecondaryelectronemissionfromthecollectorsurface,which mayaccountforupto20%ofthemeasuredionsignal[87]. 2.3.3 LangmuirProbe A Langmuir probe is a device used to measure the electron temperature T e , electron den- sity n e , and electric potential ϕ p of a plasma. A typical probe consists of one or more electrodes inserted in the plasma environment, and the electric potential of each electrode iseitherheldconstantortime-variedwithrespecttotheplasmapotential. Measuredcurrent collectionandprobepotentialdatacanthenbeanalyzedtodeterminephysicalpropertiesof theplasma. Inordertoaccuratelysizeaprobeandapplyappropriatetheory,itisnecessary toquantifythesheaththickness. Aplasmainherentlypossessestheabilitytolocallyshield appliedelectricpotentials,whichcausesacharge“cloud”tosurroundanyobject,suchasa Langmuir probe, that is perturbing quasi-neutrality. The thickness of this cloud, called the sheath,isgivenbytheDebyelength, d = √ ϵ o kT e n e e 2 (2.9) 21 whichisameasureoftheshieldingdistance. Iftheproberadiusisusedasthecharacteristic length,L o ,itcanbedeterminedifthesheathisrelativelythickorthin. Inathinsheathcase (L o >> d ) local charge concentrations are shielded out in a short distance compared to the probe radius, and all particles that enter the sheath are assumed to be collected by the probe. In a thick sheath case (L o << d ) current collection is assumed to follow orbit- motion limited (OML) theory, meaning that only a percentage of particles entering the sheatharecollected. 5.5 kΩ V Sweep Voltage Cylindrical Langmuir probe Figure2.5: Langmuirprobeelectricalschematicandphotograph The plasma generated by the electron bombardment source has a Debye length on the order of 5 mm in the plume. Based on this value OML theory applies, as the designed Langmuir probe radius is 0.5 mm. It is also a single electrode, cylindrical, non-emitting probewithalengthof5mmandasweepvoltagerangeof-30Vto60V,andisdisplayedin Fig.2.5. Thevoltagedropoveraresistorisreadandrecordedforeachsweepvoltagevalue, whichallowsanI-Vcurve(Fig.2.6)tobegenerated. Fromthiscurveϕ p isreaddirectly,T e in eV is calculated from the slope of the “T e line” by Eqn. 2.10, andn e is calculated from Eqn.2.11 T e = 1 slope(T e line) (2.10) 22 −40 −20 0 20 40 60 80 −0.5 0 0.5 1 1.5 2 2.5 3 3.5 x 10 −6 I −V curve V (V) I (A) −10 0 10 20 30 40 50 60 70 −19 −18 −17 −16 −15 −14 −13 −12 V (V) LnI (A) T e line Φ p Figure2.6: ExampleLangmuirprobeI-Vcurve n e = I esat eA probe √ 2m e eT e (2.11) whereI esat istheelectronsaturationcurrentandA probe istheLangmuirprobesurfacearea. In a flowing plasma, such as an ion beam, the Langmuir probe potential measurement maybeperturbed,necessitatingtheuseofanemissiveprobe. Previousstudieshaveshown thatcylindricalLangmuirprobessimilartotheoneusedintheseexperimentshaveageneral accuracyof20%[89]. 2.3.4 RetardingPotentialAnalyzer A retarding potential analyzer (RPA) is an electrostatic plasma diagnostic instrument used to measure ion energy distribution. An RPA consists of a current collector that is typically shieldedfromtheupstreamplasmabyaseriesofbiasedgrids. Theinstrumentonlyallows ions with energy to charge ratios higher than the retarding potential to reach the collector (soitdoesnotdistinguishbetweensinglyandmultiply-chargedions). Following standard guidelines [4], a three-grid RPA was designed and fabricated (a schematicandphotographsaregiveninFig.2.7). Thefirstgridisfloatingtoreduceplasma perturbations, the second grid is biased to -40 V to repel electrons, the third grid is swept 23 through the maximum expected ion retarding voltage, and the back plate is biased neg- atively to ensure ion collection. The mesh opening dimension is chosen to be less than the sheath thickness over the electron repelling grid, to ensure that electrons will not pass throughthegriduninfluencedbythegridbias. Thegridspacingischosentobelessthanthe sheath thickness over the ion retarding grid to mitigate space charge effects. The spacing between the RPA opening and the collector plate sets the collimation angle at 9.3 which ensures that flowing ions arrive nearly perpendicular to the collector. Ceramic insulators isolategridsfromeachotherandthecollector. Thegridsandcollectorarestainlesssteel. (a) Schematic (b) Gridsandfullassembly Figure2.7: RPAschematicandphotographs 24 0 200 400 600 800 1000 1200 1400 1600 −2 0 2 4 6 8 10 12 14 16 18 x 10 −8 RPA collector current vs. ion retarding voltage Ion retarding voltage, V Collector current, A (a) Collectorcurrent 0 200 400 600 800 1000 1200 1400 1600 −1 0 1 2 3 4 5 6 7 8 9 x 10 −10 Main beam ion energy distribution eV dI/dV (A/V) (b) Ionenergydistribution Figure2.8: RPAmainbeamexperimentaldata To obtain ion energy distribution, collector current is first plotted against ion retarding voltage, as shown in Fig. 2.8 (a). The derivative dI dV is then taken, which yields the ion energydistributionperEqn.2.12[4] dI dV =A c q i n i √ 2q i V ionr m i f(V ionr ) (2.12) where A c is the RPA collector area, q i is the ion charge, and V ionr is the ion retarding voltage. Fig. 2.8 (b) displays the resulting ion energy distribution for main beam ions. The most probable ion energy (voltage at the peak of the distribution) closely matches the inputted beam voltage ϕ o = 1100 V. Ion temperature T i is obtained directly from the ion energydistributionspread. ErrorisprimarilyintroducedbyAugeremissionfromtheRPA collectorsurface. 2.3.5 RetardingPotentialAnalyzerIonFocusingEffect Throughout RPA development and testing an anomalous phenomenon was observed. As ion retarding voltage is increased, the current collected by the back plate should theoreti- cally continually decrease. Instead, the current-voltage (I-V) curve consistently increased 25 with increasing ion retarding voltage, before decreasing sharply as the most probable ion energy was approached. This phenomenon is displayed in Fig. 2.9. In order to explain the anomalousphenomenon,severalexplanationswereconsidered,including: Townsenddischargetheory[38,47]-meaningthatsecondaryelectronsemittedfrom the stainless steel backplate by incident ions are being driven by the electric field to the respective electrodes and collide with neutral gas to produce new ions and elec- trons(similartoTownsenddischarge). Thenewionsandthesecondaryelectronswill produce extra collector current. However, the mean free path of electron collisions with the neutral gas was estimated to find that the Townsend discharge process will notoccur. Sheath over the RPA inlet- considers sheath effect on incident ion trajectories that could influence the collector current. However, simulation results show that the potential drop within the sheath is only several volts and has little effect on the tra- jectoriesoftheincidentions[60]. Secondary electron emission [71, 15] from the collector- considers that secondary electronemissionfromthestainlesssteelcollectorbyincidentionimpactsisrespon- sible for the extra collector current. A special experiment was designed to test the assumption. A negative bias was applied to grid 1 to repel electrons, grid 2 was sweptthroughtheionretardingvoltage,andgrid3wasnegativelybiasedtosuppress secondary electron emission . However, the ascending trend in resulting I-V curves remained. This experiment firmly indicated that secondary electron emission is not thereasonoftheanomalousI-Vcurvephenomenon. Finitecollectorsizeeffect[74]-considersthatsomeincidentionswillimpactthewall betweentheionretardinggridandbackplatebecauseofthefinitesizeofthecollector. 26 However, simulation results show that this effect will only reduce the descending trendoftheI-Vcurve[60]. Next, considering that I-V curves and collector current magnitude are not symmetrical asafunctionofradiallocation(Fig.2.9),ionbeamparameterssuchasdensityanddirection are therefore also not symmetrical. Since the ion density influences collector current mag- nitude,iondirection(velocityvectors)arelikelytoaffecttheshapeoftheI-Vcurve. Totest this theory the RPA was located 8 inches downstream of the source and rotated about the r axis, as shown in Fig. 2.10. Additionally, the RPA inner wall between the ion retarding gridandthecollectorwascoatedwithaluminumfoil(Fig.2.10). Thefoilwasbiasedtothe same negative potential as the collector to quantify ion current to the RPA wall. Current collectedbytheelectronrepellinggridwasalsorecorded. Eachcurrent(electronrepelling grid, wall, collector) was recorded as a function of ion retarding voltage. The results for alpha = -10 to 15 in 5 increments are shown in Fig. 2.11, where I SUM is the sum of the back plate (collector) and wall (aluminum foil) ion currents. The fact that I SUM ascends, descends, or is relatively flat as a function of incident angle alpha indicates that ion focusing between the electron repelling grid and the ion retarding grid is the driver of theobservedanomalousI-Vcurvebehavior. Aquantitativeanalysisofthe5 alphacaseisgiveninFig.2.12. Theoretically,theback plateshouldcollectcurrentI c perEqn.2.13 I c = en i v i A c %open (2.13) where %open = 52% is the grid transparency. Experimentally I c is typically only 5% of the incident flux, as opposed to 52%. The primary reason for this reduction is grid misalignment, but that does not explain the I-V curve anomalous behavior. The floating grid by definition does not record current, but test results have shown that it does absorb 60% of incident ions. Current collected by the electron repelling grid as displayed in 27 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 x 10 −6 Ion Retarding Potential, V Collector current, A RPA I−V Trace (a) -1”inradialdirection 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 x 10 −6 Ion Retarding Potential, V Collector current, A RPA I−V Trace (b) -0.5”inradialdirection 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 x 10 −6 Ion Retarding Potential, V Collector current, A RPA I−V Trace (c) 0”inradialdirection 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 x 10 −6 Ion Retarding Potential, V Collector current, A RPA I−V Trace (d) 0.5”inradialdirection 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 x 10 −6 Ion Retarding Potential, V Collector current, A RPA I−V Trace (e) 1”inradialdirection 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 x 10 −6 Ion Retarding Potential, V Collector current, A RPA I−V Trace (f) 1.5”inradialdirection Figure2.9: RPAI-Vtraces8inchesdownstreamof6mAbeam 28 ! z r Measurement Point RPA Argon Ion Source Three-dimensional Movement Stand Vacuum Chamber Biased Al foil between ion retarding grid and collector to collect current to RPA inner wall alpha Figure2.10: RPAionfocusingeffectexperimentalsetup Fig.2.12istypically15%ofincidentionsforvoltageslessthentheionretardingvoltage, and30% of incident ions for voltages greater than the ion retarding voltage (due to the fact that ions are rejected and turned around by the ion retarding grid at sufficiently high voltages). The remaining flux is adsorbed by the ion retarding grid itself, and the wall and back plate. Figure 2.12 shows that the ion retarding grid (high voltage grid) could be collecting up to twice as much current as the wall and back plate combined. Thus, ion focusing as a function of ion retarding grid voltage to the ion retarding grid itself is the likely driver of the anomalous I-V curve behavior. Note that the high voltage power supplyandDAQequipmentavailablefortheRPAexperimentswerenotcapableofdirectly recordingionretardinggridcurrentasafunctionofionretardingvoltage. 2.3.6 DataAcquisitionandTraversingSystem Dataiscollectedandrecordedwithintheexperimentaldomainusingtheprobesuite,atwo- dimensional traversing system, and a data acquisition (DAQ) and control system. Driven bysteppermotors,thetraversingsystemanditsrangeofmotionaredisplayedinFig.2.13. The system is capable of a linear resolution of 1 mil in either direction. The probe suite is mountedonthetraversingsystemarmandisdisplayedinFig.2.14. 29 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I−V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha −10 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING (a) -10 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I−V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha −5 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING (b) -5 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I−V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha 0 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING (c) 0 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I−V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha 5 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING (d) 5 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I−V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha 10 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING (e) 10 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I−V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha 15 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING (f) 15 Figure 2.11: RPA grid and collector currents for varying alpha, 8 inches downstream of 6 mAbeam 30 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 x 10 −6 I −V curve V (V) I (A) I BACK PLATE I WALL I SUM 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 x 10 −5 alpha 5 deg V (V) I (A) I FLOATING GRID I ELECTRON REPELLING 1.25e-5 A 3e-5 A 4.5e-6 A 4.15e-6 A Indicates that 12.5e-6 A – 4.15 e-6 A = 8.35e-6 A could be collected by HV grid Figure2.12: QuantitativeanalysisofRPAI-V5 alphacase Each motor is controlled through LabVIEW and a unipolar stepper motor driver board by a signal sent from a computer’s parallel port. The control program is written such that a user can input a path file describing the desired motion in both directions, and data can be collected at each point. The data acquisition and traversing system control loops are displayed in Fig. 2.15 (from [13]). Motion signals sent to the stepper motor control board andtraversingsystemoperateinanopenloopconfiguration. Meanwhile,aNIDAQcardis usedtogeneratea5VsignalthatisamplifiedandsenttotheLangmuirprobe. Scandata from the Langmuir probe, Faraday probes, emissive probe, and RPA are then sent back to the computer using an Agilent 34970A data log unit, and are written to a text file for post processing. 31 12” 16” Probe Suite Figure2.13: Traversingsystem 2.4 IonBeamandBackgroundPlasmaCharacterization Anexperimentalinvestigationonplumepotentialanddensityfortheionthrustertypeelec- tronbombardmentplasmasourceispresentedinthissection. Theobjectiveistounderstand processes and factors that influence the plume potential and density, including the plume expansionprocessandeffectsfromtheplumeCEXplasmaandfacilitybackgroundplasma. 2.4.1 ExperimentalSetup Figure 3.4 shows the testing schematic in the vacuum chamber. The plasma source was placedoppositethecryogenicpumpalongchambercenterline. A2Dscanoftheplumefield wasachievedbytraversingtheprobesuite. Theentirescanningprocessisfullyautomated. Each probe was traversed within the “Scan region” of 0.3810 m by 0.2286 m (Fig. 2.16) 32 Flowing ions Thermal electrons RPA EP LP Axial FP Radial FP Figure2.14: Probesuite whichwasdividedinto160measurementpointswithaspatialresolutionof2.54cminboth zandr. Acompletescanofthe160measurementpointstakesabout1hr. 2.4.2 PlumeMeasurements Plume measurements were obtained under different plasma source operating conditions. TheoperatingconditionsformeasurementsdiscussedinthissectionarelistedinTables2.1 and 2.3. The measured and normalized ion density, electron density, and electron tem- perature at thruster exit for the three current emission operating conditions are shown in Table2.2. Asdiscussedpreviously,theplumeplasmaiscomposedofthepropellantplasma emitted by the source and the CEX plasma generated in the plume. The values shown in 33 Figure2.15: DAQandcontrolsystem Figure2.16: Experimentalsetup Table 2.2 primarily come from the propellant plasma. Plasma measurements made down- stream of the source exit will include contributions from the propellant plasma, the plume CEX plasma, and the facility background plasma. Measurement error given in the tables andplotsisduetoprobeerrorasdiscussedinthePlasmaDiagnosticssection. Figure 2.17 shows normalized 2D contours of the plasma potential ϕ p , ion density n i , electron density n e , and space charge n i - n e obtained from the scan measurement for the 34 (a) Plasmapotentialϕ p (b) Iondensityn i (c) Electrondensityn e (d) Spacechargen i -n e Figure2.17: Plasmaplumenormalized2Dcontours,I =10mA medium current operating condition (10 mA). Figure 2.18 (a) (c) (e) shows the radial pro- files of electrical potential at three downstream distances, ∆z=3.8R b , 12.7R b , and 22.9R b , whereR b is the beam radius. Note that the mass flow rate for these three current emission settingswas4sccm. Table2.1: Massflowratevariationoperatingconditions I,mA P c ,Torr _ m,sccm n neu;exit ,m 3 cex;exit ,mm div ,deg 10. 2.310 6 1.2 11.6% 7.61:510 18 46090 21.6 10. 3.810 6 4.1 3.4% 2.90:610 19 12020 16.2 10. 8.010 6 10.0 1.4% 7.11:410 19 5010 10.8 35 Table2.2: Measuredandnormalizedplasmaparametersatthrusterexit Property I =2mA Normalized I =10mA Normalized I =18mA Normalized Exitplanen i 1.23:1410 14 m 3 0.415 6.15:6810 14 m 3 0.415 1.11:1210 15 m 3 0.415 Exitplanen e 2.97:5910 14 m 3 1.0 1.48:3010 15 m 3 1.0 2.67:5310 15 m 3 1.0 ExitplaneT e 2-7eV 1.0 2-7eV 1.0 2-7eV 1.0 Exitplane d 0.929:094mm 1.0 0.416:042mm 1.0 0.310:031mm 1.0 v i;avg 72,680m/s 0.137 72,680m/s 0.137 72,680m/s 0.137 v te;avg 532,400m/s 1.0 532,400m/s 1.0 532,400m/s 1.0 Table2.3: Plasmasourcecurrentemissionoperatingconditions I,mA P c ,Torr _ m,sccm V fil ,V I fil ,A ϕ o ,V 2 3.610 6 4.1 6.75 8.0 1100 10 3.810 6 4.1 8.0 9.1 1100 18 3.510 6 4.0 9.0 9.5 1100 2.4.3 PlumeExpansionvs. PlumePotential The propellant plasma emitted by the plasma source forms a mesothermal plasma flow along the emitting direction. First the effect of the plume expansion on plume potential is investigated. The expansion of a quasi-neutral mesothermal plasma into vacuum is well understood (for instance, see [80] and references therein). Under the assumption that the electrons in the expansion are an isothermal fluid, such expansion can be solved analytically from Eqn.2.14[80] ln n i n io = eϕ p kT e = √ M 2 o 1 2 2 (2.14) where n io is the plasma density at exit, M o = v io /C s is the Mach number of the source ion flowcalculatedusingtheionvelocityatexitv io andtheionacousticvelocityC s =(T e /m i ) 1=2 , and is the radial expansion angle with respect to the flow direction. The analytical solutionusingtheisothermalelectronfluidassumptionistheupperboundofpotentialdrop associatedwithmesothermalplasmaexpansion. Thissolutionmaybeusedtodescribethe potentialdropassociatedwiththepropellantplasmaexpansion. 36 In Fig. 2.18 (b) (d) (f), the analytical expansion solution is overlaid with the measured radial plume profile. Note that the plasma potential profiles initially follow the analytical solution. However, the measured profiles abruptly flatten outside the beam core region. This deviation from the analytical expansion solution indicates that the expansion of the propellant plasma is overtaken by other effects outside of the primary beam region. This is to be expected as the expansion process will be terminated when the density of the expanding plasma has been reduced to that of the background plasma density. Outside the beam core region, the “background plasma” includes the plume CEX plasma and the facilitybackgroundplasma. 2.4.4 EffectsofPlumeCharge-ExchangePlasma NexttheeffectoftheplumeCEXplasmaonplumepotentialisinvestigated. Theanalytical model [4] given in Eqn. 2.15 was used to estimate neutral density at the source exit plane n neu;exit , where is the ionization rate, c n is the speed of neutrals, and A orifice is the total exitgridopenarea. Thesubsequentexitplanemeanfreepath cex;exit calculationbasedon Ar-Ar + charge-exchange cross section cex = 3.02510 19 m 2 [16] is given in Eqn. 2.16, andvaluesaredisplayedinTable2.1. n neu;exit = (1) _ m m i c n A orifice (2.15) cex;exit = 1 cex n neu;exit (2.16) In each case cex;exit L chamber , where L chamber = 457.5 mm is the chamber character- istic length (chamber radius), illustrating the CEX ion production relevance. The effect of neutraldensity(henceCEXproduction)ontheplumeandbackgroundplasmaenvironment wasexperimentallytestedbyfixingtheemittedbeamcurrentI =10mAwhilevaryingthe argonmassflowrate _ mto1,4,and10sccm. 37 −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 −10 −5 0 5 10 15 20 25 30 35 r, normalizied Plasma Potential, normalized Normalized Plamsa Potential vs. r for near−axial scan plane (z1) 2 mA 10 mA 18 mA (a) Z1=3.8R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 0 10 1 10 2 r, normalizied Plasma Potential, normalized, log scale Normalized Plamsa Potential vs. r for near−axial scan plane (z1) 2 mA 10 mA 18 mA 2 mA, expansion solution 10 mA, expansion solution 18 mA, expansion solution (b) Z1=3.8R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 −10 −5 0 5 10 15 20 25 30 35 r, normalizied Plasma Potential, normalized Normalized Plamsa Potential vs. r for mid−axial scan plane (z2) 2 mA 10 mA 18 mA (c) Z2=12.7R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 0 10 1 10 2 r, normalizied Plasma Potential, normalized, log scale Normalized Plamsa Potential vs. r for mid−axial scan plane (z2) 2 mA 10 mA 18 mA 2 mA, expansion solution 10 mA, expansion solution 18 mA, expansion solution (d) Z2=12.7R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 −10 −5 0 5 10 15 20 25 30 35 r, normalizied Plasma Potential, normalized Normalized Plamsa Potential vs. r for far−axial scan plane (z3) 2 mA 10 mA 18 mA (e) Z3=22.9R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 0 10 1 10 2 r, normalizied Plasma Potential, normalized, log scale Normalized Plamsa Potential vs. r for far−axial scan plane (z3) 2 mA 10 mA 18 mA 2 mA, expansion solution 10 mA, expansion solution 18 mA, expansion solution (f) Z3=22.9R b Figure 2.18: Radial potential profiles at 3 downstream locations (left); Comparison of potentialprofilestoanalyticalexpansionsolution(right) 38 Inthebeamregion,theiondensityn i isobtainedutilizingtheaxialFaradayprobemea- surement and the thin sheath assumption from Eqn. 2.8, wherev i is the velocity of source ions. Outside of the beam region, the ion density is obtained utilizing the radial Fara- dayprobemeasurementandthethicksheathorbit-motionlimitedtheory(duetodecreased electron density) from Eqn. 2.17, where J OML is given in Eq. 2.18 and v cex is CEX ion velocity based on the electrostatic potential difference between the plume centerline and radialbackground[48]. TheparameterQ i isgiveninEqn.2.19. n i = J OML ev cex (2.17) J OML = Ji [ 2 √ Q i +exp(Q i )erfc ( √ Q i ) ] (2.18) Q i = eϕ FP kT i (2.19) whereϕ FP is the radial Faraday probe bias = -40 V and CEX ion temperatureT i =3 eV isobtainedfromradialRPAmeasurements. Figure 2.19 displays the resulting measured n i and n e radial profiles for each flow rate. The 10 sccm data set clearly indicates that background plasma density scales as a function of _ m. Furthermore, for this data set, at the thruster exit cex;exit << L chamber , which suggests that this population is primarily comprised of low energy CEX ions and accompanyingelectrons. To explain the similarity between the 1 sccm and 4 sccm data sets, a separate experi- ment was conducted for each flow rate to obtain the 95% half-angle divergence div , per Eqn.2.20[44], 0:95I = 2r 2 sweep ∫ div 0 J i ()sind (2.20) 39 where I is beam ion current = 10 mA and a 90 constant radius r sweep = 40 cm sweep through the beam plume was performed. Ion current densityJ i was obtained from a Fara- day probe mounted on a rotating arm as shown in Fig. 2.20. As an example, the resulting plotforthe4sccmcaseisgiveninFig.2.21,andthebeamhalf-anglesaretabulatedinthe right most column of Table 2.1. The half-angles confirm that the high energy source ions are primarily contained to the beam plume, and also show the 1 sccm case to be the most divergent. This is due to an increase in the acceleration gap sheath thickness which leads to increased acceleration grid interception current and beam divergence [21], causing the backgroundplasmadensityprofilesofthe1and4sccmcasestoapproachoneanother. The half-angles also show that the optimal flow rate range of this source at a 10 mA discharge current is 4 to 10 sccm (above 10 sccm the sheath will flatten to the point that interception currentandbeamdivergenceincreasesignificantly). To conclusively differentiate between high energy (primary beam) and low energy (CEX) ions, the RPA was located normal to the beam flow direction, along the radial line Z1. First, a high voltage sweep through beam potential confirmed the absence of radially flowingprimarybeamions. Second,alowvoltagesweep(-20Vto20V)underthe1sccm and 10 sccm _ m conditions confirmed the presence of CEX ions, as displayed in Fig. 2.22. CEXionsarecreatedatandflowoutradiallyatplasmapotential,andthepotentialindicated bytheRPAI-Vtraceswasindependentlyvalidatedbyemissiveprobemeasurements. Using results from a DSMC code to obtain estimates of the local neutral density n neu forthe4sccmcase,theCEXproductionrate _ n cex nearsourceexitwascalculatedfrom[76] _ n cex = n neu n i v i cex = 1:3710 19 1 m 3 s (2.21) Next, assuming a CEX line source production rate and a 1/2 sphere CEX exit surface area A halfsphere outwards over the chamber volume, an analytical model of CEX density n cex 40 −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion Density, normalized, log scale Normalized Ion Density vs. r for near−axial scan plane (z1) 10 mA (1 sccm) 10 mA (4 sccm) 10 mA (10 sccm) (a) Z1=3.8R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Electron Density, normalized, log scale Normalized Electron Density vs. r for near−axial scan plane (z1) 10 mA (1 sccm) 10 mA (4 sccm) 10 mA (10 sccm) (b) Z1=3.8R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion Density, normalized, log scale Normalized Ion Density vs. r for mid−axial scan plane (z2) 10 mA (1 sccm) 10 mA (4 sccm) 10 mA (10 sccm) (c) Z2=12.7R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Electron Density, normalized, log scale Normalized Electron Density vs. r for mid−axial scan plane (z2) 10 mA (1 sccm) 10 mA (4 sccm) 10 mA (10 sccm) (d) Z2=12.7R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion Density, normalized, log scale Normalized Ion Density vs. r for far−axial scan plane (z3) 10 mA (1 sccm) 10 mA (4 sccm) 10 mA (10 sccm) (e) Z3=22.9R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Electron Density, normalized, log scale Normalized Electron Density vs. r for far−axial scan plane (z3) 10 mA (1 sccm) 10 mA (4 sccm) 10 mA (10 sccm) (f) Z3=22.9R b Figure 2.19: Radial ion density as a function of _ m (left); Radial electron density as a functionof _ m(right) 41 0.40 m Figure2.20: Faradayprobemountedonrotatingarmforhalf-angledivergenceexperiment −80 −60 −40 −20 0 20 40 60 80 0 0.05 0.1 0.15 0.2 Ion current density as a function of angle, r = 40 cm Angle from thruster centerline, deg Ji (mA/cm 2 ) mass flow = 4.1 sccm, 95% half−angle divergence = 16.2 deg Figure2.21: Half-anglecurrentdensitysweep,10mA,4.1sccm based on Gauss’s law was derived as follows as a function of radial distance r volume from thesourcecenterline,andisgiveninEqn.2.24. A halfsphere = 2r 2 volume (2.22) flux = _ n cex = n cex ∮ dA halfsphere v cex (2.23) 42 −20 −15 −10 −5 0 5 10 15 20 −2 0 2 4 6 8 10 12 14 16 18 x 10 −9 RPA current vs. Ion Retarding Voltage Voltage, V Grid current, A −20 −15 −10 −5 0 5 10 15 20 −2 0 2 4 6 8 10 12 14 16 18 x 10 −9 RPA current vs. Ion Retarding Voltage Voltage, V Grid current, A 1 sccm 10 sccm 10 sccm cubic fit 1 sccm cubic fit Local plasma potential, 10 sccm Local plasma potential, 1 sccm Local plasma potential, 10 sccm Local plasma potential, 1 sccm Figure2.22: RPAI-Vcurvesfor1and10sccmoperatingconditions Table2.4: NSTARto4cmgriddedionsourcecomparison. Thruster _ mtotal _ mneutral _ n cex Custom4cm 4.1 3.4% 3.96 1.37:1510 19 NSTAR 23.87 88% 2.86 1.010 19 n cex = _ n cex v cex 2r 2 volume (2.24) The model is plotted against the 4 sccm background ion density in Fig. 2.23 and shows strong agreement, validating the experimental methods and assumptions included herein. Note that the CEX ion production of the custom built 4 cm electron bombardment grid- ded ion source is comparable to that of ion thrusters that have flown in space, such as NSTAR [76], despite the drastic propellant utilization efficiency difference, as shown in Table2.4. 43 −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion density, normalized, log scale Normalized Ion density vs. r for near−axial scan plane (z1) 10 mA, 4 sccm, 1/2 sphere surface area model 10 mA, 4 sccm, experimental data Figure2.23: CEXiondensitymodelcomparison 2.4.5 EffectsofFacilityBackgroundPlasma Finally,theeffectofthefacilitybackgroundplasmaonplumepotentialisinvestigated. The facilitybackgroundplasmacomefromthefollowingsources: Secondaryelectronemissionorelectronbackscatteringfromthechamberwalls Ionsputteringorioninducedelectronemissionfromthechamberwalls Ionization of beam or background (tank) neutrals by the neutralizer or ionization of tankneutralsbysourceions As the primary beam electrons generated by the neutralizer are of low energy (2 to 7 eV), both secondary electron emission and electron backscattering from the chamber wall can be neglected due to the low electron emission yield for low energy electron impinge- ment [12, 84]. Furthermore, the grounded stainless steel chamber walls act as an electron sink for the secondary electrons generated. By comparison, the primary beam argon ions are heavy and energetic (1100 eV), which can lead to ion sputtering and ion-induced 44 −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion Density, normalized, log scale Normalized Ion Density vs. r for near−axial scan plane (z1) 10 mA (4 graphite panels) 10 mA (2 graphite panels) 10 mA (0 graphite panels) (a) Z1=3.8R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Electron Density, normalized, log scale Normalized Electron Density vs. r for near−axial scan plane (z1) 10 mA (4 graphite panels) 10 mA (2 graphite panels) 10 mA (0 graphite panels) (b) Z1=3.8R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion Density, normalized, log scale Normalized Ion Density vs. r for mid−axial scan plane (z2) 10 mA (4 graphite panels) 10 mA (2 graphite panels) 10 mA (0 graphite panels) (c) Z2=12.7R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Electron Density, normalized, log scale Normalized Electron Density vs. r for mid−axial scan plane (z2) 10 mA (4 graphite panels) 10 mA (2 graphite panels) 10 mA (0 graphite panels) (d) Z2=12.7R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Ion Density, normalized, log scale Normalized Ion Density vs. r for far−axial scan plane (z3) 10 mA (4 graphite panels) 10 mA (2 graphite panels) 10 mA (0 graphite panels) (e) Z3=22.9R b −1600 −1400 −1200 −1000 −800 −600 −400 −200 0 10 −4 10 −3 10 −2 10 −1 10 0 10 1 r, normalizied Electron Density, normalized, log scale Normalized Electron Density vs. r for far−axial scan plane (z3) 10 mA (4 graphite panels) 10 mA (2 graphite panels) 10 mA (0 graphite panels) (f) Z3=22.9R b Figure2.24: Radialiondensityvs. numberofgraphitepanels(left);Radialelectrondensity vs. numberofgraphitepanels(right) 45 Figure2.25: 61cmx61cmgraphitetargetatchamberrear secondary electron emission as a function of incident ion angle and target properties. As it is a charge neutral process, sputtering in this experiment will not significantly alter the backgroundplasmadensity. Initialestimationofioninducedelectronemissionfromthealuminumgatevalveatthe rear of the chamber led to the concern that aluminum surface emission might be a signifi- cantcontributortothefacilitybackgroundplasma. Hence,a61cmby61cmgraphitetarget (primarybeamradiusatthisdownstreamlocationis15cm)wasinstalledatchamberrear during testing to investigate the effects of ion induced electron emission (Fig. 2.25). The measurements obtained with graphite panels installed are shown in Fig. 2.24. The results do not indicate a reduction of background plasma density. In fact, the results indicate that background plasma density was slightly increased due to a reduction in effective pumping speedbythegraphitepanelgeometrywhichincreasedneutraldensityandCEXproduction. Hence, it is found that the contribution from ion-induced secondary electron emission to thefacilitybackgroundplasmaisnotsignificant. 46 Lastly, measurements were performed using varied neutral argon flow over the neu- tralizer. Results showed that the ionization of both tank and beam neutrals by the neutral- izer are negligible with respect to background CEX plasma. Additionally, calculations of the mean free path cex;avg (Eq. 2.26) using a chamber-averaged neutral density n neu;avg (Eq. 2.25) show that CEX ion production due to average background neutral density (i.e. tankneutrals)isalsonegligible,as cex;avg >>L chamber , n neu;avg = 133:3 P c kT neu (2.25) cex;avg = 1 cex n neu;avg = 27:0m (2.26) whereP c is average chamber pressure in Torr andT neu is the temperature of neutral argon inKelvin. 2.5 SummaryandConclusions An experimental investigation is carried out to investigate the plasma plume potential and density. A capability for efficient 2D measurements of the plasma plume is developed and 2Dprofilesofaplumeemittedfroma4cmelectronbombardmentionsourceareobtained utilizing a Langmuir probe, Faraday probes, an emissive probe, and an RPA. It is found that the magnitude of the plume potential is controlled primarily by the factor that termi- nates the propellant plasma expansion process. In these experiments, the density of the facility background plasma is much less than that of the plume CEX plasma in the region outside of the beam. Hence, the plume CEX plasma will significantly modify the expan- sion of the propellant plasma in the region outside of the beam, and is the dominant factor for terminating the plume expansion process. The fact that the termination of the plume expansion process is controlled by the CEX plasma generated by the ion source suggests that this ground based measurement can be applied to predict the plasma plume potential 47 for in-flight conditions. For an ion thruster operating in a near vacuum environment, the magnitudeofthepotentialdropfromtheplumetotheambientiscontrolledbythepropel- lant CEX ion production rate. As the CEX ion production of the custom built gridded ion thruster used in this study is comparable to thrusters that have flown in space, such as the NSTARthruster,themeasurementspresentedheremaybeappliedtopredictplumeplasma potentialsduringin-flightoperation. 48 CHAPTER 3: CHARACTERIZATION OF SURFACE CHARGING MEASUREMENTS Having quantified the experimental mesothermal plasma environment, it was next neces- sarytoestablishandquantifysurfacechargingdetectiontechniquesandsurfacechargingin the main plasma beam. This chapter details the techniques employed for both conducting anddielectricsurfaces,andpresentsfloatingpotentialresultsasafunctionofbothangleof attack and material variation in the main plasma beam. The results obtained establish the foundationforthesubsequentlunarchargingsimulatorfacility. 3.1 Surface Potential Charging Theory and Sheath Con- siderations An object immersed in plasma will continue charging until it reaches a steady state. The floatingpotentialisdeterminedbythecurrentbalancecondition: N ∑ k=1 I k (ϕ surf ) = 0 (3.1) where ϕ surf is the surface floating potential, and I k represents the various current sources tothesurface. Typicallunarcurrentsourcesincludetheambientplasmaelectronsandions (I e , I i ), electron and ion induced secondary electrons (I se , I si ), backscattered electrons I bse , and photoelectrons I ph . The complexity of the problem varies greatly depending on thecharacteristicsoftheambientplasma,materialpropertiesandgeometry. 49 In the absence of photoemission a plasma sheath is expected to form the transition region between the quasi-neutral ambient plasma and the surface, driving current collec- tion. Mobile electrons are expected to charge the surface negatively, leading to reduced electron density within the sheath. Typical plasma sheath potential and density profiles are given in Fig. 3.1 [66]. The electric field within the sheath points toward the surface, but is reduced exponentially with distance as the potential decreases. In the case of strong secondary emission, the surface will “self-limit” i.e. reach a positive potential such that secondary electrons are returned (usually on the order of the emitted electrons, if they are thedominantspecies). When photoemission dominates current collection is controlled by a photoelectron sheath that forms above the surface. It is a region of excess electrons that at steady state is bothfreeing andcapturing emittedphotoelectrons. Forapositivesurfacepotentialeithera monotonic or non-monotonic sheath profile can exist, as shown in Fig. 3.2 [49]. A mono- tonic sheath such as B will not reflect plasma electrons or capture emitted photoelectrons, butanon-monotonicsheathsuchasAcould. Figure3.1: Plasmasheathprofiles Figure3.2: Photoelectronsheathprofiles 50 3.2 SurfacePotentialDetectionTechniques When experimentally measuring the voltage distribution on a dielectric surface, such as a lunar simulant, any measurement that requires charge transfer could modify or destroy the actual data. Therefore, a non-contacting measurement method is often preferred. The via- bility of emissive probe (non-contacting) and embedded wires (contacting) surface poten- tial measurement methods, both of which are inexpensive to implement, was considered, before a commercial Trek non-contacting electrostatic voltmeter and probe was ultimately selected. For equipotential conducting surfaces a high impedance ( 10 MΩ) voltmeter wasemployed,whichprovidesaccuratefloatingpotentialmeasurements[31]. 3.2.1 EmissiveProbeSurfacePotentialMeasurements To establish an emissive probe surface potential detection baseline [66], a 100 mm x 100 mm flat aluminum (alloy 2024) plate was installed 10 inches downstream of beam cen- terline, at an angle of 0 degrees with respect to the beam direction. The conducting plate wasbiasedinincrementsof-10V,andtheemissiveprobe(25milexposedTungstenradius, Fig.3.5)wasusedtorecordplasmapotentialdataasafunctionofdistanceabovethebiased plate for both ambient and accelerated plasma cases. Resulting surface potentials are dis- played in Fig. 3.3. The Debye length 10 inches downstream of source exit was calculated tobe1mm,yieldingasheaththicknessontheorderofafewmm’sforthebiasedpoten- tials applied. Both test cases show that the emissive probe detects a potential gradient in thesheath,butdoesnotsufficientlyapproachthebiasedplateboundaryconditionpotential (the data point at 0 mm above the surface in the plots is a boundary condition and not an emissiveprobedatapoint). The experimental setup was next reconfigured to simultaneously detect floating con- ducting plate surface potentials with both a datalog unit (high impedance voltmeter) and theemissiveprobe,asshowninFigs.3.4and3.5. Usinganaccelerated10mAionbeamto 51 0 5 10 15 20 25 −60 −50 −40 −30 −20 −10 0 10 distance above surface, mm EP potential, V Sheath potentials over a biased plate, AMBIENT plasma −10 V biased plate −20 V biased plate −30 V biased plate −40 V biased plate −50 V biased plate (a) Ambientplasma 0 5 10 15 20 25 −60 −50 −40 −30 −20 −10 0 10 distance above surface, mm EP potential, V Sheath potentials over a biased plate, 10 mA ion beam −10 V biased plate −20 V biased plate −30 V biased plate −40 V biased plate −50 V biased plate (b) Acceleratedbeam Figure3.3: Plasmapotentialaboveabiasedaluminumplate charge the plates, the recorded surface potentials and emissive probe error (datalog toler- anceisontheorderofmV’s)aredisplayedinFig.3.6. Whiletheemissiveprobeerrorwith respecttothe“correct”datalogvaluesincreasesasafunctionofdistanceabovethesurface, these results do not conclusively prove that the emissive probe is capable of detecting the platefloatingpotential,asbothobjectsmaysimplybefloatingatsimilarvaluesindependent ofeachother. Figure3.4: Emissiveprobeanddatalogexperimentalsetup(nottoscale) 52 Figure3.5: Floatingconductingplatesgeometryandemissiveprobe −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −6 −4 −2 0 2 4 6 8 10 12 Floating potential as a function of x position, 10" downstream, 10 mA beam x position, inches floating potential, phi Plate potential, average EP potential, distance = 3.3825 mm EP potential, distance = 4.7116 mm EP potential, distance = 5.4833 mm EP potential, distance = 6.3254 mm EP potential, distance = 7.2371 mm EP potential, distance = 11.5606 mm (a) Floatingpotentials 0 2 4 6 8 10 12 0 0.5 1 1.5 2 2.5 3 Floating potential error vs. distance above surface, 10" downstream, 10 mA beam distance above surface, mm floating potential error, phi (b) Error Figure3.6: Conductingplatefloatingpotentialsandemissiveprobeerror 3.2.2 EmbeddedWireSurfacePotentialMeasurements In order to further test the emissive probe’s ability to accurately detect surface potentials in a flowing ion beam, a dielectric alumina silicate test bed of the same geometry was fabricated, as shown in Fig. 3.7 (a). Due to localized charging and charge separation, the dielectric material is expected to float at different potentials than the conducting plates. As potential is a function of position, i.e. any single point must be at the same potential, embedded wire tip potential in theory may yield the local dielectric surface potential. The embedded wire schematic is shown in Fig. 3.7 (a). Again using a 10 mA beam to charge the test bed, resulting surface potentials are displayed in Fig. 3.7 (b). These results show 53 (a) Schematic −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −6 −4 −2 0 2 4 6 8 10 12 x location, inches potential, V Embedded wires and EP potential, 1 mm above surface, 10 mA beam Embedded wires Emissive probe (b) Surfacepotentials Figure3.7: Embeddedwiresschematic,geometry,andchargingresults that the emissive probe and embedded wires are not precisely detecting dielectric surface potential with respect to one another. Therefore, one of the methods must be invalid and bothmaybe. 3.2.3 Non-contactingTrekProbeSurfacePotentialMeasurements To conclusively test the accuracy of the embedded wires and hence emissive probe as sur- face potential detection techniques, a commercial Trek non-contacting electrostatic volt- meter(ESVM)andprobewaspurchased. AnESVMisavibratingcapacitiveprobeusedto detect surfacepotential [39]ata givenlocationwith acurrentnulling method[91,55, 51]. The method is illustrated in Fig. 3.8 (from [13]). A capacitor is created between the sam- pling surface with potential ϕ s and the vibrating electrode, biased to V. The upper elec- trode is vibrating with frequency ! and amplitude d 1 , forming a time dependent gap, d = d 0 +d 1 sin(!t) with d 1 < d 0 , between the sampling dielectric surface and the vibrating electrode. When the electrode potential V is biased to V = ϕ s there is zero current flow through the electrode. This establishes the surface potential measurement. In addition to not disturbing the physical surface of the dust layer, this method also eliminates typical 54 errorsassociatedwithcontactingmethodsformeasuringthesurfacepotentialandtheFara- day cup method for measuring single dust charge. The model purchased and employed Actual probe above surface Figure3.8: TrekESVMnon-contactingsurfacepotentialmeasurementmethod was a Trek 323-L-CE ESVM and model 6000B-8 side-view probe with a voltage range of 100V,anaccuracyof50mV,andaresponsetimeof300ms,providedthattheprobe spacing-to-surfaceis1to3mm. After calibrating the Trek probe with a conducting plate of known bias, the alumina silicate dielectric plates were again placed in the main plasma beam at0 degree angle of attack, and measurements were recorded for both the embedded wires and the non- contacting Trek probe. The experimental setup is given in Fig. 3.9. Note that the exposed vibrating electrode of the Trek probe must be sealed from plasma flux during source oper- ation to prevent charge build-up in the probe itself. Also, the plasma source must be shut downbeforetheTrekprobescansthesurface,ortheplasmacurrentwillsubstantiallyinter- fere with the measurement. Provided that the Trek probe scan is completed within a few minutes of source shutdown, surface charge decay is minimal (the charge decay effect is discussed in detail in Chapter 5). Surface potential results for both stationary plasma and accelerated beam cases are given in Fig. 3.10. Given that Trek probe accuracy is on the order of 50 mV, the embedded wire technique is drastically inaccurate. This inaccuracy is furtherevidencedbythelackofexpectedfloatingpotentialvariationintheembeddedwire measurementsasplasmacurrentsareadjusted. 55 x= -2” x= -1” x= 0” x= 1” x= 2” Embedded wires Trek probe Figure3.9: Floatingdielectricplates,embeddedwires,andTrekprobeexperimentalsetup −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −20 −15 −10 −5 0 Trek probe dielectric floating potential x position Floating potential, V Beam = 2 mA Beam = 6 mA Beam = 10 mA Stationary = 6 mA (a) Trekprobe,beamcurrentvariation −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −20 −15 −10 −5 0 Embedded wires floating potential x position Floating potential, V Beam = 2 mA Beam = 6 mA Beam = 10 mA Stationary = 6 mA (b) Embeddedwires,beamcurrentvariation −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −20 −15 −10 −5 0 Trek probe dielectric floating potential x position Floating potential, V Ii = 10 mA, Ie = 20 mA Ii = 6 mA, Ie = 20 mA Ii = 2 mA, Ie = 20 mA Ii = 2 mA, Ie = 30 mA (c) Trekprobe,electroncurrentvariation −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −20 −15 −10 −5 0 Embedded wire floating potential x position Floating potential, V Ii = 10 mA, Ie = 20 mA Ii = 6 mA, Ie = 20 mA Ii = 2 mA, Ie = 20 mA Ii = 2 mA, Ie = 30 mA (d) Embeddedwires,electroncurrentvariation Figure 3.10: Trek probe measured dielectric floating potentials (left); Embedded wires measureddielectricfloatingpotentials(right) 56 3.2.4 DetectionTechniqueConclusions Non-contacting surface potential probes are typically perturbed by local plasma, but by employinganemissiveprobetheplasmasheathenvironmentisusedtoapproximatesurface potentials. However,probevaliditybreaksdownasdensitydecreasesinthesheath[36]and resultspresentedhereshowthatemissiveprobesmaybeusedtoestimatesheathpotentials but not floating surface potentials. Through the purchase and use of a commercial Trek ESVM and probe, it is also found that the embedded wire method presented here does not accurately measure floating surface potentials. Therefore, the Trek ESVM and probe is employedforallsubsequentdielectricsurfacepotentialmeasurements. 3.3 ConductingPlateAngleofAttackVariation In order to assess the influence of experimental beam angle of attack on surface floating potential and to validate the mesothermal plasma source charging capability, a 50.8 mm x 50.8 mm flat aluminum conducting plate was placed 10 inches downstream of source exit alongbeamcenterline, asshowninFig. 3.11. Onlythefrontfaceoftheplatewasexposed to plasma flow (the sides and back face were covered with Kapton tape). Angle of attack was varied from -90 to 90 in 1.8 increments, and the plate floating potential was read withahighimpedancevoltmeter. Theexperimentwasconductedforbeamcurrentsettings of 6 mA and 10 mA, respectively. The resulting data is displayed in Fig. 3.12, and it is clear that electron flux dominates for negative angles of attack. The floating potential then increases through 0 degrees to 20 degrees before plateauing for higher angles. The flattening of the potential is due to ion flux sin dependency (Eqn. 3.3) and ion-induced secondaryemissionfromthealuminumplateduetoincreasedhighenergyionimpingement athighanglesofattack. Fluxsourcesanalyticallyconsideredwerehighenergy(1100eV)ionflow,lowenergy (2 - 7 eV) thermal electron flux, and ion-induced secondary electron emission. The flat 57 Figure3.11: Flatplatechargingexperimentalsetup(nottoscale) plate charging deductions are confirmed by the analytical probe theory charging model displayedinFig.3.13andEqns.3.2-3.9 e = n e 2 [ v te p exp ( ( v de v te ) 2 ) + v de ( 1+erf ( v de v te )) ] (3.2) where e iselectronfluxandv de iselectrondriftvelocity(0forthermalelectrons), i = n i v i ;sin() (3.3) where i isionflux, si = si n i v i (3.4) where si ision-inducedsecondaryelectronfluxand si =.1istheyieldof1100eVargon ionsonaluminum[10], crit = sin 1 ( e i + si ) (3.5) < crit : e exp ( eϕ surf kT esec ) i sin() si sin() = 0 (3.6) ϕ surf = T e ln ( i sin()+ si sin() e ) (3.7) 58 > crit : e ( i sin()+ si sin())exp ( eϕ surf kT esec ) = 0 (3.8) ϕ surf =T esec ln ( e i sin()+ si sin() ) (3.9) and where crit is the angle of attack at whichϕ surf = 0 andT esec = 3 eV is the secondary electron temperature (from on-going Marshall Space Flight Center charging research). Sheath thickness is on the order of d 6 mm. Plasma parameters for the analytical equations presented were recorded during beam characterization. Results show that the experimental surface potential is generally higher than what the analytical solutions pre- dict. This effect is due to either the CEX ion population or the high speed neutral popu- lation (created by CEX and not detected by plasma probes), or both. In order to quantify experimental surface charging results it is therefore necessary to ensure separation of the twoionpopulationsandreduceneutralimpingementinanygiventestregion. −100 −50 0 50 100 −15 −10 −5 0 5 10 15 20 Angle of attack, deg Plate floating potential, V Plate floating potential vs. angle of attack 6 mA 6 mA 10 mA 10 mA ANALY 6 mA ANALY 10 mA Figure3.12: Flatplatefloatingpotentialasafunctionofangleofattack 59 Φ surf α = 90° α = 0° α < α crit α > α crit α = α crit Figure3.13: Analyticalsurfacechargingmodelasafunctionofalpha 3.4 ConductingPlateMaterialVariation High energy main beam ions are confined to the axial plume as previously defined by the divergence angle div . CEX ions are primarily born near source exit and flow outwards bothaxiallyandradially. Therefore,lowenergyCEXionsandlowenergymainbeamions that expand radially outward at the ion acoustic velocity are the dominant ion species at 0 degreeangleofattackinthemainbeamandatlowanglesofattackoutsideoftheplume. Forlowenergy(0-1keV)Ar + bombardmentofmetals,“PotentialEjection”dominates theion-inducedsecondaryelectronemissionprocess[10]. ThisemissionistypicallyAuger or resonance. For high energy (>1 keV) Ar + bombardment of metals, “Kinetic Ejection” dominates[10]. Tostudy“PotentialEjection”theCEXionpopulationwasthenselectedas theenergyrangeis 5-30eV. The experimental setup in Fig. 3.14 was chosen to quantify conducting plate charging: First,withonlyCEXionflux(0degreeinmainbeam)andsecond,withneithersignificant mainbeamnorCEXionflux(90degreeoutsidemainbeam). Threeconductingmaterials were used: stainless steel 304, aluminum alloy 2024, and copper alloy 110 , as displayed in Fig. 3.15. Each sample is 1 ′′ x 1 ′′ and 20 mil thick, with 0.53 ′′ x 0.71 ′′ exposed area. Thetestbedcontainingeachsamplewastraversedsuchthatsampleswereheldatthesame locationsalongthey-axis(-1.5 ′′ y,0 ′′ y,+1.5 ′′ y),andthefloatingpotentialswererecorded 60 by a Datalog unit. The plasma beam setting was the nominal: 1100 eV beam energy, 10 mA beam current, 10 mA neutralizer current, and 4 sccm argon flow rate. Results are given in Fig. 3.16. To ensure steady state measurements, five data points were recorded for each sample material at each location. The left column shows that when exposed to low energy (CEX) ion flux and low energy electron flux, different conducting materials of the same size and thickness will float to different potentials. The right column shows that when exposed to primarily low energy electron flux (mobile electrons are the dominant flux source at 90 outside of the main beam), the floating potential solution for different conductingmaterialswillconverge. Figure3.14: Conductingplatematerialvariationexperimentalsetup(nottoscale) SS Al Cu +y -y Figure3.15: Conductingplatematerials: Stainlesssteel(SS),aluminum(Al),copper(Cu) 61 0 1 2 3 4 5 6 −5 0 5 10 15 Data point Conducting plate floating potential, V Conducting plate floating phis, 10 mA beam, 0 deg alpha, −1.5" y SS Al (a) SSandAl,CEXionflux,-1.5 ′′ y 0 1 2 3 4 5 6 −5 0 5 10 15 Data point Conducting plate floating potential, V Conducting plate floating phis, 10 mA beam, 90 deg alpha, −1.5" y SS Al (b) SSandAl,limitedflux,-1.5 ′′ y 0 1 2 3 4 5 6 −5 0 5 10 15 Data point Conducting plate floating potential, V Conducting plate floating phis, 10 mA beam, 0 deg alpha, 0" y SS Al Cu (c) SSandAlandCu,CEXionflux,0 ′′ y 0 1 2 3 4 5 6 −5 0 5 10 15 Data point Conducting plate floating potential, V Conducting plate floating phis, 10 mA beam, 90 deg alpha, 0" y SS Al Cu (d) SSandAlandCu,limitedflux,0 ′′ y 0 1 2 3 4 5 6 −5 0 5 10 15 Data point Conducting plate floating potential, V Conducting plate floating phis, 10 mA beam, 0 deg alpha, 1.5" y Al Cu (e) AlandCu,CEXionflux,1.5 ′′ y 0 1 2 3 4 5 6 −5 0 5 10 15 Data point Conducting plate floating potential, V Conducting plate floating phis, 10 mA beam, 90 deg alpha, 1.5" y Al Cu (f) AlandCu,limitedflux,1.5 ′′ y Figure 3.16: Conducting plate floating potentials at 0 in main beam (left); Conducting platefloatingpotentialsat90 outsidemainbeam(right); 62 This demonstrates the Auger effect, which is well defined for conductors: when the filling of an inner shell electron vacancy of an atom (say for example when an inner shell electron neutralizes an incident ion) is accompanied by emission of an electron from the sameatom(whenanelectronfromahigherenergylevel“falls”intotheinnershellvacancy energy is released and transferred to an outer electron which can escape). Auger emission yield Auger is a function of atomic number, work function, and inner shell transition, as displayed in Fig. 3.17. Under constant low-energy ion flux, higher Auger predicts higher floatingsurfacepotentials. Thistrendisconfirmedbytheexperimentaldata,inwhichϕ surf Al>ϕ surf SS>ϕ surf CuwhenthematerialsareunderCEXionflux. Additionally,Auger emissionpartiallyexplainsthedeviationbetweentheexperimentalandanalyticalsolutions inFig.3.12. Low-energyCEXionfluxandassociatedAugeremission(whichtheanalytical solutiondoesnotaccountfor)drivestheexperimentalsolutiontohigherfloatingpotentials for0 . Metal Z Work func0on (eV) γ Auger (K-‐shell) Φ surf under ion flux Al 13 4.06 – 4.26 .95 13 V SS ~26 ~4.4 .65 10 V Cu 29 4.53 – 5.10 .5 8 V Figure3.17: ConductingmaterialAugeremissionproperties 3.5 DielectricMaterialProperties The floating surface potential charging process is much different for dielectrics than for conductors. Asafunction ofdielectricconstant(dielectricconstantis zerofrequencyrela- tivedielectricpermittivity),dielectricshavetheabilitytokeepelectricalchargesphysically separated by a distance. This property is controlled by material bulk density. Dielectrics 63 havelowelectricalconductivity(quantifieshoweasilyamaterialallowstheflowofelectri- cal current) and low loss tangents (inherent dissipation of electromagnetic energy). These twopropertiesarebothmaterialandtemperaturedependent. As a result of dielectric properties, charge accumulates at the surface of dielectrics. Differentialchargingwilloccurovertheentiresurfaceasafunctionofincidentflux,mate- rial properties, and the nature of the surface itself. Surface properties (solid, regolith, dust, porosity, roughness, etc.) influence the charge state of local patches and individ- ual particles, and hence the surface potential. To experimentally introduce these effects, the following dielectrics were selected: alumina silicate, Mykroy/Mycalex R ⃝ glass mica, andJSC-1A(alunarregolithsimulant). Availabledielectricpropertyvaluesforthemateri- als[45,41,9,29]aregiveninFig.3.18,andcompositionsanddescriptionsaregivenbelow. Each material possesses relatively high dielectric strength and low electrical conductivity andlowdielectriclosses. Thesepropertiesareverysimilartothoseofactuallunarsoil,and theypermitthematerialstobothreadilychargeandremainelectrostaticallycharged. Property( Mykroy/Mycalex®( Alumina(silicate( JSC:1A*( Lunar(Soil( Dielectric(constant( 6.7(–(6.9( 5.6(–(6.8( 3.6(–(4.22( 2(7(11( Electrical(conduc<vity((S/m)( 7( 7( 7( 10 713( –(10 710( Loss(tangent( 7( 7( .11(7(.29( 0.001(7(.1(( *at 23 deg C Figure3.18: Dielectricmaterialproperties 64 3.5.1 AluminaSilicate Alumina silicate is a common name for chemical compounds primarily derived from alu- minum oxide (Al 2 O 3 ) and silicon dioxide (SiO 2 ). The compounds may occur naturally (andalusite, kyanite, sillimanite) or synthetically (alumina silicate ceramics), and may be anhydrousorhydrated[56]. Analuminasilicateceramicwasselectedduetoitshighoxide content(similartothelunarsoil)anditsabilitytobeeasilycrushedandsieved. Thecrush- ing and sieving process was performed in the laboratory and is shown in Fig 3.19. Alu- mina silicate dust of 100 m 40 m was obtained. The chemical composition (from McMaster-Carr data sheet, approximate) of the alumina silicate used in these experiments isgiven. 59%SiO 2 +29.2%Al 2 O 3 +3%Fe 2 O 3 +1.4%TiO 2 +1.2%K 2 O 3.5.2 Mykroy/Mycalex R ⃝ Mykroy/Mycalex R ⃝ is a molded glass mica composite with high dielectric strength. It is composed of finely powdered glass and mica. It was selected to compare charging results of a second solid dielectric relative to solid alumina silicate. Approximate values of mica (muscovite)chemicalcomposition[56]aregiven. 45.3%SiO 2 +38.4%Al 2 O 3 +11.8%K 2 O 3.5.3 JSC-1A JSC-1A bulk lunar mare regolith simulant was created to match the composition of the previous JSC-1 simulant as closely as possible, in order to support NASA’s future lunar surface exploration. Like the original JSC-1 material, JSC-1A was created from glass- rich (50 vol%) basaltic tuff mined from a commercial quarry at Merriam Crater, which is a volcanic cinder cone near Flagstaff, AZ [30]. Typical lunar soils have mean grain 65 (a) Solidmaterialbrokenwithhammer (b) Segmentsfurthercrushed (c) Sieves (d) 100m40mdust Figure3.19: Crushingandsievingaluminasilicate sizes between 45 and 100 microns, and analysis of JSC-1A samples show that its particle distribution falls within one standard deviation of these values [90]. Composition has beenestimatedas 46.2% SiO 2 , 17.1% Al 2 O 3 , 11.2% FeO, 9.43% CaO, 6.87% MgO, 3.33% Na 2 O, 1.85%TiO 2 andothertraceoxides,whichiscomparabletoApollo17soilsamples[30]. Whilealllunar simulantscurrentlydonotreproducemanyofthespecialcharacteristicsoflunarsoil,such 66 asthepresenceofagglutinatesandnp-Fe 0 (nanometer-scalemetallicFe)particles,JSC-1A doesincludealargeproportionofglassshardsandangulargrains. Thelackofagglutinates and metallic Fe is thought to only be a drawback when simulating the heating or melting oflunarsoil,orwhenmagneticpropertiesarestudied. 3.6 Main Beam Low Angle of Attack Dielectric Surface ChargingResults Chargingmeasurementsofthethreeselecteddielectricmaterialsinthemainplasmabeam at 0 were obtained via the experimental setup shown in Fig. 3.20. Each test bed as viewed from the top is 2 ′′ x 5 ′′ . Each test bed as viewed from the side is 1 8 ′′ thick and backed by a 20 mil thick aluminum grounded plate. One test bed is placed on the +y side ofthezaxis,andonetestbedisplacedonthe-ysideofthezaxis,foreachexperimentaltest case. Test beds were held under vacuum and baked with a Watlow polyimide sheet heater to outgas moisture. This 5 ′′ x 5 ′′ heater is shown in Fig. 3.21 and was located directly beneath and in contact with the grounded 20 mil thick aluminum plate. The plasma beam is operated at the nominal setting (1100 eV beam energy, 10 mA beam current, 10 mA neutralizercurrent,4sccmargonflowrate)foratotaltimeof10minutestoensurethatthe dielectric surface potential reaches steady state (confirmed by testing). During this time period the Trek probe is covered to prevent charge buildup in the probe itself. Following source shutdown, the Trek probe scans 1D surface potential of both test beds along the y axis. Mainbeam0 chargingresultsanddiscussionforaconductingaluminumtestcase andforeachsetofdielectrictestbedsisprovidedbelow. 67 Figure3.20: Mainbeamdielectricmaterialchargingexperimentalsetup Figure3.21: 5 ′′ x5 ′′ Watlowpolyimideheaterformoistureoutgassing 3.6.1 Aluminumvs. Aluminum First, two aluminum strips of equal dimensions and exposure to plasma were installed at the Trek probe scan locations on either side of the z-axis (Fig. 3.20), in order to quantify beamchargingsymmetry. Beamcurrentwassetto2,6,and10mA.Allotherbeamsettings werenominal. ThealuminumstripsandfloatingpotentialresultsaredisplayedinFig.3.22. As beam current increases, aluminum plate floating potential increases. This is additional confirmation of Auger emission. Also, floating potentials for -Y are slightly more positive thanfloatingpotentialsfor+Y,underidenticalbeamcurrentflux. Thisdemonstratesthatthe 68 plasmabeamisslightlystrongerin-Y.IndependentFaradayprobeandRPAmeasurements haveshownthe-Yfluxtobe25%>+Yflux. −60 −40 −20 0 20 40 60 −5 0 5 10 15 Y, mm Al strip floating potential, V Measured Al strip potentials, varied beam current −y plate, 2mA +y plate, 2mA −y plate, 6mA +y plate, 6mA −y plate, 10mA +y plate, 10mA -Y +Y Figure3.22: Aluminumconductingstripssetupandmainbeamchargingresults 3.6.2 AluminaSilicatevs. Mykroy/Mycalex R ⃝ Secondly, solid alumina silicate and solid Mykroy/Mycalex R ⃝ were both charged with 10 mA beam current on either side of the z axis, as shown in Fig. 3.23 (a) and (b). This was donetodemonstratethechargingnatureofasoliddielectricat0 inthemainbeam,com- paredtothepreviousconductingaluminumstripresults. Trekprobescanswereperformed 1 minute and 6 minutes after source shutdown, respectively. Floating surface potential results are given in Fig. 3.23 (c) and (d). The floating potentials are over an order of mag- nitude higher than those recorded for the aluminum strips. Clearly, charging in the main 69 plasma beam varies considerably between dielectrics and conductors. Further, the dielec- tric materials charge to surface potentials that vary 20% with respect to one another. Thisdemonstratestheeffectofsoliddielectricmaterialpropertiesonthesurfacepotential. Solid alumina silicate potentials were less than solid Mykroy/Mycalex R ⃝ potentials for bothalignments,showingthatmaterialpropertiescontributedmoretothefloatingpotential resultsthanbeamasymmetry. Limitedchargedecayoverthe5minutetimeperiodbetween scans demonstrates the ability of the dielectrics to maintain surface charge separation and limitelectricalcurrentflow. -Y +Y (a) Aluminasilicate. Mykroy/Mycalex R ⃝ . -Y +Y (b) Mykroy/Mycalex R ⃝ . Aluminasilicate. −60 −40 −20 0 20 40 60 −20 0 20 40 60 80 100 120 140 160 180 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (c) Alignment(a)results −60 −40 −20 0 20 40 60 −20 0 20 40 60 80 100 120 140 160 180 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (d) Alignment(b)results Figure3.23: AluminasilicateandMykroy/Mycalex R ⃝ mainbeamcharging 3.6.3 AluminaSilicatevs. JSC-1A Thirdly, solid alumina silicate and JSC-1A particles were both charged with 10 mA beam current on either side of the z axis, as shown in Fig. 3.24 (a) and (b). This was done to demonstrate charging potential differences between a solid dielectric and a particulate 70 lunar simulant. Trek probe scans were performed 1 minute and 6 minutes after source shutdown, respectively. Floating surface potentials are given in Fig. 3.24 (c) and (d). For bothalignments,JSC-1Apotentialsare35%lowerthansolidaluminasilicatepotentials. Thiseffectisduetoeitherorbothmaterialpropertiesdifferencesanddustvs. solidcharge depositiondifferences. Notethatasurgeinbeamcurrentto15mApriortosourceshutdown contributed to the heightened potentials in Fig. 3.24 (d). Charge decay occurred nearly twiceasquicklyforJSC-1Athanforsolidaluminasilicate. Interestingly,chargedecayfor solid alumina silicate occurred more quickly in these two data cases than when charged nexttoMykroy/Mycalex R ⃝ (Fig.3.23). -Y +Y (a) Aluminasilicate. JSC-1A. -Y +Y (b) JSC-1A. Aluminasilicate. −60 −40 −20 0 20 40 60 −20 0 20 40 60 80 100 120 140 160 180 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (c) Alignment(a)results −60 −40 −20 0 20 40 60 −20 0 20 40 60 80 100 120 140 160 180 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (d) Alignment(b)results Figure3.24: AluminasilicateandJSC-1Amainbeamcharging 3.6.4 AluminaSilicateSolidvs. AluminaSilicateDust Finally, solid alumina silicate and alumina silicate dust were both charged with 10 mA beam current on either side of the z axis, as shown in Fig. 3.25 (a) and (b). This was 71 done to demonstrate charging potential differences between a solid and dust material of thesamechemicalcomposition. Trekprobescanswereperformed1minuteand6minutes after source shutdown, respectively. Floating potential results are given in Fig. 3.25 (c) and (d). It appears as though the stronger beam in -Y drives the surface potential higher when the test bed is either alumina silicate solid or alumina silicate dust. However, this trendisinconclusive,astheplasmafluxconsistsofhighenergyions,lowenergyions,and low energy electrons. Charge deposition processes of each population may independently contributetoandgoverntherecordedsteadystatesurfacepotentials. -Y +Y (a) Aluminasilicatesolid. Aluminasilicatedust. -Y +Y (b) Aluminasilicatedust. Aluminasilicatesolid. −60 −40 −20 0 20 40 60 −20 0 20 40 60 80 100 120 140 160 180 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time Strong beam Weak beam (c) Alignment(a)results −60 −40 −20 0 20 40 60 −20 0 20 40 60 80 100 120 140 160 180 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time Strong beam Weak beam (d) Alignment(b)results Figure3.25: Aluminasilicatesolidandaluminasilicatedustmainbeamcharging 3.7 SummaryandConclusions Emissive probes, embedded wires, and a commercial Trek ESVM and probe have been considered as dielectric surface potential detection techniques. Only the Trek probe has proven to be viable, provided that charge-build up in the probe itself is mitigated, and 72 surface potential scans are performed after plasma shutdown. Experimental conducting plate floating potential measurements have been recorded as a function of plasma beam angle of attack. An analytical model and additional experiments in which three different conducting materials were charged under identical plasma conditions have demonstrated theAugereffect. Solid alumina silicate, solid Mykroy/Mycalex R ⃝ , JSC-1A lunar simulant, and alumina silicate dust have been charged under 10 mA beam flux at a low (0 ) angle of attack in the main plasma beam. Results demonstrate the significance of both electrical properties and surface properties on dielectric surface floating potentials and surface charge decay. However, well-defined emission (electron backscattering, ion and electron induced sec- ondary electron emission) coefficients and charge deposition models do not exist for these materials,makingitdifficulttoanalyticallycalculatethesurfacechargingpotentials. Addi- tionally,highenergyionflux,highspeedneutralflux,lowenergyionflux,andlowenergy electronfluxareallpresentintheexperimentalmainplasmabeam,makingitverydifficult to establish individual particle flux effects on the various dielectric surfaces. That being said, a theory to explain the high (+100 V with respect to plasma potential) dielectric surfacepotentialsrecordedinthemainbeamisgiveninChapter6(followingasystematic studyofdielectricsurfacecharginginChapter5). 73 CHAPTER 4: EXPERIMENTAL AND NUMERICAL INVESTIGATIONS OF THE NEAR SURFACE LUNAR PLASMA FIELD Dielectric charging in the experimental main plasma beam does not well represent lunar surface plasma charging, due to combined high energy ion flux, low energy ion flux, low energy electron flux, and limited plasma wake expansion. In this chapter a “Lunar Sim- ulator Facility” is developed and characterized both experimentally and numerically for mesothermal plasma flow a low angle of attack. The facility has the ability to control low energy CEX ion flux to the test bed region, to control electron kinetic energy in the test bed region, and to create an expansion region for high energy ions. These controls make it possible to carry out a systematic study of dielectric surface charging in the subsequent chapter. 4.1 PlasmaDensityandLengthScaling Simulatinglunarsurfacecharginginavacuumchamberrequirescarefulscaling. Theaver- agesolarwinddensityisontheorderof4particles/cm 3 ,whichwithaplasmatemperature of20eVwouldresultinaDebyelengthof16m. Asthevacuumchamberusedisabout1 mindiameter,theexperimentalDebyelengthmustbeontheorderofcmormm,meaning that the plasma density must be at least 5 to 6 orders of magnitude greater than that of the solar wind. Density scaling will result in significantly increased experimental ion and electron current flux to any incident surface as compared to lunar conditions, as shown in 74 Fig. 4.1. Photoemission current flux J ph cannot be scaled so readily, due to the relatively fixed nature of photoelectron yield. Therefore, the lunar simulator facility and surface chargingtestbedshereinareinvestigatedfromaplasmaonlyperspective,whichsimulates wake and shadowed regions downstream of lunar topography or surface objects when the sunangleislow. dmoon = 740 √ 20eV 4cm 3 = 16:54m (4.1) dchamber = 740 √ 4eV 1e7cm 3 = 4:68mm (4.2) dchamber L chamber = 4:68mm 101:6mm = 0:046 (4.3) L Moon = dMoon 0:046 = 360m (4.4) Lunar dimensions are scaled according to the vacuum chamber Debye length-to- characteristic length ratio dchamber /L chamber per Eqn. 4.3, where dchamber = 4.68 mm per Eqn. 4.2 at the front face of the simulated lunar object and L chamber = 101.6 mm is the size of the simulated lunar object. Using that ratio and the calculated lunar dMoon per Eqn 4.1 results in a lunar characteristic length L Moon of360 m per Eqn. 4.4, which is equivalent to a small lunar hill or crater (Shackleton crater near the lunar south pole has a radius of10,500 m). Therefore, the experimental characteristic length ratio is scaled to lunarobjectsontheorderof100’sofmeters. 4.2 LunarSimulatorFacilityDevelopment Anexperimentalsetuphasbeenestablishedinthevacuumchambertosimulatesolarwind plasma flow over the lunar surface at low angles of attack such as at the lunar terminator and polar regions. This section presents the design and testing process used to select a simulatedlunarobjectanditslocationandorientationwithrespecttothemainbeamplasma 75 Figure4.1: Experimentalplasmadensityandcharacteristiclengthscaling flow. Also,theregionsinwhichhighenergyandlowenergyionsarerespectivelydominant areidentified. 4.2.1 SimulatedLunarObjectSelectionandDimensions Design and development of the lunar simulator facility was iterative. The first step was to determine the appropriate size and location with respect to the main plasma beam of the simulated lunar object. Two aluminum cubes were tested, one 2 ′′ x 2 ′′ x 2 ′′ and one 4 ′′ x 4 ′′ x 4 ′′ . These sizes were chosen for testing because they are on the order of the main beam plasma plume radius. Only one cube at a time was installed in the vacuum chamber. An alumina silicate dielectric plate was located immediately downstream of the cube location. The initial experimental setup is given in Fig. 4.2 and the two cubes are displayedinFig.4.3. The emissive probe, axial Faraday probe, and Langmuir probe were employed in the scan area plane, which was located at y=0 (the center of each cube, respectively). The top of the ion source exit grid was aligned with the top of the cube in each case, such that the plasma beam flowed directly into the front face and directly over the top face of each cube. Each cube was fixed to a small positive potential 3 to 5 V on the order of the electron temperature to simulate positive floating potential on the front face. In each case 76 .915 m 1.05 m .040 m .230 m 0 V Neutralizer -15.5 V 0 V Scan area .1905 m .2794 m .1524 m .330 m .0254 m .2032 m X Z Cube Dielectric Figure4.2: Initiallunarsimulatorfacilityexperimentalsetup 2” x 2” x 2” 4” x 4” x 4” Figure4.3: Cubessimulatinglunarobjectsinlunarsimulatorfacilitydevelopment the plasma beam setting was nominal (1100 eV beam energy, 10 mA beam current, 10 mA neutralizer current, 4 sccm argon flow rate). Resulting scan area 2D contour plots for both cube test cases of plasma potential ϕ p , ion density n i , electron density n e , and space charge n i - n e are given in Figs. 4.4 - 4.7. It is clear that main beam plasma expansion is occurring over the top face of the cube in each case. However, the experimental region of interest behind and below each cube does not show clear expansion. This region does shownegativespacecharge,indicatingthatelectronsmaybedominantinthis“shadowed” region. But, the Faraday probe measurements taken were axial only, and high and low energy ion populations were not distinguished from each other. It is therefore necessary 77 to further characterize and control plasma flow over the simulated lunar object, which was chosen to be the 4 ′′ cube as it provides a larger “shadowed” scan and test region than the 2 ′′ cube,whilestillremainingimmersedinthemainbeamplasmaplume. 4.2.2 IonSpeciesCharacterizationbyRPA Inordertomodifytheinitiallunarsimulatorfacilityexperimentalsetuptobetterrepresent solar wind flow at a low angle of attack, it was first necessary to identify the dominant ion species in both the axial flow direction above the block and the radial flow direction behind the block. To do so the RPA was located in the main beam and “shadowed region” as shown in Fig. 4.8. Note that measurements were taken in the y=0 (source and block y centerline)x-zplane. Highvoltage(0to1500V)andlowvoltage(-20to20V)scanswereperformedineach location under nominal plasma beam flux. The high voltage main beam scan results are given in Fig. 4.9. The RPA collector current given is the sum of the aluminum foil wall and back plate currents, and the two peaks offer additional evidence of ion focusing as a function of ion retarding grid voltage (discussed previously under RPA diagnostics). The resultingionenergydistributioniscenteredaround1100V(beamenergy)whichconfirms that axially flowing high energy ions are dominant in the region above the cube. The low voltage main beam scan was overwhelmed by high energy ion flux, such that the main beamCEXpopulationcouldnotbedistinguished. A 2D radial Faraday probe scan of the shadowed region confirmed the presence of radially inflowing ions. The resulting contour plot is given in Fig. 4.11 (a), and the peak ion current flux is 3 orders of magnitude lower than that in the main beam. The RPA high voltage shadowed region scan did not detect high energy ions, indicating that only CEX ions were radially inflowing. The RPA low voltage shadowed region scan results are shown in Fig. 4.11 (b), where ion retarding grid voltages greater than 5 V reduce the 78 X,mm Z,mm 350 400 450 50 100 150 phiV 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 2” x 2” x 2” Alumina silicate 1” 1” Cube +5V (a) 2 ′′ cube X,mm Z,mm 300 350 400 450 500 -50 0 50 100 phi V 5.0 4.0 3.0 2.0 1.0 0.0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 -7.0 -8.0 4” x 4” x 4” 1” 1.5” Alumina silicate Cube +3V (b) 4 ′′ cube Figure4.4: Initiallunarsimulatorfacilityplasmapotentialϕ p 79 X,mm Z,mm 350 400 450 50 100 150 niaxim -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 Alumina silicate 2” x 2” x 2” 1” 1” Cube +5V (a) 2 ′′ cube X,mm Z,mm 300 350 400 450 500 -50 0 50 100 niaxim -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 1” Alumina silicate 4” x 4” x 4” 1” 1.5” Cube +3V (b) 4 ′′ cube Figure4.5: Initiallunarsimulatorfacilityaxialiondensityn i 80 X,mm Z,mm 350 400 450 50 100 150 nem -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 Alumina silicate 2” x 2” x 2” 1” 1” Cube +5V (a) 2 ′′ cube X,mm Z,mm 300 350 400 450 500 -50 0 50 100 nem -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 Alumina silicate 4” x 4” x 4” 1” 1.5” Cube +3V (b) 4 ′′ cube Figure4.6: Initiallunarsimulatorfacilitytotalelectrondensityn e 81 X,mm Z,mm 350 400 450 50 100 150 niaxi-nem -3 1.00E+13 9.33E+12 8.67E+12 8.00E+12 7.33E+12 6.67E+12 6.00E+12 5.33E+12 4.67E+12 4.00E+12 3.33E+12 2.67E+12 2.00E+12 1.33E+12 6.67E+11 0.00E+00 -6.67E+11 -1.33E+12 -2.00E+12 Alumina silicate 2” x 2” x 2” 1” 1” Cube +5V (a) 2 ′′ cube X,mm Z,mm 300 350 400 450 500 -50 0 50 100 niaxi-nem -3 1.00E+13 9.33E+12 8.67E+12 8.00E+12 7.33E+12 6.67E+12 6.00E+12 5.33E+12 4.67E+12 4.00E+12 3.33E+12 2.67E+12 2.00E+12 1.33E+12 6.67E+11 0.00E+00 -6.67E+11 -1.33E+12 -2.00E+12 Alumina silicate 4” x 4” x 4” 1” 1.5” Cube +3V (b) 4 ′′ cube Figure4.7: Initiallunarsimulatorfacilityspacechargen i -n e 82 4” x 4” x 4” Alumina silicate Cube +3V RPA Main beam axial flow “Shadowed region” radial flow RPA 0.5” 2” X Z Figure4.8: RPAlocationswithrespecttosimulatedlunarobject 0 200 400 600 800 1000 1200 1400 1600 0 0.5 1 1.5 2 2.5 3 3.5 4 x 10 −7 RPA current vs. Ion Retarding Voltage Voltage, V Grid current, A Wall + Back plate (a) Collectorcurrent 0 200 400 600 800 1000 1200 1400 1600 −1 0 1 2 3 4 5 6 7 8 9 x 10 −10 Ion energy distribution eV dI/dV Back plate dI/dV (b) Ionenergydistribution Figure4.9: RPAmainbeamaxialflowhighvoltagescan 260 280 300 320 340 360 380 400 −2 0 2 4 6 8 10 x 10 −4 X, mm FP radial Ji, A/m 2 Measured FP radial Ji 2.6 sccm 4.1 sccm 10.1 sccm (a) J i radial 260 280 300 320 340 360 380 400 10 9 10 10 10 11 10 12 X, mm LP ne, m−3 Measured LP electron density 2.6 sccm 4.1 sccm 10.1 sccm (b) n e Figure4.10: 1DJ i radialandtotaln e 0.75 ′′ abovedielectricsurface 83 collector current to the noise threshold of the measurement circuit, confirming that low energy CEX ions comprise the radially inflowing ion population to the dielectric surface. Additionally, 1D measurements in the x direction of the radial ion current flux and total electron density above (a constant 0.75 ′′ ) the dielectric surface decrease as a function of mass flow rate, as shown in Fig. 4.10. As decreasing mass flow rate (while holding all other beam settings constant) directly decreases CEX ion flux for this plasma source, the reductioninmeasureddensitiesisfurtherproofthatCEXionsarethedominantionspecies intheshadowedregion. 4.3 Plasma Species Control with CEX Plate and Final ExperimentalSetup Having identified the dominant ion population in both the main beam axial flow and shad- owedradialflowregions,amethodwasdevelopedtosimultaneouslycontrolCEXionflux and electron kinetic energy to the dielectric surface. The method is displayed in Fig. 4.12. TheCEXplatenegativebiascontrolsthemagnitudeofCEXionfluxbehindthesimulated lunarobjectbycollectingorredirectinglowenergyionsthatflowintotheopeningbetween thetopfaceofthesimulatedlunarobjectandtheCEXplate. Conversely,electronsthatflow into the opening are rejected from the plate and acquire kinetic energy of approximately theCEXplate bias magnitude in the -zdirection. Thesheaththickness overthe negatively biasedCEXplateisdeterminedasfollowsinEqns.4.5-4.11 ϕ edge = kT e 2e (4.5) whereϕ edge isthesheathedgepotentialfromcontinuityandconservationofenergy, ^ ϕ edge = eϕ edge kT e = 1 2 (4.6) 84 X,mm Z,mm 300 350 400 450 500 -80 -60 -40 -20 0 20 40 60 80 100 120 Ji_radial A/m 2 5.000E-04 4.687E-04 4.375E-04 4.063E-04 3.750E-04 3.437E-04 3.125E-04 2.813E-04 2.500E-04 2.188E-04 1.875E-04 1.563E-04 1.250E-04 9.375E-05 6.250E-05 3.125E-05 0.000E+00 4” x 4” x 4” Alumina silicate Cube +3V (a) FaradayprobeJ i radial2Dcontour −25 −20 −15 −10 −5 0 5 10 15 20 25 −2 −1 0 1 2 3 4 5 6 7 8 x 10 −9 RPA current vs. Ion Retarding Voltage Voltage, V Grid current, A Wall + Back plate 4th degree Noise threshold (b) RPAlowvoltagescan Figure4.11: Shadowedregionradialflowscans 85 ^ n edge = n edge n e = n e exp ( ^ ϕ edge ) n e = 0:6065 (4.7) where ^ ϕ edge isthenormalizedsheathedgepotentialand ^ n edge isthenormalizedsheathedge density, edge d = √ 1 ^ n edge = 1:2840 (4.8) d sh edge = √ 4 p 2 9 ( V cexplate T e )3 2 1 M psh = 3:7 (4.9) whereV cexplate =-31VistheCEXplatebias,T e =4eV,andM psh =1isthepre-sheathion Machnumber, ^ d sh = d sh edge edge d = 3:71:2840 = 4:75 (4.10) where ^ d sh isthenormalizedsheaththicknessovertheCEXplate, d sh = ^ d sh d = 4:751:1cm = 5:23cm = 2:06inches (4.11) andwhered sh isthesheaththicknessoverthenegativelybiasedCEXplateand d iscalcu- latedfromelectrondensityn e =210 12 m 3 atthefrontedgeoftheCEXplate. The2 ′′ gap between the top face of the simulated lunar object (cube) and the CEX plate was chosen based on this calculated sheath thickness. The CEX plate is 36 gauge (5 mil) thick alu- minum, and 4 ′′ wide and 5 ′′ long. From the top of the final experimental setup (Fig. 4.13) facing downwards, the CEX plate dimensions match and cover the dielectric test bed sur- face. Thecubeandtestbedbackingplate(5milthick)weregroundedforconsistency. 4.3.1 PlasmaSpeciesControlTestResults MeasurementsoftheradialCEXionfluxandtotalelectronenergyintheshadowedregion as a function of CEX plate bias have confirmed the CEX plate control method. One- dimensional measurements in the x direction of radial CEX ion flux and total electron 86 4” x 4” x 4” Alumina silicate Cube X Z 2” 5” V cexplate Φ surf Φ surf Φ surf Φ surf Main beam ions CEX ions Electrons CEX plate Figure4.12: LunarsimulatorfacilityandCEXplatesketch .915 m 1.05 m .040 m .230 m 0 V Neutralizer -15.5 V 0 V 2D scan area .1143 m .2794 m .1524 m .330 m .0254 m .2286 m X Z CEX plate Cube Dielectric Figure4.13: Finallunarsimulatorfacilityexperimentalsetup density above (a constant 0.75 ′′ ) the dielectric surface for multiple CEX plate biases are given in Fig. 4.14. The plasma beam setting was 1100 eV beam energy, 10 mA beam current, 10 mA neutralizer current, and 4 sccm argon mass flow rate for each data case. Fig. 4.14 (a) shows that the CEX plate actually increases CEX ion flux to the dielectric surfaceforCEXplatebias>-25V.Thisisduetobeamconfinementinthe+zdirection. However, for CEX plate bias < -30 V, CEX ion flux is reduced below the nominal (flux 87 with CEX plate removed) value, effectively screening out CEX ions. Fig. 4.14 (c) and (d) show repeated scans for -31 V and -40 V CEX plate biases. Fig. 4.14 (b) shows that electron density for the maximum CEX ion flux case (CEX plate = floating) and nominal case (CEX plate removed) are approximately equal. This indicates the thermal nature of theelectronswhentheCEXplateiseithernotpresentornotbiased. 260 280 300 320 340 360 380 400 −2 0 2 4 6 8 10 x 10 −4 X, mm FP radial Ji, A/m 2 Measured FP radial Ji, FP bias = −40V CEX Ji nominal CEX plate = floating CEX plate = −10V CEX plate = −20V CEX plate = −30V CEX plate = −40V (a) J i radialCEXionflux 260 280 300 320 340 360 380 400 10 9 10 10 10 11 10 12 X, mm LP ne, m−3 Measured LP electron density Ne nominal CEX plate = floating (b) n e 260 280 300 320 340 360 380 400 −2 0 2 4 6 8 10 x 10 −4 X, mm FP radial Ji, A/m 2 Measured FP radial Ji, FP bias = −40V CEX Ji nominal CEX plate = −31V (c) J i radialCEXionflux,-31V 260 280 300 320 340 360 380 400 −2 0 2 4 6 8 10 x 10 −4 X, mm FP radial Ji, A/m 2 Measured FP radial Ji, FP bias = −40V CEX Ji nominal CEX plate = −40V (d) J i radialCEXionflux,-40V Figure4.14: 1DJ i radialCEXionfluxandtotaln e 0.75 ′′ abovedielectricsurface WhentheCEXplateisbiasedtheelectronsgainkineticenergyapproximatelyequalto the plate bias. Figure 4.15 gives experimental confirmation of this effect in the shadowed region. Figure4.15(a)showsthatmeasuredLangmuirprobeplasmapotentialincreasesfor morenegativeCEXplatebias. However,iftheplasmaisnolongerthermal,stationaryand 88 Maxwellian(i.e. if theCEXplateisincreasingelectronkineticenergy)thentheLangmuir probe measured plasma potential is actually estimating electron kinetic energy [3]. Fig- ure4.15(a)additionallyshowsanindependentmeasurementoftrueplasmapotentialatthe samelocationbytheemissiveprobe,whichisnotaffectedbyincreasedkineticenergy. The divergenceofthetwoplasmapotentialmeasurementsconfirmsthekineticenergyincrease. Electronthermalenergyremainslow(2-7eV)asshowninFig.4.15(b). Electrondensity isreducedbynearlyanorderofmagnitudeasshowninFig.4.15(c). Figure4.15(d)shows that the radial Faraday probe in the shadowed region must be biased at least as negatively astheCEXplatetorejectelectronsthathavegainedkineticenergy. −50 −40 −30 −20 −10 0 −25 −20 −15 −10 −5 0 5 10 15 20 25 CEX plate voltage, V Measured plasma potential, V Measured plasma potential vs. CEX plate voltage Langmuir probe Emissive Probe (a) EmissiveandLangmuirprobeϕ −50 −40 −30 −20 −10 0 0 1 2 3 4 5 6 7 8 9 10 CEX plate voltage, V LP electron temperature, eV Measured LP electron temperature (b) LangmuirprobeT e −50 −40 −30 −20 −10 0 10 9 10 10 10 11 10 12 CEX plate voltage, V LP ne, m−3 Measured LP electron density (c) Langmuirproben e −60 −50 −40 −30 −20 −10 0 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 x 10 −3 FP bias, V FP radial Ji, A/m 2 FP radial current vs. FP bias, CEX plate = −40V (d) FaradayprobeJ i radial Figure4.15: ExperimentalelectronkineticenergyincreasebyCEXplatebias 89 4.4 ExperimentalPlasmaWakeExpansion Previously presented lunar simulator facility plasma wake expansion (Figs. 4.4 - 4.7) has beenin+z,awayfromthedielectricsurface. However,forlunarsurface-plasmainteraction studies, the plasma wake expanding in the direction of the surface is of primary interest. Therefore, the desired experimental plasma wake expansion direction is -z, towards the dielectric surface. In addition to controlling CEX ion flux and electron kinetic energy, the CEX plate accomplishes this result by preventing main beam plasma expansion in the +z directiononcehighenergyionsflowintotheopeningbetweenthetopfaceofthesimulated lunarobjectandtheCEXplate. 4.4.1 ExperimentalWakeExpansionResults Using the final experimental setup and 2D scan area given in Fig. 4.13, the plasma diag- nostics probe suite was employed to obtain plasma potential, total electron density, and axialiondensityinthelunarsimulatorfacilityplasmawake. Theplasmabeamsettingwas nominal,CEXplatebiaswas-31.2V,andthesimulatorlunarobject(cube)wasgrounded. ResultsaregiveninFigs.4.16-4.17. Thelowerleftcornerofthe2Dscanareais1 ′′ behind thecubeand2.5 ′′ abovethedielectricsurface. Figure4.16(a)showsthatthelunarsimula- torfacilitymainbeamisquasi-neutral(ϕ p 0V)butthatelectronsdriveϕ p negativeinthe shadowedregion,duetoeffectiveCEXfluxscreeningbytheCEXplate. Thespacecharge transitiontonegativeintheshadowedregionisclearlyevidentinFig.4.16(b). Figure4.17 (a) shows that main beam ions are expanding into the void region opposite the CEX plate. Thisexpansiondirectionisalsooppositeinitiallunarfacilitysimulatorexpansiondirection (Fig. 4.5). Figure 4.17 (b) shows an electron density decrease outside of the main beam plume. Anexpansioncomparisonisnextmadetoanalyticalandnumericalmodels. 90 X,mm Z,mm 300 350 400 450 500 -100 -50 0 50 100 phiV 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 -13 -14 -15 -16 -17 -18 -19 -20 CEX plate = -31.2 V Alumina silicate 4” x 4” x 4” 1” 2.5” Cube GND (a) ϕ p X,mm Z,mm 300 350 400 450 500 -100 -50 0 50 100 niaxi-nem -3 1.00E+13 9.33E+12 8.67E+12 8.00E+12 7.33E+12 6.67E+12 6.00E+12 5.33E+12 4.67E+12 4.00E+12 3.33E+12 2.67E+12 2.00E+12 1.33E+12 6.67E+11 0.00E+00 -6.67E+11 -1.33E+12 -2.00E+12 Alumina silicate 4” x 4” x 4” 1” Cube GND 2.5” CEX plate = -31.2 V (b) n i -n e Figure4.16: Finallunarsimulatorfacilityplasmapotentialϕ p andspacechargen i -n e 91 X,mm Z,mm 300 350 400 450 500 -100 -50 0 50 100 niaxim -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 6.00E+10 1.00E+10 6.00E+09 1.00E+09 Alumina silicate 4” x 4” x 4” 1” Cube GND 2.5” CEX plate = -31.2 V (a) n i axial X,mm Z,mm 300 350 400 450 500 -100 -50 0 50 100 nem -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 6.00E+10 1.00E+10 6.00E+09 1.00E+09 Alumina silicate 4” x 4” x 4” 1” Cube GND 2.5” CEX plate = -31.2 V (b) n e total Figure4.17: Finallunarsimulatorfacilityaxialiondensityn i andtotalelectrondensityn e 92 4.4.2 AnalyticalWakeExpansionComparison MainbeamionexpansionintheshadowedregionwasmodeledanalyticallyperEqn.2.14, where n io = 3.3 10 13 m 3 is the axial ion density at the lunar facility simulator inlet (opening between cube and CEX plate), C s andM o are given in Eqns. 4.12 - 4.13, and istheradialexpansionangleinthe-zdirection. C s = √ kT e m i (4.12) M o = v io C s = 33:16 (4.13) Analytical ion expansion in the experimental 2D scan area is given in Fig. 4.18 (a). This plot indicates that the experimental main beam ion expansion shown in Fig. 4.17 (a) gen- erally follows the ideal (analytical) result. Note that the analytical solution assumes a Maxwell-Boltzmann distribution of electrons, while the experimental electron distribution is modified by the CEX plate. Additionally, the analytical solution assumes ion flow par- allel to the cube top face, while experimental flow is not precisely parallel due to slight sourcemisalignmentandbeamhalf-angledivergence. Deviationbetweentheexperimental andanalyticalresultsisprimarilydrivenbythesetwofactors. 4.4.3 NumericalWakeExpansionComparison Main beam ion expansion in the shadowed region was also modeled with a 2D numerical simulation(development,validationandspecificsarediscussedinthesubsequentsection). All simulation variables were normalized. Normalized simulation 2D domain dimensions and ion density were converted into physical units by multiplying by experimental Debye length and axial ion density, respectively, at the lunar simulator facility inlet. Ion expan- sion in the experimental 2D scan area from the numerical simulation is given in Fig. 4.18 (b). Thisresultwellfollowstheanalyticalsolutionandgenerallyfollowstheexperimental 93 X,mm Z,mm 300 350 400 450 500 -100 -50 0 50 100 niANLm -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 6.00E+10 1.00E+10 6.00E+09 1.00E+09 Alumina silicate 4” x 4” x 4” 1” Cube GND 2.5” (a) Analyticaln i X,mm Z,mm 300 350 400 450 500 -100 -50 0 50 100 niSIMm -3 1.00E+13 9.45E+12 8.90E+12 8.35E+12 7.80E+12 7.25E+12 6.70E+12 6.15E+12 5.60E+12 5.05E+12 4.50E+12 3.95E+12 3.40E+12 2.85E+12 2.30E+12 1.75E+12 1.20E+12 6.50E+11 1.00E+11 6.00E+10 1.00E+10 6.00E+09 1.00E+09 Alumina silicate 4” x 4” x 4” 1” Cube GND 2.5” (b) Numericalsimulationn i Figure4.18: Analyticalandnumericalsimulationionexpansionintheshadowedregion 94 solution. Note that the numerical model does not include CEX ions, the CEX plate, or the dielectric surface. Also, the simulated lunar object is modeled as a thin plate and not a cube. Whilehavinglittleeffectonthenumericaltoanalyticalwakeexpansioncomparison, theseaspectscontributetothedeviationbetweenthenumericalandexperimentalresults. 4.5 Numerical Plasma Wake PIC Code Development and Validation A 2D electrostatic full particle Particle-In-Cell (PIC) code has been developed, validated, and applied to the lunar simulator facility experimental setup. The code models both ions and electrons as macro-particles and solves the electrostatic Poisson’s equation self- consistently. Biasedandfloatingplatesareselectivelyincorporatedintothedomaintorep- resent lunar objects. Plasma flow is mesothermal, and the angle of attack is variable. The general numerical recipe is displayed in Fig. 4.19. The domain is established at timestep = 0. Then for every timestep, particles are injected, charge is distributed to domain nodes, potential and electric fields are solved, particles are moved, charge is deposited to simu- lated lunar objects, boundary conditions are applied, and particle statistics are performed. Thiscyclecontinuesuntilsteadystateortimestep=final. Inject particles Setup domain (t = 0) Node charge Solve potential field Solve electric field Move Particles Charge deposition Particle statistics Apply boundary conditions Steady state (t = final) Figure4.19: Generalnumericalrecipe 95 4.5.1 DomainSetup The domain is two dimensional and comprised of uniform square cells. Cell size and domain length and width are normalized by the Debye length, such that the domain is composed of unit cells. The simulated lunar object is incorporated into the domain by biasing or floating a set of domain nodes. An example 20x20 cell domain with a biased platefromthepoints(4,5)to(16,5)isgiveninFig.4.20(a). X Y 0 5 10 15 20 0 5 10 15 20 phi hat 0.0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 -0.7 -0.8 -0.9 -1.0 (a) 20x20celldomainandbiasedplate (b) Mesothermalplasmaflowintobiasedplate Figure4.20: Numericaldomainsetupandmesothermalplasmaflow 4.5.2 ParticleInjection The simulation ions are protons, and both ion mass and electron mass are normalized to electron mass, meaning that the electrons and ions have a real mass ratio of 1:1836. The flowismesothermalsuchthat ^ v te > ^ v beam > ^ v ti perEqns.4.14-4.16 ^ v te = v te v te = 1 (4.14) ^ v beam = v beam v te = 0:307 (4.15) ^ v ti = v ti v te = 0 (4.16) 96 where ^ v te , ^ v beam ,and ^ v ti arethenormalizedelectronthermalvelocity,normalizedionbeam velocity, and normalized ion thermal velocity for the nominal experimental plasma beam setting, respectively. Particles are injected to a random position within the first 1 10 of the loading cells (left and bottom domain edge cells) for each timestep. Injected electrons are sampledfromaMaxwellianvelocitydistributionandinjectedionsaresampledfora“cold” nearly constant ion velocity distribution. An example flow domain is given in Fig. 4.20 (b), where ions and electrons are ejected into the bottom domain cells, such that ion flow velocityisnormaltotheplateandelectronvelocityisthermal. 4.5.3 NodeCharge Particle charge within each cell is distributed to the surrounding 4 nodes, as shown in Fig. 4.21. Particle position with respect to the surrounding cell nodes determines the size of each “area”, and thus the percentage of charge to allocate to each node. The charge of all particles within the domain is cumulatively allocated to the domain nodes for every timestep. i =1 j =1 Area 1 Area 2 Area 3 Area 4 Area 1 Area 3 Area 2 Area 4 Figure4.21: Nodechargeallocationscheme 97 4.5.4 SORPotentialandElectricFieldSolver The potential field is obtained by numerically solving the electrostatic Poisson’s equation given in Eqn. 4.17. The right-hand side (total charge density) of Eqn. 4.17 is calculated from node charge. The left-hand side is discretized with a second order central difference methodasshowninEqn.4.18foreachnode (i;j). r 2 ϕ = " 0 (4.17) r 2 ϕ i;j = ϕ i1;j +ϕ i;j1 4ϕ i;j +ϕ i;j+1 +ϕ i+1;j ∆ 2 (4.18) Next,thecentraldifferenceschemeisputintothematrixequationformofEqn.4.19 A ⇀ ϕ = ⇀ B (4.19) where A is the coefficient matrix from central difference and the B matrix is node charge. The ϕ matrix is solved with the simultaneous over-relaxation (SOR) algorithm given in Fig. 4.22. The SOR over-relaxation parameter is 1 ! SOR 2. The electric field is obtained by taking the negative gradient of potential per Eqn. 4.20. The electric field in eachdimensioniscalculatedforeachnode (i;j)perEqns.4.21and4.22. ⇀ E =rϕ (4.20) E i;j j x = ϕ i+1;j ϕ i1;j 2∆ (4.21) E i;j j y = ϕ i;j+1 ϕ i;j1 2∆ (4.22) 98 φ=φ − ω SOR ⋅residual −4 residual= A⋅φ −B choose initialφ Converge if convergence criterion met (residual < tolerance) Figure4.22: SORalgorithm 4.5.5 ParticleMover Particleaccelerationisobtainedfromtheelectrostaticforceappliedbytheelectricfieldper Eqn. 4.23. The acceleration in each simulation dimension is calculated independently for ions and electrons. Acceleration is then used to update particle velocity, which in turn is usedtoupdateparticleposition,perEqns.4.24and4.25,respectively. Velocityandposition ofallparticlesareupdatedineachsimulationdimensionforeverytimestepdt. ⇀ a = q ⇀ E m (4.23) ⇀ v f = ⇀ v i + ⇀ adt (4.24) ⇀ x f = ⇀ x i + ⇀ v f dt (4.25) 4.5.6 ChargeDeposition In the numerical simulation the simulated lunar object is modeled as a plate. The plate can be biased or floating, and is either a line of nodes in the domain interior as shown in Fig. 4.23 (a), or a two line set of nodes at the top of the domain as shown in Fig. 4.23 (b). In the simulation a biased plate is at constant potential and absorbs all particles that “hit” 99 theplate. ChargeiscumulativelydepositedtofloatingplatenodessuchthattheSORsolver yieldsadirectsolutiontofloatingplatepotential. inlet inlet absorption absorption α (a) Plateaslineofnodes inlet inlet absorption symmetric α (b) Plateastwolinesetofnodes Figure4.23: Plategeometryanddomainboundaryconditions 4.5.7 ApplyBoundaryConditions Potential field boundary conditions are either inlet, absorption, or symmetric, as shown in Fig. 4.23. Particles enter the domain along the inlet boundaries (potential held at 0). Absorptionboundariesarealsoheldat0,whilesymmetricboundariesmirrortheinnerline node potential. It is noted that absorption field boundary conditions may over constrain small domains (when the potential gradient between the field and absorption boundary conditionsintroducesinstabilities). Inthesesituationsitisnecessarytousereflection(such as from a chamber wall or object) or open (with one-sided flux from infinity) boundary conditions. Particle boundary conditions surrounding the domain are either absorption (particles that leave domain are discarded) or symmetric (mirror-like specular reflection). Particle plateboundaryconditionsareabsorption. 100 4.5.8 ParticleStatistics Thenumberofparticles,nodecharge,depositedcharge,kineticenergy,andfield(potential) energy in every cell and in the entire domain are recorded for every timestep and cumu- latively. These statistics are used to confirm a steady state numerical solution and energy conservation. 4.5.9 PICcodeValidation Thenumericalcodewasvalidatedbytestingitagainsttheanalyticalexpansionsolutionin Eqn.2.14. Thisanalyticalsolutionisonlyvalidwhenthesheathovertheobjectcausingthe expansioncanberegardedasinfinitelythincomparedtotheexpansionfanregion(i.e. when theleadingedgeeffectcanbeneglected). ThisrestrictionisencompassedbyEqn.4.26 = L 2 L plate = M o d sh L plate 1:3M o j ^ ϕ plate j 3=4 ^ L plate < 1 (4.26) where ^ ϕ plate isthenormalizedsimulationplatepotentialand ^ L plate isthenormalizedsimu- lationplatelength(lengthofplateinnumberofsimulationcells). Thesimulationtestsetup is displayed in Fig. 4.23 (b) and Fig. 4.24, where angle of attack = 0 and where o is calculatedfromEqn.4.27. o = sin 1 ( 1 M o ) = 1:728 (4.27) The numerical inputs used to test against the analytical solution are given in Fig. 4.25. After 50,000 timesteps both test cases are converged, as evidenced by energy time history in Fig. 4.26. The sharp initial energy increase occurs as particles fill the domain (empty at t=0). Steadystateisreachedwhenparticlefluxinandoutofthedomainisapproximately equal. FieldenergyisdominantinCase2duetothemuchlargerplatebias. Simulationand analyticalexpansionpotentialsolutionsaregiveninFig.4.27. Case1 =.591inFig.4.27 101 Biased plate φ ∧ plate L ∧ plate n ∧ e ~ n ∧ i θ o δθ Figure4.24: Simulationvalidationtestsetup (a)and(b)showsthatthesimulationresultsfollowtheanalyticalexpansionsolution. Slight deviation between the two solutions is attributed to absorption field boundary conditions. Case 2 = 3.321 in Fig. 4.27 (c) and (d) shows that the analytical solution is not valid under the inputted conditions, as expected. Therefore, the numerical code is validated by showing that it correctly follows and deviates, respectively, from the analytical solution as a function of the restriction parameter . Note that the analytical solution is identical for Case1andCase2asthedomainsizeandplatesizeisheldconstant. Variables Xcell Ycell L hat plate Φ hat plate ξ 3mestep Converged Case 1 100 12 73 -‐1 .591 50000 Yes Case 2 100 12 73 -‐10 3.321 50000 Yes Figure4.25: Simulationtestcasesinputs 4.5.10 LunarSimulatorFacilityPICcodeApplication Thenumericalcodewasnextadaptedtosimulateexpansionoverabiasedorfloatingplate at a variable angle of attack as displayed in Fig. 4.23 (a). This test setup with 90 represents flow into and over the experimental simulated lunar object. Domain size, angle of attack, plate length, plate potential, ion flow velocity and electron thermal velocity are variable code inputs. Figure 4.28 is the flow domain with normalized nominal plasma 102 (a) Case1 (b) Case2 Figure4.26: Simulationtestcaseenergyhistory X Y 20 30 40 50 60 70 80 0 2 4 6 8 10 phiV 0.00 -0.20 -0.40 -0.60 -0.80 -1.00 -1.20 -1.40 -1.60 -1.80 -2.00 (a) Simulationpotentialcase1, =.591 X Y 0 10 20 30 40 50 60 70 0 2 4 6 8 10 phiV 0.00 -0.20 -0.40 -0.60 -0.80 -1.00 -1.20 -1.40 -1.60 -1.80 -2.00 (b) Analyticalpotentialcase1 X Y 20 30 40 50 60 70 80 0 2 4 6 8 10 phiV 0.00 -1.00 -2.00 -3.00 -4.00 -5.00 -6.00 -7.00 -8.00 -9.00 -10.00 -11.00 -12.00 -13.00 -14.00 -15.00 -16.00 -17.00 -18.00 -19.00 -20.00 (c) Simulationpotentialcase2, =3.321 X Y 0 10 20 30 40 50 60 70 0 2 4 6 8 10 phiV 0.00 -0.20 -0.40 -0.60 -0.80 -1.00 -1.20 -1.40 -1.60 -1.80 -2.00 (d) Analyticalpotentialcase2 Figure4.27: Numericalandanalyticalexpansionpotentialsolutionsoverabiasedplate 103 beam settings and a normalized biased plate potential of -1. The numerical domain size was normalized to the shadowed region of the lunar simulator facility by the experimental Debye length at facility entry. The numerical ion expansion results are given in Fig. 4.18 (b)anddiscussedintheNumericalWakeExpansionComparisonsub-section. Figure4.28: Numericalflowdomain 4.6 SummaryandConclusions The mesothermal plasma source bombards a given surface with high energy ions, low energy ions and low energy electrons. Charging analysis of the given surface requires accurate incident flux knowledge. The design, testing, and validation of an experimental “Lunar Simulator Facility” has been carried out to provide this requirement. The facil- ityiscapableofdifferentiatinglowenergyCEXfluxfromhighenergyprimarybeamflux, whilesimultaneouslycontrollingelectronkineticenergy. Additionally,axiallyflowinghigh energy ions that enter the facility expand into a shadowed region behind a simulated lunar object, inthe directionof a dielectricsurface. This conditionis directly applicableto solar 104 wind plasma flow over the lunar surface at low angles of attack, such as at the lunar ter- minator or in the lunar polar regions. Measured ion beam expansion in the experimental facility well matches analytical theory. Additionally, a full particle electrostatic PIC code has been developed to numerically model the plasma flow and expansion. These results well match both the analytical and experimental data, further validating the experimental lunarsimulatorfacility. 105 CHAPTER 5: EXPERIMENTAL INVESTIGATIONS OF LUNAR SURFACE CHARGING Likeanyobjectinaplasma,thelunarsurfacechargestoanelectrostaticpotentialthatmini- mizesthetotalincidentcurrent. Therearefiveprimarychargingcurrentsources: solarwind plasma electrons, solar wind plasma ions, UV induced photo-electrons, particle induced secondary electrons, and backscattered primary electrons. The “Lunar Simulator Facility” developed in the previous chapter is used to carry out a systematic study of dielectric sur- facechargingbyexperimentallysimulatedsolarwindplasmafluxatalowangleofattack. Thisfluxiscapableofsimulatingallprimarylunarchargingcurrentsourcesexceptphotoe- mission,whichwillnotoccurinlunarregionsofshadow. Surfacechargingintheseregions isexperimentallyconsideredherein. Planar dielectric material, dust dielectric material, and the lunar simulant JSC-1A are charged as a function of ion flux, electron flux, and electron kinetic energy. Steady state surface potentials are recorded and surface charge decay is tracked over time. Dielec- tric charging results are quantified and previously unreported ion-driven charge deposition dependenciesarefound. Surfacegrainchargingresultsarecomparedtothesingle,isolated grainchargingcase. 5.1 LengthScalingandExperimentalConsiderations Ideal simulation of lunar surface charging in a vacuum chamber environment is not possi- ble. Experimentalplasmadensitymustbeordersofmagnitudegreaterthanthatofthesolar wind,suchthattheexperimentalDebyelengthismuchlessthanthechambercharacteristic 106 length. Realistic physical dimensions are not simulated for obvious reasons. The lunar surfacepressureisontheorderof10 11 to10 13 Torr,whilechamberbackgroundpressure is on the order of 10 5 to 10 6 Torr. Lunar simulants are charged in place of actual lunar soil. Implicationsoftheseexperimentalconstraintsarediscussed. 5.1.1 LengthScaling The lunar simulator facility dimensions have been scaled per Eqn. 4.4 to lunar dimensions ontheorderof100’sofmeters. Thesimulatedlunarobject(cube)representsasmalllunar hill,rock, oroutcropping,suchasinFig. 5.1. Underlocalsunrise/sunsetconditionsandin thelunarpolarregionstheseobstaclesblocksolarwindplasmaflow,creatinglocalplasma depletionandplasmafluxdrivensurfacepotentialsandelectricfieldsonleewardandfloor faces,suchasinFig.5.1(a). Astheseobstaclesarecorollariestopermanentlylit/shadowed regionssuchasShackletonCrater,aproposedlunarsouthpoleoutpostlocation,analyzing thesurfacepotentialsformedisofparticularinterest. 5.1.2 BackgroundPressureImplications Due to the low solar wind plasma density and low lunar atmosphere neural density, CEX flux to the lunar surface is negligible. In the lunar simulator facility CEX flux is either mitigated with the CEX plate, or increased to simulate radial lunar ion flow at the ion acousticvelocityC s . Itisnotedthatscavengingofsurfacechargeduetoneutralcollisionsis muchmorelikelyintheexperimentthanonthelunarsurface. AsurfacechargingPICcode capableofselectivelyincludingornotincludinganeutralpopulationisunderdevelopment toassessthisexperimentalconstraint. 107 (a) Illustration[18] (b) Apolloastronautandrover(NASA) Figure5.1: Lunartopographicobstructiontosolarwind 5.1.3 Argonplasmavs. HydrogenplasmaImplications Argonisusedintheexperimentbecauseitisinertandcost-effective. Itisalsoeasytoionize (relative to hydrogen) by electron bombardment as argon ionization energy is 15.76 eV, electron-argoncollisioncrosssectionishigh,andargonmobilityislow. Intheexperiment argonionsareacceleratedto73km/s,whiletheaveragesolarwindionflowspeedis400 -600km/s. However,kineticenergyofthetwospeciesissimilarasgiveninEqns.5.1-5.2 KE argon = 1 2 m argon v argon 2 = 1100 eV (5.1) KE proton = 1 2 m proton v proton 2 = 835 eV (5.2) 108 where v argon is the average experimental argon ion velocity and v proton is average solar wind ion flow velocity = 400 km/s. A solar wind flow velocity of 600 km/s results in KE proton =1880eV. Ions at these energies that strike a surface are very likely to deposit charge (when an electronfromthesurfaceistransferredtoanincidention,creatingafreeneutral). Sputter- ingmayoccurbutisneglectedintheexperimentwithrespecttomeasuredsurfacepotential because it is a charge-neutral process. Ion implantation is extremely unlikely for ion ener- gies<10keV.Ion-inducedsecondaryemissionmaybesignificantatboththeexperimental and solar wind ion velocities (the ion-induced kinetic secondary electron emission energy threshold for both argon and hydrogen ions incident on clean metal surfaces and semi- conductors has been predicated to be300 eV [10, 72, 53], and this threshold is much lower for insulators [10, 6]). Ion-induced secondary electron emission is further discussed insubsequentsections. Duetosolarflares,traversethroughEarth’smagnetotailandplasmasheet,andelectron acceleration in the lunar wake, the lunar surface is frequently subject to electrons from a Kappa distribution [27, 22, 35]. However, average solar wind electrons are often modeled asMaxwellianwithathermalenergy=10-20eVandadriftvelocity=400km/s(average solar wind flow velocity) [57, 83]. The experimental lunar simulator facility follows this approach as electron thermal energy is2 - 7 eV and electrons are accelerated to at most 30 eV. In this energy range electrons that strike a surface are very likely to be absorbed withoutinducingsecondaryelectronemission. 5.1.4 LunarSimulantImplications Alumina silicate, Mykroy/Mycalex R ⃝ and JSC-1A are used in the experimental facility as lunarsimulants. Aspreviouslydescribed,eachmaterialpossessesrelativelyhighdielectric strengthandlowelectricalconductivityandlowdielectriclosses. Thesepropertiesarevery 109 similar to those of actual lunar soil, and they permit the materials to both readily charge andremainelectrostaticallycharged. Experimentalstudieshaveindicatedthatelectron-inducedsecondaryelectronemission from both lunar soil and lunar simulants is negligible below incident electron energies of 50eV[33,34]. PhotoemissioncreatedbythesolarspectrumisdominatedbyLymanalpha emission, which has properties of = 121.6 nm and E = 10.2 eV [11]. Under these con- ditions both lunar soil samples and JSC-1 simulant grains have been measured to yield photoelectron current density on the order of 1 to 4.5 A/m 2 [88, 2]. The lunar simulants usedhereinareexpectedtofollowtheseproperties(also,theexperimentalsamplesarenot exposedto50eVelectronsorUVflux). Itisnotedthattheselunarsimulantsdonotcon- tainthemagneticpropertiesoflunarsoil. Further,bothlunarmagneticpropertiesandlunar regolithDCconductivityvarywithtemperature. However,thelackofbothmagneticprop- erties and simulated lunar soil temperatures will not significantly affect the electrostatic surfacechargingprocessitself,althoughitmaysignificantlyalterchargedecay. 5.2 ExperimentalSetupandExperimentalFluxSpecies ThelunarsimulatorfacilitydielectricchargingexperimentalsetupisdisplayedinFig.5.2. Figure 5.2 (a) shows a top and side view sketch of the simulated lunar object (cube) and dielectrictestsurfaceswithrespecttotheionsource. Trekprobe1Dscansaretakeninthey directionoverbothsamplesurfaces. Plasmafluxaboutthezaxisissymmetrical(confirmed by testing), which allows a charging comparison under the same plasma conditions to be madebetweenthe-yand+ysurfaces. Figure5.2(b)showsatopview(CEXplateremoved) of the experimental setup with alumina silicate in the -y position and JSC-1A in the +y position. Asummaryofelectron-surfaceinteractionsingiveninthefollowinglist: 1. Primaryelectronsareabsorbed 110 (a) Sketch(nottoscale) 4” Cube Z Y 2” x 5” Al Silicate 2” x 5” JSC-1A Trek probe -Y +Y (b) Topview(CEXplateremoved) Figure5.2: Lunarsimulatorfacilitydielectricchargingexperimentalsetup 2. Primaryelectronsarebackscattered 3. Specimenemitssecondaryelectronsfromorbitals(notexclusive) 4. Specimenemitsx-raysorlight(cathodoluminescense) Electrons that strike a surface are light (with respect to ions) and absorbed as a function of energy and specimen thickness/composition. In the experiment absorption (without inducedemission)dominatestheelectron-surfaceinteractionprocess,asexperimentalelec- trons are low energy (2 - 40 eV) and the dielectric specimens are very thick with respect totheelectronpaththroughthespecimens. 111 Asummaryofion-surfaceinteractionsisgiveninthefollowinglist: 1. Ionreflection(chargenotdeposited) 2. Exclusive charge deposition (specimen transfers electron to incident ion which is neutralized) 3. Sputteringofspecimen 4. Secondaryelectronemissionfromspecimen 5. Ionimplantationinspecimen High energy ions (such as primary beam ions1100 eV) that strike a surface are heavy (with respect to electrons) and unlikely to reflect without depositing charge. Low energy ions(suchasCEXions30eV)aremuchmorelikelytoreflectwithoutdepositingcharge intheabsenceofAugerneutralization(seeEqn.5.3). Intheexperimentsputteringistaken as a charge-neutral process and ion implantation is negligible for energies < 10 keV. Ion- inducedsecondaryemissioncanoccurthrougheither“PotentialEjection”or“KineticEjec- tion”. Potentialejection(Augeremission)fromadielectricsurfaceoccurswhenanelectron gainssufficientenergyE k fromtheincidentionneutralizationprocesstoovercomethesur- facebarrierenergyperEqn.5.3 E k = E i 2(E g +) > 0 (5.3) where E i is the ionization potential of the gas ion, E g is the band gap energy of the solid andistheelectronaffinity. NotethatE g +isthephotoelectricthresholdforaninsulator or semiconductor. As shown in Fig. 5.3, E k for both argon and hydrogen ions incident on silicon oxide and aluminum oxide (primary constituents of most analyzed lunar soils and simulants) is negative, indicating that neither Auger neutralization nor potential ejection 112 occur. Molecular hydrogen is included as it may be used in the future as the experimental sourcegas. Dielectric E g eV χ eV SiO 2 8.9 1.0 Al 2 O 3 9.8 1.0 (a) Bandgapandelectronaffinity[75] Gas E i eV E k eV (SiO 2 ) E k eV (Al 2 O 3 ) Ar 15.76 -‐4.04 -‐5.84 H 2 15.4 -‐4.4 -‐6.2 H 13.6 -‐6.2 -‐8.0 (b) Ionizationpotentialandejectionenergy Figure5.3: Potentialejectionenergies The theory of ion-induced secondary electron emission from dielectrics (and met- als) by kinetic ejection is much less understood. Relevant projectile variables include species, mass, excitation, ionization state and whether the projectile arrives as a single atom, molecule or cluster of molecules [6]. Relevant target variables include elemental composition, atomic and electronic surface structure, magnitude of surface electrostatic fields,temperature,anddegreeofmagnetization(formagneticmaterials)[6,72]. Relevant geometricalfactorsincludeprojectileincidenceanglewithrespecttothesurfaceandcrys- tallographicplanes[10,6]. Aminimumincidention(argonorhydrogen)energythreshold of300eVhasbeenpredictedforkineticejectionfromcleanmetalsurfacesandsemicon- ductors[10,72,53],butthisthresholdmaybeaslow15-20eVforinsulators[10]. Oneof themainreasonsforthisisasurfacebarrierpotentialthathasbeenfoundtobeontheorder of11-15eVinmetals,butistheelectronaffinityininsulators(whichcanbeafractionof aneVorevennegative)[6]. 113 Figure5.4: Experimentalfluxsourcestodielectricsolidanddustsurfaces Therefore, in the experimental lunar simulator facility, three flux sources to the dielec- tric test surfaces are considered: low energy ion flux (simulating radially inflowing solar wind ions), low energy electron flux (simulating thermal solar wind electrons), and ion- induced kinetic secondary electron flux. These experimental flux sources are displayed in Fig. 5.4 over a solid dielectric surface and a dust dielectric surface. Any two given sur- facesarechargedsimultaneouslyinthelunarsimulatorfacilityunderthesameplasmaflux conditionstoyieldadirectsurfacechargingstatecomparisonasafunctionofmaterialand surfaceproperties. 5.3 Dielectric Surface Charging: Alumina Silicate Plate vs. Mykroy/Mycalex R ⃝ Plate The lunar simulator facility with one alumina silicate plate and one Mykroy/Mycalex R ⃝ plateinstalledispicturedinFig.5.5. TheCEXplatewasthecontrolvariable. Itwaseither floating, -5 V, -10 V, -20 V, -30 V, or removed. The CEX plate setting controlled low energy ion flux and electron kinetic energy to the dielectric surfaces (Figs. 4.12 - 4.15). Theplasmabeamsettingwasnominal(Table5.1)foreachdataset. Chargingtimeforeach datasetwas10minutes. IndependentTrekprobescansconfirmedsteadystatepotentialsfor chargingtimegreaterthan5minutes,i.e. dielectricsurfacepotentialrespondsdynamically to incident currents. After 10 minutes of charging time the plasma beam was directly shut 114 Table5.1: Nominalplasmabeamsettings I,mA _ m,sccm ϕ o ,V 10 4 1100 down (no change in discharge current or voltage prior to shut down), and the Trek probe scanned1Ddielectricsurfacepotentialalong-yand+yasshowninFig.5.5. Mykroy/Mycalex® Alumina silicate -Y +Y Figure5.5: AluminasilicateandMykroy/Mycalex R ⃝ platesforlunarsimulatorfacility Figure 5.6 gives the resulting surface potentials 1 minute and 6 minutes after plasma beam shutdown (the Trek probe scan takes 1 minute and the 6 minute scan was performed to ensure minimal charge decay). All values are referenced to GND (the starting and end- ing Trek probe potential reference during the scans). The plasma potential reference is also 0 V (plasma potential in the main beam above the dielectric surfaces). The results show that ion flux is driving the surface potential difference between alumina silicate and Mykroy/Mycalex R ⃝ fortheCEXplate=FLOATINGandCEXplate=-5Vcases. AsCEX ion flux is reduced by increasing CEX plate negative bias, the surface potential solutions forthetwomaterialsconverge. Surfacepotentialconvergenceisnotduetoelectronkinetic energy increase (electron kinetic energy increases with CEX plate negative bias). This is 115 proven by Fig. 5.6 (f), which shows that the solutions converge when the CEX plate is removed(i.e. forlowelectronkineticenergyandlowionflux). Notethatforeachdatacasethesamechargingprocedurewasfollowed: 1. Ensure0Vdielectricpotentialsunderroomconditions 2. Holdsamplesundervacuumfor1hour 3. BakewithWatlowpolyimidesheetheater 4. 10minutechargingofsamplesinlunarsimulatorfacilityundernominalbeamsetting 5. Shutdownplasmabeam 6. Trekprobesurfacepotentialsscans 5.4 Dielectric Surface Charging: Alumina Silicate Plate vs. AluminaSilicateDust The lunar simulator facility with one alumina silicate plate and one alumina silicate dust bed (crushed and sieved to 100 m 40 m) installed is pictured in Fig. 5.7. Again, the CEX plate was the control variable. It was either floating, -5 V, -10 V, -20 V, -30 V, or removed. The plasma beam setting and charging procedures used for the alumina silicate andMykroy/Mycalex R ⃝ platechargingcaseswererepeatedforeachCEXplatesetting. Figure 5.8 gives the resulting surface potentials 1 minute and 6 minutes after plasma beam shutdown. When ion flux is relatively strong (CEX plate = FLOATING and CEX plate = -5 V cases) the alumina silicate dust surface potentials are driven higher than the aluminasilicatesolidsurfacepotentials. Dustandsolidsurfacepotentialsolutionsconverge forreducedionfluxatbothincreasedelectronenergyandnominallowelectronenergyset- tings, as shown in Fig. 5.8 (e) and (f). This proves that ion flux is responsible for the 116 −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, CEX plate = FL Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (a) CEXplate=FLOATING −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, CEX plate = −5 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (b) CEXplate=-5V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, CEX plate = −10 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (c) CEXplate=-10V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, CEX plate = −20 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (d) CEXplate=-20V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, CEX plate = −31 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (e) CEXplate=-30V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, NO CEX plate, 4 sccm Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time Mykroy/Mycalex, 1 min decay time Mykroy/Mycalex, 6 min decay time (f) CEXplate=REMOVED Figure 5.6: Alumina silicate and Mykroy/Mycalex R ⃝ surface potentials as a function of CEXplatesetting 117 -Y +Y Alumina silicate dust Alumina silicate solid Figure5.7: Aluminasilicateplateandaluminasilicatedustinlunarsimulatorfacility potential difference between the surfaces in Fig. 5.8 (a) and (b). If the dust and solid sur- faceshadthe same ion chargedeposition mechanismsthesolutions wouldhaveconverged in Fig. 5.8 (a) and (b) as well. Therefore, ion charge deposition is shown to be much more significant for a dust surface than a solid surface of the same chemical composition. Note thatthedielectrictestsurfaceswere“mirrored”inthelunarsimulatorfacilitytoensurethat charging results were consistent when the alumina silicate dust bed was moved to -y and the alumina silicate plate to +y. The experimental surfaces and mirrored results are given in Fig.5.9, and confirm the symmetryassumption (aCEXplate edgeeffectdrovethe high +ydustedgesurfacepotentialsinFig.5.9(c)). 118 −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 30 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = FL Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (a) CEXplate=FLOATING −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −5 V Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (b) CEXplate=-5V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −10 V Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (c) CEXplate=-10V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −20 V Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (d) CEXplate=-20V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −30 V Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (e) CEXplate=-30V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, NO CEX plate Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (f) CEXplate=REMOVED Figure5.8: Aluminasilicateplateandaluminasilicatedustsurfacepotentialsasafunction ofCEXplatesetting 119 -Y +Y (a) Aluminasilicateplate. Aluminasilicatedust. -Y +Y (b) Aluminasilicatedust. Aluminasilicateplate. −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 30 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = FL Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (c) Alignment(a)results −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 30 Y, mm Dielectric floating potential, V Measured dielectric plate potentials, 10 mA beam, CEX plate = FL Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 6 min decay time (d) Alignment(b)results Figure 5.9: Alumina silicate plate and alumina silicate dust mirrored charging in the lunar simulatorfacility 5.5 Dielectric Surface Charging: Alumina Silicate Plate vs. JSC-1ADust ThelunarsimulatorfacilitywithonealuminasilicateplateandoneJSC-1Adustbed(distri- bution10mto1mm)installedispicturedinFig.5.10. Onceagain,theCEXplatewasthe control variable. It was either floating, -5 V, -10 V, -20 V, -30 V, or removed. The plasma beamsettingandchargingproceduresusedforthealuminasilicateandMykroy/Mycalex R ⃝ platechargingcaseswererepeatedforeachCEXplatesetting. Figure 5.11 gives the resulting surface potentials 1 minute and 6 minutes after plasma beam shutdown. The alumina silicate solid and JSC-1A dust surface potentials converge under reduced ion flux for both the increased electron energy and nominal low electron energy settings, as shown in Fig. 5.11 (e) and (f), similar to the previous two charging experimentalsetups(aluminasilicate-Mykroy/Mycalex R ⃝ andaluminasilicatesolid-dust). 120 -Y +Y JSC-1A Alumina silicate Figure5.10: AluminasilicateplateandJSC-1Adustinlunarsimulatorfacility However, the charge decay rate between the alumina silicate solid and JSC-1A dust is increased as compared to the previous two charging setups. This effect will be discussed in a subsequent section. Also, the surface potential solutions only slightly diverge for increased ion flux, as shown in Fig. 5.11 (a) for the CEX plate = FLOATING case. The previous two charging setups and results showed that both material properties (chemical composition,etc.) andsurfaceproperties(solidvs. dust)maysignificantlyinfluencetheion chargedepositionprocess. Therefore,theresultsinFig.5.11cannotbeusedtoconclusively determine if material properties or surface properties cause the convergence/divergence of thealuminasilicatesolidandJSC-1Adustsurfacepotentials. 121 −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = FL Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (a) CEXplate=FLOATING −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −5 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (b) CEXplate=-5V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −10 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (c) CEXplate=-10V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −20 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (d) CEXplate=-20V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −30 V Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (e) CEXplate=-30V −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, NO CEX plate Alumina Silicate, 1 min decay time Alumina Silicate, 6 min decay time JSC−1A, 1 min decay time JSC−1A, 6 min decay time (f) CEXplate=REMOVED Figure 5.11: Alumina silicate plate and JSC-1A dust surface potentials as a function of CEXplatesetting 122 5.6 IonChargeDepositionMechanisms: Ion-inducedSec- ondaryEmissionYieldandIonReflection The electron flux J e and measured ion flux J i to the dielectric test bed Trek probe scan regionisgiveninTable5.2asafunctionofCEXplatesetting. Equation5.4givesJ e J e = en e 2 [ v te p exp ( ( v de v te ) 2 ) + v de ( 1+erf ( v de v te )) ] (5.4) whereelectrondriftvelocityv de isgiveninEqn.5.5 v de = √ 2eV cexplate m e (5.5) andV cexplate istheCEXplatebias. Electrondensityn e ismeasuredbytheLangmuirprobe. TheFaradayprobeyieldsadirectmeasurementofJ i . Next,theplanaraluminasilicatesurfacepotentialϕ AlSilplate foreachCEXplatesetting was set as the “baseline” or “zero ion-induced secondary electron emission” value. This isappropriatesincethesurfacepotentialcomparisonsaremadebetweenonedielectrictest bed (planar alumina silicate was the control surface) and the other, under identical lunar simulator facility plasma flux. In other words, the same plasma environment (independent ofplasmaprobeaccuracy)ischargingdifferentmaterialsanddifferentsurfacestodifferent potentialswithrespecttoeachother. Havingestablishedϕ AlSilplate asthebaseline(ion-inducedsecondaryelectronyield si = 0), the relative yields for Mykroy/Mycalex R ⃝ , alumina silicate dust, and JSC-1A were solvedfromEqn.5.6(forϕ measured >0)andEqn.5.7(forϕ measured <0) si = J e J i J i exp ( ϕ measured Tesec ) (5.6) 123 Table5.2: CEXplatesettingsandlunarsimulatorradialplasmaflux CEXplate Je,A=m 2 Ji,A=m 2 floating 8.01.610 4 4.50.510 4 -5V 4.30.910 3 3.20.410 4 -10V 6.01.210 3 2.70.310 4 -20V 8.51.710 3 0.90.110 4 -30V 1.10.210 2 0.90.110 5 removed 8.01.610 4 4.50.510 5 si = J e exp ( ϕ measured Te ) J i J i (5.7) where J i is set from measured ϕ AlSilplate and calculated J e , and T esec = 3 eV. The sur- face potential results and calculated ion-induced secondary electron emission yield si for each lunar simulator facility experimental charging setup and data case are displayed in Table 5.3. Each yield is calculated with respect to an alumina silicate plate reference si = 0 and the plasma flux for each particular data case. The calculated yields are clearly not absolute. Yields are not calculated when the ion flux is insufficient to drive a detectable surface potential difference between any two given dielectric test beds (CEX plate = -20 V, -30 V, and removed). Also, an experimental in-situ measurement of secondary electron fluxisnotperformedduetothenatureoftheexperimentalsetup. In Table 5.3 it is immediately obvious that si values> 3 are non-physical for an inci- dent 30 eV ion, as approximately 10 eV (the photoelectric threshold) is required to kineti- callyextractanelectron. Alikelyion-inducedsecondaryelectroncoefficientis1(although this value could be much higher for high energy ions). The mechanism to explain the surfacepotentialresultsisdiscussedinthefollowingthreesub-sections. 124 Table5.3: Tabulatedlunarsimulatorfacilitychargingresultsandcalculated si (—indicatesvalueisnon-physical) CEXplate ϕ AlSilplate ,V ϕ AlSildust ,V ϕ MMplate ,V ϕ JSC1A ,V siAlSildust siMM siJSC1A floating -1.9 12.9 5.8 0.1 39.6 3.7 0.6 -5V -11.1 -7.1 -8.0 -10.8 4.6 1.0 0.1 -10V -14.3 -12.3 -13.1 -13.0 1.4 0.3 0.3 -20V -21.2 -21.5 -20.3 -20.0 - - - -30V -31.6 -31.9 -31.2 -30.2 - - - removed -13.0 -13.5 -12.9 -13.1 - - - 5.6.1 Aluminasilicatesolidvs. aluminasilicatedust In the experiment, when low energy (30 eV) argon ion flux J i is of the same order as lowenergy(2-40eV)electronfluxJ e ,ionchargedepositionmechanismsonanalumina silicate dust layer drive an order of magnitude increase in surface potential, compared to an alumina silicate plate of the same chemical composition and thickness. Ion-surface interactions include charge deposition, ion-induced secondary electron emission and ion reflection. Ion-inducedsecondaryelectronemission Previous work by Richterov´ a et al. [62, 63, 64] has demonstrated enhanced electron- induced secondary electron yield from dust grains with respect to large planar surfaces. Thisenhancementmayoccurforatleasttworeasons: 1. Grain surface curvature that results in incidence angle variation along the grain sur- face 2. Primaryelectronpenetrationdepthontheorderofgrainsize Implantation (penetration) of the low energy experimental argon ions on the scale of the experimental dust grains (100 m) is negligible. However, the experimental surface topography is such that spacing and openings between surface grains are generally much greaterthanionsize. Thisallowsionstostrikeor“graze”surfacegrainsatincidenceangles 125 between0and90degrees. Ion-inducedsecondaryemissionyieldsbyeitherfastheavyions orfastprotonsincidenceonorgrazingplanarsurfacesmaybeashighas100’sofelectrons per incident ion [37, 6, 72, 5]. An example of this effect is shown in Fig. 5.12, where the grazingionincidenceangleislessthan1 withrespecttothesurface. 234 R.A. Baragiola / Ion induced EE . . Grazing ion . . . - l . . . * . . . . . . . * . _ l / . ._-. l . . . . *a _ 0. .._-a.. . Solid Fig. 8. Schematic drawing showing electron could originating from multiple-electron emission by a fast ion at grazing inci- dence on a surface. ion incidence of less than 1” to the surface [127]. One should expect even higher yields at lower energies, close to the stopping power maximum. One can visual- ize that the trajectory of the ion is accompanied by a cloud of ejected electrons (fig. 8). A convoy electron which would normally be ejected along the ion path with a velocity close to that of the ion, would be pushed to higher energies. It is proposed here that this is the case for energy shifts observed [128-1321 in convoy EE at grazing incidence which has been at- tributed to acceleration by the image charge induced by the projectile [133,134]. 5. Yields To calculate differential and total yields one has to develop quantitative descriptions of the mechanisms shown above and assemble them to derive observable quantities. No complete theory exists at this time al- though some special cases have been derived where simplifications are possible. For instance, at high veloc- ities, one can neglect recoil effects and consider that the projectile does not lose much energy over the escape depth. In this case, the number of primary events is independent of depth z. One then needs energy and angular distributions of these 1G electrons as input to a calculation of the electron transport and multiplication in the cascade. These can be obtained from atomic physics calculations or experiments, and one can use electron gas models for the valence band, which gives the most important contribution at low impact velocities. At high velocities, most inner shells can be excited, each contributing approximately in proportion to its number of electrons, and so it is important to describe well inner-shell excitations and the subsequent Auger decay. Once the excited electron sources are obtained, calculation of the electron cascade can proceed using transport theory or Monte Carlo simulations starting with electron scattering data from gas-phase collisions or free-electron models, as appropriate to the type of solid. Here the main difficulty is the lack of an ade- quate description of elastic scattering at the lower electron energies where quantum effects are impor- tant. Once the flow of the electrons at the surface is obtained one can calculate the transport through the barrier. In this case, the main difficulty is the estima- tion of the inner potential and quantum reflection effects. Calculations at low velocities need to take into account excitations by fast target recoils, which require an evaluation of the atomic cascade. Scattering of the projectile near the surface makes the ionization path longer inside the escape region. If the scattering is very strong, the projectile can backscatter producing addi- tional excitation. The maximum excitation is caused by a projectile which scatters and follows a trajectory parallel to the surface, just below it. This process is unlikely and will not contribute to any significant ex- tent to the total yield. It can, nevertheless, be observed as a finite probability of ejecting a very large number of electrons n (n >> 7). A good description of valence excitations by projec- tiles and recoils becomes more important since inner- shell contributions are small. To calculate the distribu- tion of 1G electrons one has to follow the trajectories of projectile and target atoms (e.g., by using Monte Carlo or molecular dynamics simulations). For each collision one calculates the inelastic energy transfer by using, for instance, Firsov’s formula or improved ver- sions [135,136]. If this calculated energy transfer is larger than U, one counts an excited electron from which one builds a distribution of ionization events, n(z) as a function of distance z to the surface. At low enough energies, cascade multiplication can be ne- glected, and one can equate n(z) to the depth distribu- tion of 1G electrons. If one neglects inelastic electron collisions that do not attenuate the electron flux, one can calculate the energy distribution of ejected elec- trons from g = T(E’)/n(z, E’) exp(-z/L(E’)) dz, (14) where E = E’ - I, from which one can deduce the total yield by integration. Another simple case is that of KEE near threshold. When the mass of the projectile is larger than that of the target atoms, the threshold behavior of KEE will be dominated by recoil effects, since the recoils can be faster than the projectile [53]. When the projectiles are Figure 5.12: Schematic showing multiple-electron emission by a fast ion at grazing inci- denceonasurface[6] A similar grazing effect may occur on the experimental dust surface when ion flux is onthesameorderaselectronflux. ThiseffectisillustratedinFig.5.13. Radiallyinflowing ions are normally incident on the planar dielectric surface shown in Fig. 5.13 (a), and ion grazing is unlikely to occur. When the surface is comprised of dust such that the spacing (a) Planarsurfacenormalincidence (b) Dustsurfacegrazingincidence Figure 5.13: Experimental normal incidence on planar surface and grazing incidence on dustsurface 126 and openings between surface grains is greater than the ion size as shown in Fig. 5.13 (b), ion grazing may occur along one or more grain surfaces. However, for the low energy argonionsintheexperiment,grazing(whenagivenionpassesthroughelectroncloudsand causes multiple-electron emission) emission is limited to at most 3 electrons per incident ion. Also, ion-induced secondary emission is mitigated when electron flux is dominant. Under this condition the surface potential is driven by the electron energy and becomes negative. Thenegativesurfacepotentialattractsthelowenergyionsdirectlytodustgrains, anddoesnotallowgrazingalonggrainsurfaces. Ionreflection The surface potential deviation between the planar and dust surfaces must then be primar- ily caused by ion charge that is not trapped (not deposited). Auger neutralization will not occur for argon ions incident on alumina silicate, and resonance neutralization is unlikely as dielectric orbitals are tight, compared to metals. Therefore, the incident 30 eV ions are likely to reflect without depositing charge. All of these processes/effects are valid on the granularsurfacecomparedtotheplanarsurface. However,ionreflectionwilltendtogofor- ward, due to the shape of the grains, as shownin the scanning electron microscopy (SEM) images in Fig. 5.14. The ions will then ricochet between grains and their positive charge will most likely be deposited on a solid dust grain. This results in heightened positive charge(potential)onthedustsurfacewithrespecttotheplanarsurface. Both ion-induced secondary electron emission and ion reflection explain the sharp reduction of siAlSildust in Table 5.3 as function of CEX plate negative bias. As the CEX plateisbiasedmorenegativelytheelectronsbothgainkineticenergyandbecomethedom- inantfluxspecies(aslowenergyionsarecollectedbytheCEXplate),andionchargedepo- sition mechanism differences between the planar and dust surfaces, respectively, become negligible. 127 (a) Planaraluminasilicatesurface,100mscale (b) Dustaluminasilicatesurface,100mscale Figure5.14: SEMimagesofplanaranddustaluminasilicateexperimentalsurfaces 5.6.2 Aluminasilicatesolidvs. Mykroy/Mycalex R ⃝ Compared to si =0 for the alumina silicate plate, the calculated ion-induced secondary electron emission yield siMM = 3.7 for the Mykroy/Mycalex R ⃝ plate when experimental J i isofthesameorderasexperimentalJ e . Whiletheyieldvalueitselfisnotvalid,thecalcu- lationdemonstratesthationchargedepositionvariessignificantlybetweenplanarinsulator surfaces of different chemical compositions under low energy argon ion flux. However, determination between ion-induced secondary emission and ion reflection is not made. Also, bandgapE g and electronaffinity valuesarenotavailableforMykroy/Mycalex R ⃝ , so whether the electron ejection would be potential or kinetic is not conclusive. Similar to aluminasilicatedustyield siAlSildust ,Mykroy/Mycalex R ⃝ plateyield siMM alsoappears to be a function of surface potential/dominant flux species, as the values decrease rapidly asJ i isreducedandJ e isincreasedbyincreasingCEXplatenegativebias. 5.6.3 Aluminasilicatesolidvs. JSC-1Adust Compared to si =0 for the alumina silicate plate, the calculated ion-induced secondary electronemissionyield siJSC1A valuesforJSC-1Aarealllessthanone. Itisnotpossible 128 to conclude whether material properties or surface properties or both are driving the ion charge deposition mechanisms and resulting surface potential deviation. The yield value drops off rapidly once J e >> J i , again confirming surface potential dependency on ion chargedeposition. 5.7 ChargeDecay Dielectric surface charge decay is a function of electrical resistivity, dielectric permittivity and time. Resistivity and permittivity are influenced by moisture, temperature, grain size, packing,andotherfactors. Intheexperimentchargedecaywasrecordedoveraneighthour time period for the alumina silicate plate, alumina silicate dust and JSC-1A -30 V (CEX platesetting)chargingcases. Thechargedtestsampleswereheldundervacuum(110 6 Torr) during this time period. The respective surface potentials as a function of time and the average potential difference between the alumina silicate plate and dust surfaces as a functionoftimearegiveninFigs.5.15and5.16. Figure 5.15 (a) shows that charge decay from both the alumina silicate plate and alu- mina silicate dust layer is minimal over the eight hour time period. Figure 5.16 (a) shows thattheaveragepotentialdifferencebetweenthealuminasilicateplateandaluminasilicate dust layer is less than one volt over the eight hour time period. These trends are due to homogenous chemical composition and the narrow dust size distribution (100 m 40 m). Figure 5.15 (b) shows that charge decay from the JSC-1A surface is much more rapid over the eight hour time period. The JSC-1A charge decay rate also increases the alumina silicateplatechargedecayrate. Figure5.16(b)showsthattheaveragepotentialdifference betweenthealuminaplateandJSC-1Adustlayerconvergesto30%oftheinitialsurface potential after eight hours. These trends are due to a dielectric constant difference and a widerJSC-1Adustsizedistribution(10mto1mm). 129 −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −30 V Alumina Silicate SOLID, 1 min decay time Alumina Silicate SOLID, 2.5 min decay time Alumina Silicate SOLID, 4 min decay time Alumina Silicate SOLID, 6 min decay time Alumina Silicate SOLID, 10 min decay time Alumina Silicate SOLID, 15 min decay time Alumina Silicate SOLID, 20 min decay time Alumina Silicate SOLID, 30 min decay time Alumina Silicate SOLID, 60 min decay time Alumina Silicate SOLID, 120 min decay time Alumina Silicate SOLID, 240 min decay time Alumina Silicate SOLID, 480 min decay time Alumina Silicate DUST, 1 min decay time Alumina Silicate DUST, 2.5 min decay time Alumina Silicate DUST, 4 min decay time Alumina Silicate DUST, 6 min decay time Alumina Silicate DUST, 10 min decay time Alumina Silicate DUST, 15 min decay time Alumina Silicate DUST, 20 min decay time Alumina Silicate DUST, 30 min decay time Alumina Silicate DUST, 60 min decay time Alumina Silicate DUST, 120 min decay time Alumina Silicate DUST, 240 min decay time Alumina Silicate DUST, 480 min decay time (a) Aluminasilicateplateandaluminasilicatedust −60 −40 −20 0 20 40 60 −40 −30 −20 −10 0 10 20 Y, mm Dielectric floating potential, V Measured dielectric potentials, 10 mA beam, CEX plate = −30 V Alumina Silicate, 1 min decay time Alumina Silicate, 2.5 min decay time Alumina Silicate, 4 min decay time Alumina Silicate, 6 min decay time Alumina Silicate, 10 min decay time Alumina Silicate, 15 min decay time Alumina Silicate, 20 min decay time Alumina Silicate, 30 min decay time Alumina Silicate, 60 min decay time Alumina Silicate, 120 min decay time Alumina Silicate, 240 min decay time Alumina Silicate, 480 min decay time JSC −1A, 1 min decay time JSC −1A, 2.5 min decay time JSC −1A, 4 min decay time JSC −1A, 6 min decay time JSC −1A, 10 min decay time JSC −1A, 15 min decay time JSC −1A, 20 min decay time JSC −1A, 30 min decay time JSC −1A, 60 min decay time JSC −1A, 120 min decay time JSC −1A, 240 min decay time JSC −1A, 480 min decay time (b) AluminasilicateplateandJSC-1Adust Figure5.15: Measureddielectricsurfacepotentialsasafunctionoftime 130 0 100 200 300 400 500 −2 0 2 4 6 8 10 time, min delta V Potential difference between Alumina Silicate Solid and Dust vs. time, −30 V charging (a) Aluminasilicateplateandaluminasilicatedust 0 100 200 300 400 500 −2 0 2 4 6 8 10 time, min delta V Potential difference between Alumina Silicate and JSC−1A vs. time, −30 V charging (b) AluminasilicateplateandJSC-1Adust Figure5.16: Averagedielectricsurfacepotentialdifferenceasafunctionoftime 5.8 Alumina Silicate Surface Dust Charge to Single Iso- latedDustChargeComparison The capacitance (between a grain and its sheath boundary) of a single, isolated, spherical dustgrainchargedinplasmawithdustradiusr d << d isgivenbyEqn.5.8. C d = 4" 0 1 ( 1 r d 1 D ) 4" 0 r d (5.8) This form is the same as that of a single isolated sphere in free space, and is widely used to estimate charge accumulated on dust particles [7, 66, 83]. However, this assumption is not valid for a dusty plasma where the inter-dust distance is much less than the Debye length. It is also not valid for a dusty surface system. The capacitance of a dusty surface system (i.e. the lunar surface and experimental dust surfaces herein) is expected to reduce the amount of charge on a given surface dust grain compared to an isolated dust grain at 131 the same potential. To quantify this effect, the chargeQ dexp accumulated on experimental aluminasilicatedustgrainsiscalculatedfromEqn.5.9 Q dexp = r 2 d (5.9) wherer d =50m20m(sphericalgrainsareassumed)andsurfacechargedensity is calculatedfromEqn.5.10 = dQ dA = dC dA ϕ AlSildust = " 0 " rd d layer ϕ AlSildust (5.10) wherealuminasilicatedielectricconstantϵ rd =5.6-6.8,dustlayerthicknessd layer =3.175 mm( 1 8 inch)andmeasuredsurfacepotentialϕ AlSildust isfromTable5.3. The resulting Q d values for both the experimental dust grains and isolated dust grains aregiveninTable5.4. TheisolateddustgrainchargeQ diso iscalculatedfromEqn.5.11 Q diso = C d ϕ AlSildust (5.11) where C d is given in Eqn. 5.8. The Q d values show that individual dust charge of the 100 m 40m alumina silicate dust layer is one to two orders of magnitude lower than the calculated charge for isolated dust grains of the same size at the same potential. This charge reduction effect demonstrates the effect of neighboring dust particles on dusty sur- face system capacitance. Accurate surface potential, dust levitation and transport models mustaccountfordustysurfacelayercapacitance. 5.9 SummaryandConclusions Lunarlowangleofattackplasmaflowinthelunarsimulatorfacilityhasbeenusedtocharge planar alumina silicate and Mykroy/Mycalex R ⃝ , alumina silicate dust surface layers, and 132 Table5.4: Dustysurfacegrainchargetoisolatedgrainchargecomparison ϕ AlSildust ,V Q dexp ,C Q diso ,C 12.9 +(2.21.6)10 15 +(7.22.9)10 14 -7.1 -(1.20.8)10 15 -(4.01.6)10 14 -12.3 -(2.11.5)10 15 -(6.82.7)10 14 -21.5 -(3.62.6)10 15 -(1.20.5)10 13 -31.9 -(5.33.9)10 15 -(1.80.7)10 13 -13.5 -(2.31.7)10 15 -(7.53.0)10 14 JSC-1A dust surface layers. Low energy argon ions (30 eV) and low energy electrons (2 - 40 eV) are the primary incident plasma flux species as a function of CEX plate setting. Primary electron backscattering and electron-induced secondary emission from eachdielectricsurfacetestedisfoundtobenegligibleforthegivenenergyrange. However, significantion-inducedsecondaryelectronemissionyield/ionreflectionisfound. Alumina silicate dust surface potential is recorded to be an order of magnitude higher than alumina silicate planar surface potential when low energy ion flux of the same order as low energy electronfluxisincidentontherespectivetestsurfaces. Ion“grazing”ofsurfacedustgrains alonggrainsurfacecurvatureandforwardionreflectionfromdustgrainsisthoughttodrive this effect. Ion charge deposition is shown to vary as a function of plasma flux, material propertiesandsurfaceproperties. Additionally, charge decay from a JSC-1A dust (10 m to 1 mm grain size) surface hasbeenshowntobemuchmorerapidthanchargedecayfromplanaraluminasilicateand alumina silicate dust (100 m 40 m grain size) surfaces. This trend follows electrical conductivity, as JSC-1A has a lower dielectric constant (3.6 - 4.22) than alumina silicate (5.6 - 6.8). The dust charge of individual dust grains comprising an alumina silicate dust layerhasalsobeencalculated. Whencomparedtotheisolatedgrainchargingcase,surface grains are shown to hold one to two orders of magnitude less charge for the measured experimentalsurfacepotentials. 133 CHAPTER 6: CONCLUSIONS, IMPLICATIONS AND FUTURE RESEARCH As an airless body lacking a global magnetosphere, the Moon is exposed to solar radia- tion and various space plasma environments. A direct consequence is surface charging. Five primary current sources are generally considered: solar wind plasma electrons, solar wind plasma ions, UV induced photo-electrons, particle induced secondary electrons, and backscattered primary electrons. In lunar terminator regions (lunar polar regions and the day-night transition line) differential surface charging is especially complicated due to combinedplasmafluxandsolarillumination. Asthelunarpolarregionsarepreferredlunar baseandsurfacemissionlocations,anaccuratelunarterminatorsurfacechargingmodelis required. Model development is also crucial for dust levitation and transport predictions, as well as relevant to many natural lunar surface processing, such as solar wind saturation ofdustgrainsandlunaratmosphereevolution. Accordingly,thisdissertationexperimentallyinvestigatedthenearsurfaceplasmafield and charging at the lunar terminator, in an effort to contribute to an accurate and compre- hensivelunarterminatorchargingmodel. 6.1 ConclusionsandImplications Thisresearchsetouttoprovideanswersandinsightstothefollowingquestions: How different is the plasma wake expansion process in the vacuum chamber com- paredtoinspace? 134 How different is the charging process of the lunar regolith surface compared to that ofaplanardielectricsurface? How different is the charge of dust grains as part of a regolith surface compared to thatofasingle,isolateddustgraininplasma? All questions were addressed in the context of the preceding four chapters. The con- clusionsreachedandresultingimplicationsaregiveninthissection. 6.1.1 PlasmaWakeExpansion: VacuumChambervs. inSpace A flowing plasma will expand continuously to an angle of 180 degrees (the negative flow direction)unlessiteitherhitsasurfaceortheexpansiondensityhasdecreasedtobelowthe ambient density. When expansion is “terminated” by the ambient density, oncoming ion flowdoesnotturnandexpandasitwouldiftherepellingterminationdidnotexist. Forthe argonionsourceoperatedandcharacterizedinChapter2,expansionterminationissetbya radiallyoutflowinglowenergycharge-exchange(CEX)ionpopulation. Intheexperiment, the CEX ions render the facility effect (vacuum chamber vs. in space) negligible with respect to plume expansion. As the CEX ion production rate of the argon ion source is comparabletotheDeepSpace1NSTARthruster,themeasurementspresentedinChapter2 maybeappliedtopredictplumeplasmapotentialsandexpansionduringin-flightoperation ofsimilarthrusters. Inthesolarwindplasmaflowcasetheneutraldensityismuchlowerthanthethreshold for significant CEX ion flux. Therefore, plasma expansion over lunar objects is controlled by radial flow at the ion acoustic velocity and the lunar surface itself. To experimentally simulate lunar low angle of attack plasma flow, an experimental lunar simulator facility was developed in Chapter 4. The lunar simulator facility is capable of mitigating CEX ion flux such that beam expansion in the direction of the dielectric test surfaces is at the ion 135 acoustic velocity. Experimental expansion validity was confirmed using the analytical and numericalmodelsgiveninChapter4. A consequence of CEX ion flux mitigation was a simultaneous increase in electron kinetic energy flux to the dielectric test surfaces. This effect allowed study of dielectric surface potentials driven by electron energy; however, it did not allow nominal quasi- neutral plasma expansion at the ion acoustic velocity. In other words, when CEX ions were removed and main beam ions flowed out at the ion acoustic velocity, electron flux wasmuchgreaterthanionflux. Therefore,CEXionfluxtothedielectricsurfaceswasalso purposefully increased to simulate solar wind radial expansion with ion and electron flux ofthesameorder. 6.1.2 SurfaceCharging: PlanarDielectricvs. DustDielectric Conducting plates were charged in the main plasma beam as a function of angle of attack andmaterialvariationinChapter3. Thefloatingpotentialsforangleofattackvariationfol- lowedananalyticalprobetheorymodel,andtheAugereffectwasdemonstratedinthemate- rial variation case. These results validated the argon mesothermal plasma source charging capability. However,conductingplatechargingdoesnotwellrepresentlunardielectricsur- face charging. In conductors, surface charge not only remains at the surface, but can be passedtoothermaterialsincontacttominimizerepulsiveforcesbetweenexcesselectrons. Ifaconductoracquiresanexcessofpositivecharge,bothneighboringanddistantelectrons respondtoneutralizechargesuchthatrepulsiveeffectsbetweenelectronsacrossthewhole surface of the object are minimized (and may even draw electrons from other materials in contact). Aschargedoesnotflowbetweenanygivenpointswithinaconductor,conductors are equipotential such that the interior potential is uniform and the surface potential con- stant. Accordingly, local charge build up will not occur and surface charge may even be quicklytransferredaway. 136 Theseconductorpropertiespartiallyexplainwhydielectricmaterialsplacedinthemain plasma beam were charged an order of magnitude more positively than aluminum plates at the same beam setting, angle of attack, and spatial position (see Chapter 3). The other major factor that drove the charging potential difference was that the main plasma beam contained a combination of low energy electron flux, low energy ion flux, high energy ion flux, and high speed neutral flux. Conductors as opposed to dielectrics may respond differentlytoeachfluxsource. Therefore,asystematicstudyofdielectricsurfacepotentials inthelunarsimulatorfacilityplasmaenvironmentwasperformed. Chapter 5 gives the study results and shows significant ion-driven charge deposition differencesbetweenanaluminasilicatedustsurfaceandanaluminasilicateplanarsurface of the same thickness and chemical composition. The dust surface potential is recorded to be over an order of magnitude higher than the planar surface potential. Ion-charge deposi- tion was also shown to vary as a function of planar material type. Argon ions incident on aluminasilicatearebelowthepotentialejectionion-inducedsecondaryemissionthreshold, implyingthatkineticejectionion-inducedsecondaryemission/ionreflectionofthe30eV argonionsis occurring. Aniongrazing mechanismalongdust grainsurfacecurvatureand increased ion absorption due to dust surface topography is proposed. These mechanisms may also explain the high (+100 V with respect to plasma potential) dielectric surface potentialsrecordedatalow(0 )angleofattackinthemainbeaminChapter3. Onthelunarsurfaceiongrazingmayoccurwhenflowingionsareeitheratanextremely low(1 )ornormalincidenceanglewithrespecttothelunarsurface. Theeffectismost likely when high energy protons “graze” the surface in lunar terminator regions or when expandingprotonsstrikethesurfaceatnormalincidenceattheionacousticvelocity. Inthe formercaseionkineticenergyis835eV,whileinthelattercaseionkineticenergyis15 eV(sameorderasCEXargonionsintheexperiment). Inthislattercasetheionabsorption rate may be much less than 1. The kinetic energies may of course vary significantly due to the dynamic lunar plasma environment. Also, ion-induced potential ejection from lunar 137 materialcannotberuledoutwithoutspecificknowledgeoflunarsoilbandgapandelectron affinity. Thereareatleastthreemajorimplicationsofheightenedpositivechargingoflunardust grains. Theimplicationsarelistedbelow: 1. Large positive potentials due to high kinetic energy proton grazing immediately on thedaysideofthelunarterminatorthatdrivedustmotionacrosstheterminator. 2. Less negative charging in shadowed regions if ion-induced secondary emission by normalincidenceprotonsattheionacousticvelocityoccurs. 3. Dustsurfacesadjacenttoplanarsurfacesmaychargemorepositivelythantheplanar surfaces due to dust surface topography increasing ion absorption, driving charged dustmotionbetweentherespectivesurfaces. It is noted that when electron-induced secondary electron emission is > 1, the surface potential must become positive, which is not an absolute condition for ion-induced sec- ondaryelectronemission. 6.1.3 GrainCharge: DustySurfaceGrainsvs. IsolatedGrain The free space capacitance is frequently used in dust charging models. However, a dusty surface layer and reference ground is a capacitance system that depends on the surface potential, thickness, and dielectric relative permittivity, and cannot be accurately modeled using only the isolated grain capacitance. In Chapter 5 the charge of individual alumina silicate dust grains comprising a dusty surface layer has been shown to be one to two ordersofmagnitudelowerthanthechargecalculatedforanisolateddustgrainatthesame potential. This is to say that dust charge and dust charge-to-mass ratio varies not only as a function of size, charge density, and material properties, but also as a function of self-capacitance. Lower dust grain self-capacitance due to neighboring dust interactions 138 consequentlylowersdustgraincharge. Oneofthemajorimplicationsisthatthisdifference industchargingwillleadtosignificantlydifferentpredictionsofelectrostaticlevitationand dusttransportdynamicsonthelunarsurface. Itisnotedthatforafixedcapacitancesystemthicknessandreferenceground,dielectric relativepermittivitymaydirectlyinfluencethesurfacepotential. Forexample,intheplanar alumina silicate vs. dust alumina silicate charging case of Chapter 5, a dielectric relative permittivitydifference(duetosolidmaterialasopposedtograinymaterial)couldtheoreti- callycausetheobservedsurfacepotentialdifferencethatwastakentobefromion-induced secondary electron emission. However, the solution converges as a function of ion flux, which proves that the dielectric relative permittivity difference is slight, and at most alters thesurfacepotentialwithinthemeasurementaccuracy. Further,planaraluminasilicateand dustaluminasilicatesurfacechargedecayrateswereshowntobealmostidentical. Thatbeingsaid,theJSC-1AchargedecayrateinChapter5waspronouncedcompared toaluminasilicatedust. Thisisduetoalowerdielectricconstantthatvariesasafunctionof bulkdensity. The100m40maluminasilicatedustlayeriswellpacked(asevidenced by charge decay very similar to that of the alumina silicate solid surface). The 10m to 1 mm JSC-1A dust layer has a dust distribution that is more variable in both shape and size. This indicates that the unique lunar soil nature (highly elongated grains, irregular shapes, etc.) maydrivesignificantchargedecaydeltasandcorrespondingpotentialdifferencesthat coulddrivedustmotion. 6.2 FuturePlasmaResearch Follow-upplasmaresearchquestionsareoutlinedinthissection. 139 6.2.1 RetardingPotentialAnalyzerIonFocusing AnadequatepowersupplyisrequiredtoconclusivelyconfirmthationfocusingtotheRPA ionretardinggridisdrivingtheobservedanomalousI-Vcurvebehavior. 6.2.2 CEXandFacilityEffects The 4 cm argon electron bombardment ion source is well-characterized in the vacuum chamber,butpropellantionizationefficiencyisverylow(<10%). Amoreefficientsource is required to further model and study the impact of CEX ions and facility effects on ground-basedionsourcedevelopment,testingandcharacterization. 6.2.3 ElectronEnergyDistribution The underlying physics of electron-ion kinetic coupling in the main plasma beam is not wellunderstood. Specificknowledgeoftheelectronenergydistributioninsideandoutside ofthemainplasmabeammayprovideanswers. PreciseLangmuirprobingisrequired. 6.3 FutureUVResearch The photoelectric work function for Apollo sample lunar fines has been measured to be 5 eV [19], and the lunar simulant JSC-1 work function has been deduced at 5.8 eV [67]. Ultraviolet(UV)lightiselectromagneticradiationintherangeof10nmto400nm,which is shorter than visible light, but longer than X-rays. Photoemission created by the solar spectrum is dominated by Lyman alpha emission, which has properties of = 121.6 nm andE =10.2eV[11]. UndertheseconditionsbothlunarsoilsamplesandJSC-1simulant grains have been measured to yield photoelectron current density on the order of 1 to 4.5 A/m 2 [88,2]. Adeuteriumlampcouldbeusedasanexperimentalsourceofhighintensity 140 UVradiation,ascertainlampsarecapableofprovidingemissionbetween115and400nm, withnegligiblevisibleandinfraredemission. UVchargingresultsofJSC-1Acouldthenbeusedtofurtherenhancethedesiredcom- prehensivelunarterminatorregionchargingmodel. 6.4 FutureSurfaceChargingResearch Follow-upsurfacechargingresearchquestionsareoutlinedinthissection. 6.4.1 Kinetic Energy Ion-Induced Secondary Electron Emission and IonAbsorptionRates Ion-induced secondary electron emission yields from both planar dielectrics and dielectric dust surfaces are expected to vary widely as a function of incidence particle species and energy, among numerous other factors. A proton gun or hydrogen source would better simulatesolarwindplasmaflow. Development of a secondary electron collector system would allow direct differentia- tionbetweensecondaryemissionandionabsorption/reflection. 6.4.2 ChargeDecay Temperature, moisture, dust surface layer packing, and other factors affect dielectric elec- trical conductivity, relative permittivity, and charge decay. These aspects could be studied withrespecttosurfacepotentialdifferences. 141 6.4.3 DustLayerGrainCharge Thegrainsizevariationeffectondustlayerindividualgrainchargeshouldbefurtherquan- tified. Various dust size and shape distributions could be obtained from sieving alumina silicatedust,JSC-1A,orsomeotherlunarsimulant. 6.4.4 ChargedDustMotion Onceitispossibletoexperimentallysimulatesufficientelectricfieldstrength(withplasma flux or UV flux or both) to accelerate dust particles, dust motion can be tracked and quan- tified. Factors to consider include adhesion, cohesion, gravity, electrostatic forces and tribo-electriccharging. 6.4.5 LocalTopography Craters, mountains, spacecraft and other simulated lunar objects could be included in the lunarsimulatorfacility. 6.5 FutureNumericalResearch Numericalcodeshouldbevalidatedagainsttheexperimentalresultspresentedherein. Once the code is validated, it can be extended to better represent true solar wind plasma condi- tions,UVconditions,andlunarsurfaceproperties. 142 ReferenceList [1] MM Abbas, D Tankosic, PD Craven, RB Hoover, LA Taylor, JF Spann, and A LeClair. Photoelectric emission measurements on apollo 17 lunar dust grains. In 37th Annual Lunar and Planetary Science Conference, volume Abstract no. 1415, LeagueCity,Texas,2006. [2] MM Abbas, D Tankosic, PD Craven, JF Spann, A LeClair, and EA West. Lunar dust chargingbyphotoelectricemissions. PlanetaryandSpaceScience,55(7-8):953–965, 2007. [3] O.AucielloandD.L.Flamm.PlasmaDiagnostics,volume1ofDischargeParameters and Chemistry. AcademicPress,SanDiego,1989. [4] YassirAzziz. Instrument development and plasma measurements on a 200-watt Hall thruster plume. PhDthesis,MassachusettsInstituteofTechnology,2003. [5] RA Baragiola, M Shi, RA Vidal, and CA Dukes. Fast proton-induced electron emission from rare-gas solids and electrostatic charging effects. Physical Review B,58(19):13212,1998. [6] Ra´ ul A Baragiola. Principles and mechanisms of ion induced electron emission. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms,78(1):223–238,1993. [7] A Barkan, N d’Angelo, and Robert L Merlino. Charging of dust grains in a plasma. Physical reviewletters,73(23):3093–3096,1994. [8] O.Berg,H.Wolf,andJ.Rhee. Lunarsoilmovementregisterdbytheapollo17cosmic dustexperiment. InterplanetaryDust and Zodiacal Light,pages233–237,1976. [9] Om Prakash Narayan Calla and Inder Singh Rathore. Study of complex dielectric properties of lunar simulants and comparison with apollo samples at microwave fre- quencies. Advances in Space Research,2012. [10] GeorgeCarterandJohnSmallwoodColligon. Ionbombardmentofsolids.Heinemann EducationalLondon,1968. 143 [11] JE Colwell, S Batiste, M Hornyi, S Robertson, and S Sture. Lunar surface: Dust dynamicsandregolithmechanics. Reviews of Geophysics,45(RG2006),2007. [12] RogerF.Dashen. Theoryofelectronbackscattering. Phys. Rev.,134:A1025–A1032, May1964. [13] N. Ding. Experimental and Numerical Investigations of Charging Interactions of a DustySurfaceinSpacePlasma. PhDthesis,UniversityofSouthernCalifornia,2012. [14] N. Ding, J. Polansky, and J. Wang. Low cost fabrication of ion optics grids. In 47th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit.AIAA,2011. [15] FB Dunning, RD Rundel, and RF Stebbings. Determination of secondary electron ejection coefficients for rare gas metastable atoms. Review of Scientific Instruments, 46(6):697–701,1975. [16] C.C. Farnell, M.M. Schmidt, R.L. Millot, D.J. Warner, K. Siler-Evans, and J.D. Williams. Multi-axis plasma profiling system for characterization of plasma thruster plumes. 42nd AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, 2006-4818. [17] W. Farrell, T. Stubbs, J. Halekas, M. Killen, G. Delory, Collier M., and Vondrak R. Anticipated electrical environment within permanently shadowed lunar craters. Journalof Geophysical Research,115:doi:10.1029/2009JE003464,2010. [18] WM Farrell, TJ Stubbs, RR Vondrak, GT Delory, and JS Halekas. Complex electric fields near the lunar terminator: The near-surface wake and accelerated dust. Geo- physical ResearchLetters,34(L14201),2007. [19] B. Feuerbacher, M. Anderegg, B. Fitton, LD Laude, RF Willis, and RJL Grard. Pho- toemission from lunar surface fines and the lunar photoelectron sheath. In Proceed- ings of the ThirdLunar Science Conference,volume3,pages2655–2663,1972. [20] JW Freeman and M Ibrahim. Lunar electric fields, surface potential and associated plasmasheaths. Earth,Moon, and Planets,14(1):103–114,1975. [21] D.M. Goebel and I. Katz. Fundamentals of Electric Propulsion: Ion and Hall Thrusters,volume1. Wiley,2008. [22] JS Halekas, SD Bale, DL Mitchell, and RP Lin. Electrons and magnetic fields in the lunarplasmawake. Journal of Geophysical Research,110(A07222),2005. [23] JS Halekas, GT Delory, DA Brain, RP Lin, MO Fillingim, CO Lee, RA Mewaldt, TJ Stubbs, WM Farrell, and MK Hudson. Extreme lunar surface charging during solarenergeticparticleevents. Geophysical Research Letters,34(L02111),2007. 144 [24] JSHalekas,GTDelory,RPLin,TJStubbs,andWMFarrell. Lunarprospectorobser- vations of the electrostatic potential of the lunar surface and its response to incident currents. Journal of GeophysicalResearch,113(A09102),2008. [25] JS Halekas, GT Delory, RP Lin, TJ Stubbs, and WM Farrell. Lunar surface charg- ing during solar energetic particle events: Measurement and prediction. Journal of Geophysical Research,114(A05110),2009. [26] JS Halekas, RP Lin, and DL Mitchell. Large negative lunar surface potentials in sunlightandshadow. Geophysical Research Letters,32(L09102),2005. [27] JS Halekas, DL Mitchell, RP Lin, LL Hood, MH Acu˜ na, and AB Binder. Evidence for negative charging of the lunar surface in shadow. Geophysical research letters, 29(10):77–1,2002. [28] Makoto Hayashi. Bibliography of Electron and Photon Cross Sections with Atoms and Molecules Published in the 20th Century: Water Vapour. National Institute for FusionScience,2003. [29] G Heiken, D Vaniman, and BM French. Lunar sourcebook: A user’s guide to the Moon. CambridgeUniversityPress,1991. [30] E Hill, MJ Mellin, B Deane, Y Liu, and LA Taylor. Apollo sample 70051 and high- andlow-tilunarsoilsimulantsmls-1aandjsc-1a: Implicationsforfuturelunarexplo- ration. Journalof Geophysical Research,112(E02006),2007. [31] MB Hopkins. Langmuir probe measurements in the gaseous electronics conference rfreferencecell. J.Res. Natl Inst. Stand. Technol,100:415,1996. [32] M Hor´ anyi, B Walch, S Robertson, and D Alexander. Electrostatic Charging Proper- tiesofApollo17LunarDust. JournalofGeophysicalResearch,103(E4):8575–8580, 1998. [33] M.Hornyi,S.Robertson,andB.Walch.Electrostaticchargingpropertiesofsimulated lunardust. GeophysicalResearchLetters,22(16):2079–2082,1995. [34] M.Hornyi,B.Walch,S.Robertson,andD.Alexander. Electrostaticchargingproper- ties of apollo 17 lunar dust. Journal of Geophysical Research, 103(E4):8575–8580, 1998. [35] J Kaˇ sparov´ a and M Karlick` y. Kappa distribution and hard x-ray emission of solar flares. Astronomyand Astrophysics,497(3):L13–L16,2009. [36] R.F. Kemp and JM Sellen. Plasma potential measurements by electron emissive probes. Reviewof Scientific Instruments,37(4):455–461,1966. 145 [37] K Kimura, M Hasegawa, Y Fujii, M Suzuki, Y Susuki, and M Mannami. Specular reflectionoffastionsatasinglecrystalsurface. NuclearInstrumentsandMethodsin PhysicsResearchSectionB:BeamInteractionswithMaterialsandAtoms,33(1):358– 364,1988. [38] Lyon B King and Alec D Gallimore. Gridded retarding pressure sensor for ion and neutral particle analysis in flowing plasmas. Review of Scientific Instruments, 68(2):1183–1188,1997. [39] A Kumada, S Okabe, and K. Hidaka. Resolution and Signal Processing Technique of Surface Charge Density Measurement with Electrostatic Probe . Dielectrics and Electrical Insulation, IEEE Transactions on,11(1):122–129,2004. [40] PLee. Dustlevitationonasteroids. Icarus,124(1):181–194,1996. [41] Jung Wook Lim, Sun Jin Yun, and Jin Ho Lee. Electrical properties of aluminum silicate films grown by plasma enhanced atomic layer deposition. Electrochemical and solid-state letters,9(1):F8–F11,2006. [42] RP Lin, DL Mitchell, DW Curtis, KA Anderson, CW Carlson, J McFadden, MHAcuna,LLHood,andABinder. Lunarsurfacemagneticfieldsandtheirinterac- tionwiththesolarwind: Resultsfromlunarprospector. Science,281,1998. [43] RH Manka. Plasma and potential at the lunar surface. Photon and particle interac- tions with surfaces in space,pages347–361,1973. [44] DavidHManzellaandJohnMSankovic. Hallthrusterionbeamcharacterization. In Presentedatthe31stJointPropulsionConferenceandExhibit,SanDiego,CA,10-12 Jul.1995;sponsoredbyAIAA,ASME,SAE,andASEE,volume1,pages10–12,1995. [45] MatWeb. Matweb, your source for materials information http://www.matweb.com, April2013. [46] J.E. McCoy and DR Criswell. Evidence for a high altitude distribution of lunar dust. In Proceedingsof the 5th Lunar Conference,volume3,pages2991–3005,1974. [47] Earl W McDaniel. Collision phenomena in ionized gases. New York: Wiley, 1964, 1, 1964. [48] BradfordMorris. Charge-exchangecollisiondynamicsandionenginegridgeometry optimization. PhDthesis,CaliforniaInstituteofTechnology,2007. [49] T Nitter, O Havnes, and F Melands. Levitation and dynamics of charged dust in the photoelectron sheath above surfaces in space. Journal of Geophysical Research, 103(A4):6605–6620,1998. 146 [50] Tore Nitter, Ove Havnes, and Frank Melandsø. Levitation and Dynamics of Charged Dust in the Photoelectron Sheath above Surfaces in Space. Journal of Geophysical Research,103:6605,April1998. [51] M.A.NorasandA.Pandey. Surfacechargedensitymeasurements. IndustryApplica- tions Magazine,IEEE,16(4):41–47,2010. [52] KW Ogilvie, JT Steinberg, RJ Fitzenreiter, CJ Owen, AJ Lazarus, WM Farrell, and RB Torbert. Observations of the lunar plasma wake from the wind spacecraft on december27,1994. Geophysical Research Letters,23(10):1255–1258,1996. [53] K Ohya and T Ishitani. Target material dependence of secondary electron images inducedbyfocusedionbeams. Surfaceand Coatings Technology,158:8–13,2002. [54] GR Olhoeft, DW Strangway, and AL Frisillo. Lunar sample electrical properties. In Fourth Lunar Science Conference,volume3,pages3133–3149,1973. [55] A.Pandey,J.Kieres,andM.A.Noras. Verificationofnon-contactingsurfaceelectric potential measurement model using contacting electrostatic voltmeter. Journal of Electrostatics,67(2):453–456,2009. [56] DaleLPerry. Handbook of inorganic compounds. CRCPressILlc,2011. [57] A.PoppeandM.Hornyi. Simulationsofthephotoelectronsheathanddustlevitation onthelunarsurface. Journal of Geophysical Research,115(A08106):A08106,2010. [58] A. Poppe, M. Piquette, A. Likhanskii, and M. Horanyi. The effect of surface topog- raphy on the lunar photoelectron sheath and electrostatic dust transport. Icarus, 221:135–146,2012. [59] D.L.ReasonerandW.J.Burke. Characteristicsofthelunarphotoelectronlayerinthe geomagnetictail. Journal of Geophysical Research,77(34):6671–6687,1972. [60] J. Ren, J. Polansky, and J. Wang. Analysis of the anomalous phenomenon in the retardingpotentialanalyzermeasurements. Plasma Science and Technology,2013. [61] JJ Rennilson and DR Criswell. Surveyor observations of lunar horizon-glow. Earth, Moon, and Planets,10:121–142,1974. [62] IRichterov´ a,JPavlu,ZNemecek,andJ ˇ Safr´ ankov´ a. Amodelofsecondaryemission anditsapplicationonthechargingofdustgrains. 2005. [63] I Richterov´ a, J Pavlu, Z Nemecek, and J ˇ Safr´ ankov´ a. Model of secondary emis- sion and its application on the charging of gold dust grains. Physical Review B, 74(23):235430,2006. 147 [64] IvanaRichterov´ a,Jir´ ıPavlu,ZdenekNemecek,JanaSafrankova,andMartinBer´ anek. Secondary emission from glass grains: Comparison of the model and experiment. Plasma Science,IEEE Transactions on,35(2):286–291,2007. [65] AASickafoose,JEColwell,MHornyi,andSRobertson. Experimentalinvestigations onphotoelectricandtriboelectricchargingofdust. JournalofGeophysicalResearch, 106(A5):8343–8356,2001. [66] AA Sickafoose, JE Colwell, M Hornyi, and S Robertson. Experimental levitation of dust grains in a plasma sheath. Journal of Geophysical Research, 107(A11):1408, 2002. [67] Z. Sternovsky, S. Robertson, A. Sickafoose, J. Colwell, and M. Hornyi. Contact chargingoflunarandmartiandustsimulants. J. Geophys. Res,107(E11),2002. [68] TJ Stubbs, JS Halekas, WM Farrell, and RR Vondrak. Lunar surface charging: A globalperspectiveusinglunarprospectordata. In Dust in Planetary Systems,2005. [69] TJ Stubbs, RR Vondrak, and WM Farrell. Impact of dust on lunar exploration. In Dust in Planetary Systems,2005. [70] TJ Stubbs, RR Vondrak, and WM Farrell. A dynamic fountain model for lunar dust. Advances in Space Research,37(1):59–66,2006. [71] B Szapiro and JJ Rocca. Electron emission from glow-discharge cathode materials due to neon and argon ion bombardment. Journal of applied physics, 65(9):3713– 3716,1989. [72] EW Thomas. Particle-impact induced electron ejection from surfaces. International Nuclear Data Committee,1995. [73] Steve Trigwell, James G Captain, Ellen E Arens, Jacqueline W Quinn, and Carlos I Calle. The use of tribocharging in the electrostatic beneficiation of lunar simulant. Industry Applications, IEEE Transactions on,45(3):1060–1067,2009. [74] BE Troy and EJ Maier. Effect of grid transparency and finite collector size on deter- mining, ion temperature and density by the retarding potential analyzer. Journal of Geophysical Research,80(16):2236–2240,1975. [75] TJVink,ARBalkenende,RGFAVerbeek,HAMVanHal,andSTDeZwart. Materi- als with a high secondary-electron yield for use in plasma displays. Applied physics letters,80(12):2216–2218,2002. [76] J.Wang,D.Brinza,D.Young,J.Nordholt,J.Polk,M.Henry,R.Goldstein,J.Hanley, D. Lawrence, and M. Shappirio. Deep space one investigations of ion propulsion plasmaenvironment. Journalof Spacecraft and Rockets,37(5):545–555,2000. 148 [77] J. Wang, D. Brinza, and M. Young. Three-dimensional particle simulation modeling of ion propulsion plasma environment for deep space 1. Journal of Spacecraft and Rockets,38(3):433–440,2001. [78] J. Wang, Y. Cao, R. Kafafy, J. Pierru, and V.K. Decyk. Simulations of ion thruster plume-spacecraft interactions on parallel supercomputer. IEEE transactions on plasma science,34(5):2148–2158,2006. [79] J. Wang, O. Chang, and Y. Cao. Electron–ion coupling in mesothermal plasma beam emission: Full particle pic simulations. Plasma Science, IEEE Transactions on,40(2):230–236,2012. [80] J. Wang and DE Hastings. Ionospheric plasma flow over large high voltage space platforms. ii: The formation and structure of plasma wake. Physics of Fluids B: Plasma Physics,4:1615–1629,1992. [81] JWang,XHe,andYCao. Modelingelectrostaticlevitationofdustparticlesonlunar surface. PlasmaScience,IEEE Transactions on,36(5):2459–2466,2008. [82] J Wang, Xiaoming He, and Yong Cao. Modeling Electrostatic Levitation of Dust ParticlesonLunarSurface. IEEETransactionsonPlasmaScience,36(5):2459–2466, October2008. [83] J.WangandXMHe. Numericalandexperimentalinvestigationsofchargingonlunar surface. In AIAA-2010-0076,2010. [84] J. Wang, P. Wang, M. Belhaj, and J.C. Mateo Velez. Modeling facility effects on secondary electron emission experiment. Plasma Science, IEEE Transactions on, 40(10):2773–2780,2012. [85] X Wang, J Colwell, M Horanyi, and S Robertson. Charge of Dust on Surfaces in Plasma. IEEE Transactions on Plasma Science,35(2Part2):271–279,2007. [86] X. Wang, M. Hornyi, Z. Sternovsky, S. Robertson, and GE Morfill. A laboratory model of the lunar surface potential near boundaries between sunlit and shadowed regions. GeophysicalResearchLetters,34:L16104,2007. [87] John D Williams, Casey C Farnell, Paul B Shoemaker, Jason A Vaughn, and Todd A Schneider. Ground-based simulation of low earth orbit plasma conditions: Plasma generationandcharacterization. In 8th Spacecraft Charging Technology Conference, 2003. [88] RFWillis,M.Anderegg,B.Feuerbacher,andB.Fitton. Photoemissionandsecondary electron emission from lunar surface material. Photon and particle interactions with surfaces in space,pages389–401,1973. 149 [89] SDYou,RDodd,AEdwards,andJWBradley. Astudyoftheplasmaelectronegativ- ity in an argon–oxygen pulsed-dc sputter magnetron. Journal of Physics D: Applied Physics,43(50):505205,2010. [90] XZeng,CHe,HOravec,AWilkinson,JAgui,andVAsnani.Geotechnicalproperties ofjsc-1alunarsoilsimulant. Journalof Aerospace Engineering,23(2),2010. [91] W A Zisman. A New Method of Measuring Contact Potential Differencesin Metals. Reviewof Scientific Instruments,3(7):367,1932. 150

Abstract (if available)

Abstract The objective of this dissertation is to study the near surface plasma environment and surface charging environment at the lunar terminator through experimental, analytical and numerical investigations. Specifically, this dissertation investigates: 1. the plasma wake expansion process, 2. regolith surface charging, and 3. regolith grain charging. ❧ In the first area, the potential and density of the plume emitted from a gridded ion source is investigated. A capability for efficient two-dimensional measurements of the plume profile is developed. The effects of plume expansion, plume charge-exchange plasma, and facility background plasma on plume characteristics are quantified. It is found that the propellant charge-exchange plasma is the primary factor that terminates the plume expansion process, and thus largely controls the magnitude of the plume potential with respect to the ambient. The plasma plume wake expansion process is then controlled to experimentally simulate low angle of attack lunar plasma flow and surface charging. The control method is validated by analytical and numerical models. ❧ In the second area, conductors, planar dielectrics, and dust dielectric surfaces are charged under mesothermal plasma flux. Charging experiments are conducted in both the main plasma beam and the simulated lunar plasma environment. The effects of high energy ion flux, low energy ion flux, and low electron flux are quantified as a function of angle of attack, material properties, and surface properties. It is found that a dielectric dust surface will charge to a significantly more positive surface potential than a planar dielectric surface of the chemical same composition and thickness, when the ion and electron flux is of the same order. Further, this effect is found to be driven by ion-induced secondary electron emission and ion charge deposition mechanisms. The high ion-driven potential of dust surfaces compared to planar surfaces may be contributing to lunar terminator region charged dust motion and transport. ❧ In the third area, dusty surface layer individual grain charge is compared to the charge of a single, isolated dust grain at the same potential. Dusty surface layer potentials are recorded experimentally, and individual grain charge is derived using a capacitance system model. The results are compared to isolated grain charge that is calculated using the free space capacitance model and the experimental surface potential. It is shown that the charge of dust grains comprising a dusty surface layer is one to two orders of magnitude lower than the charge of an isolated dust grain. This demonstrates the packed dust grain capacitive effect on dusty surface layers such as the lunar regolith, and will lead to significantly different predictions of electrostatic levitation and dust transport dynamics on the lunar surface.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Experimental and numerical investigations of charging interactions of a dusty surface in space plasma

PDF

Experimental investigation on dusty surface charging in plasma

PDF

Numerical and experimental investigations of dust-plasma-asteroid interactions

PDF

Grid-based Vlasov method for kinetic plasma simulations

PDF

Kinetic studies of collisionless mesothermal plasma flow dynamics

PDF

The space environment near the Sun and its effects on Parker Solar Probe spacecraft and FIELDS instrumentation

PDF

Interactions of planetary surfaces with space environments and their effects on volatile formation and transport: atomic scale simulations

PDF

An experimental investigation of cathode erosion in high current magnetoplasmadynamic arc discharges

PDF

Systems engineering and mission design of a lunar South Pole rover mission: a novel approach to the multidisciplinary design problem within a spacecraft systems engineering paradigm

PDF

The cathode plasma simulation

PDF

Numerical and experimental investigations of ionic electrospray thruster plume

PDF

Three-dimensional exospheric hydrogen atom distributions obtained from observations of the geocorona in Lyman-alpha

PDF

On the Feynman path into the sun

PDF

Modeling the hot-dense plasma of the solar interior in and out of thermal equilibrium

PDF

Molecular dynamics and virial expansions for equilibrium thermodynamics in the solar interior

PDF

Particle-in-cell simulations of kinetic-scale turbulence in the solar wind

PDF

An investigation into magnetically induced extractor-less electrospray propulsion devices

PDF

Lobster eye optics: a theoretical and computational model of wide field of view X-ray imaging optics

PDF

Investigation of thin film aluminum filter contamination by monomethylhydrazine (MMH)–nitrogen tetroxide (NTO) plumes

PDF

Thermal and deformation analysis of multiphase sulfur concrete extrusion for planetary construction

Asset Metadata

Creator Polansky, John L. (author)

Core Title Laboratory investigations of the near surface plasma field and charging at the lunar terminator

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Astronautical Engineering

Publication Date 07/31/2013

Defense Date 05/24/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag ion beam expansion,ion reflection,lunar dust charging,lunar surface charging,OAI-PMH Harvest,plasma physics

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wang, Joseph (committee chair), Daeppen, Werner (committee member), Dappen, Werner (committee member), Erwin, Daniel A. (committee member), Goodfellow, Keith (committee member), Gruntman, Michael (committee member), Kunc, Joseph (committee member)

Creator Email john.l.polansky@gmail.com,jpolansk@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-309382

Unique identifier UC11294776

Identifier etd-PolanskyJo-1915.pdf (filename),usctheses-c3-309382 (legacy record id)

Legacy Identifier etd-PolanskyJo-1915.pdf

Dmrecord 309382

Document Type Dissertation

Format application/pdf (imt)

Rights Polansky, John L.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA