1

CosmicOrigins(COR)

ProgramTechnologyDevelopment

2018

https://ntrs.nasa.gov/search.jsp?R=20180006911 2019-12-29T13:50:31+00:00Z

2

Program Technology Development QuadCharts

UV/Vis/IR

ShoulehNikzad–“AdvancedFUV/UV/VisiblePhoton‐Countingand Ultralow‐NoiseDetectors” 3

ZoranNinkov–“DevelopmentofDigitalMicromirrorDevicesforUseinFuture SpaceMissions” 4

BabakN.Saif–“Ultra‐StableStructures:DevelopmentandCharacterization UsingSpatialDynamicMetrology” 5

PaulScowen–“ImprovingUltravioletCoatingsandFiltersUsingInnovative MaterialsDepositedbyALD” 6

OswaldSiegmund–“High‐PerformanceSealed‐TubeCross‐StripPhoton‐Counting SensorsforUV‐VisAstrophysicsInstruments” 7

H.PhilipStahl–“AdvancedUVOIRMirrorTechnologyDevelopment forVeryLargeSpaceTelescopes” 8

H.PhilipStahl–“PredictiveThermalControlTechnologytoEnable ThermallyStableTelescopes” 9

JohnVallerga–“DevelopmentofLarge‐Area(100cm2)Photon‐CountingUVDetectors” 10

Far‐IR

CharlesBradford–“Ultra‐SensitiveBolometersforFar‐IRSpectroscopyatthe BackgroundLimit” 10

QingHu–“RaisingtheTechnologyReadinessLevelof4.7‐THzLocalOscillators” 11

AdrianLee–“AdvancingFocalPlaneTRLforLiteBIRDandotherNextGeneration CMBSpaceMissions” 12

JohannesStaguhn–“DevelopmentofaRobust,EfficientProcesstoProduce Scalable,SuperconductingKilopixelFar‐IRDetectorArrays” 13

JamesTuttle–“High‐EfficiencyContinuousCoolingforCryogenicInstruments andSub‐KelvinDetectors” 14

3

4

5

6

7

8

9

10

11

12

13

14

15

16

ProgramTechnologyDevelopmentStatus

UV/Vis/IR

ShoulehNikzad–“AdvancedFUV/UV/VisiblePhoton‐Countingand Ultralow‐NoiseDetectors” 17

ZoranNinkov–“DevelopmentofDigitalMicromirrorDevicesforUseinFuture SpaceMissions” 27

BabakN.Saif–“Ultra‐StableStructures:DevelopmentandCharacterization UsingSpatialDynamicMetrology” 33

PaulScowen–“ImprovingUltravioletCoatingsandFiltersUsingInnovative MaterialsDepositedbyALD” 39

OswaldSiegmund–“High‐PerformanceSealed‐TubeCross‐StripPhoton‐Counting SensorsforUV‐VisAstrophysicsInstruments” 48

H.PhilipStahl‐‐AdvancedUVOIRMirrorTechnologyDevelopment forVeryLargeSpaceTelescopes 59

H.PhilipStahl–“PredictiveThermalControlTechnologytoEnable ThermallyStableTelescopes” 68

JohnVallerga–“DevelopmentofLarge‐Area(100cm2)Photon‐CountingUVDetectors” 78

Far‐IR

CharlesBradford–“Ultra‐SensitiveBolometersforFar‐IRSpectroscopyatthe BackgroundLimit” 93

QingHu–“RaisingtheTechnologyReadinessLevelof4.7‐THzLocalOscillators” 100

AdrianLee–“AdvancingFocalPlaneTRLforLiteBIRDandotherNextGeneration CMBSpaceMissions” 103

JohannesStaguhn–“DevelopmentofaRobust,EfficientProcesstoProduce Scalable,SuperconductingKilopixelFar‐IRDetectorArrays” 114

JamesTuttle–“High‐EfficiencyContinuousCoolingforCryogenicInstruments andSub‐KelvinDetectors” 120

17

AdvancedFUVUV/VisiblePhotonCountingandUltralowNoiseDetectorsPreparedby:ShoulehNikzad(PI;JPL),ChrisMartin(Caltech),DavidSchiminovich(ColumbiaU),andMichaelHoenk(JPL)

SummaryInthis three‐yearSATeffort thatbegan in January2016,wedevelopandadvancethetechnologyreadinesslevel(TRL)ofsolar‐blind,high‐efficiency,photon‐counting,low‐noisesolid‐statedetectorsin ultraviolet (UV) with emphasis on far ultraviolet (far‐UV) with < 200 nm. We combinesuperlatticedoping(SL,Fig.1),integratedsolarblind(SB)filters(Fig.2),antireflection(AR)coatings,and ultra‐low‐noise scientific CMOS (e.g., sCMOS) and large format electron multiplying CCDs(EMCCDs).Wefabricate,characterize,andvalidatethesedetectorsinarelevantspaceenvironment.

Fig.1.LeftandMiddle:Banddiagramofdelta‐dopedandsuperlattice‐dopedsiliconsurfaces.Right:Superlatticedopingisamolecularbeamepitaxy(MBE)processperformedinthe8”‐wafer‐capacitysiliconMBEatJPL.

Al

CH3CH3

CH3

Al

CH3CH3

CH3

Silicon wafer

O O

Al

CH3CH3

Al

CH3CH3

CH4

CH4

Silicon wafer

O O

Al

CH3CH3

Al

CH3CH3

Silicon wafer

O O

Al

Al

O

O O

COO

Plasma

OH OH

Silicon wafer

O O

Al

Al

OOH OH

Precursor Dose Precursor Purge Reactant Purge

Number of Cycles

Thi

ckne

ss

18

Fig.2.TopLeft:Calculatedperformanceofafive‐layermetal‐dielectricfilteronSiorquartzsubstrates,showingin‐bandQE>60%andout‐of‐bandrejectionapproaching10‐4.Thisisanexampleofdetector‐integratedfiltersthatcanrendersilicon“solarblind.”BottomLeft:DiagramofALDprocessshowingitslayer‐by‐layer‐controllednature.Right:JPLALDsystemusedfordepositionofthesemetal‐dielectriclayers.Thedetectorsunderdevelopment in thiseffortofferpotentiallygame‐changingperformanceandcapabilities, directly addressing the Technology development for Cosmic Origins (TCOR)requirementsoutlinedintheSATcallfordetectorswithhighquantumefficiency(QE)in<200nm,large format,photon‐countingcapability, andultra‐lownoise.High‐performanceUVdetectorandcoatingtechnologieswillbeanessentialpartoftheLargeUV/Optical/Infrared(LUVOIR)Surveyor,aleadingcontendertobeoneofNASA’snextastrophysicsflagshipmissions.Additionally,thephoton‐countingcapabilityandultralownoiseineitherUVorvisiblewillhaveahighimpactontheHabitableExoplanetObservatory(HabEx).Becauseofthedramaticefficiencyincreaseinthedetector,flagship‐classsciencewillbepossiblewithsmallersizeapertures.Thiseffort isalso likelytohaveagreatimpactonfutureProbe‐andExplorer‐classmissions.Ourcoatingprocesses,withfilmspreparedbyatomiclayerdeposition(ALD),alsoadvanceUVcoatingsforoptics.

Superlatticedoping(ormultilayerdeltadoping,Fig.1)wasinventedatJPL[1]andproventoachievehighinternalQEonelectronEMCCDsandCMOSarrays[2‐6].Superlattice‐dopedEMCCDsandsuperlattice‐dopedCMOSareatTRL4.Integratedvisible‐rejectionfiltershavebeendemonstratedandprovenonavalanchephotodiodesandextensivelytested[7];theyareatTRL3.Theresultingcombinationofthesesuperlattice‐doped,AR‐coated(SLAR)andSolar‐Blind(SB)EMCCD[8,9]andsCMOSarraysisatTRL3.Furthermore,wewillextendthecapabilityto<200nm.AlldeviceswillbeproducedtomeetourQE,noise,andvisiblerejectionrequirements;willbethermallycycledandcharacterized;andvalidatedinarelevantradiationenvironmenttoraisethemtoTRL4‐5bytheendofthisthree‐yeareffort.

Our team comprisesmemberswith complementary expertise inmaterials, detectors, instrumentbuilding,andobservationalscience.Theteamisuniquelyqualifiedtocarryouttheproposedwork.By forming alliances between technology developers, instrument builders, andmission PrincipalInvestigators(PIs)suchasProf.SchiminovichandProf.Martin(PIsofGALEXaswellassuborbitalmissionssuchasFIREBALL),ourteamhasanaturalpathforTRLadvancementandflightinsertion.

BackgroundThe2010DecadalSurvey,“NewWorlds,NewHorizonsinAstronomyandAstrophysics”(NWNH)[10]recommends, as a priority, path‐finding work towards a 4m+ UV/Optical flagship mission as asuccessortotheHST.GreatemphasisonExplorermissionsisalsoanticipatedinthisdecade.

Recently,mid‐decadalstudieshavebeenongoingthatbuildontheworkoftheNWNH.Thesestudieshaveconsistentlysetforthtechnologydevelopmentgoalsaimedatenablingafuturelarge‐aperture,UVOIR flagshipmissionasasuccessor to theHubbleSpaceTelescope,and increasing thescientificreachofsmallermissionsinthisdecade.TheCosmicOriginsProgramAnalysisGroup(COPAG)isnowevaluatingandrecommendingtechnologyinvestmentstowardsthesegoalsthroughScienceInterestGroups(SIGs). Inaddition,anewTechnologyInterestGroup(TIG)isreviewingtechnologygaps. Inthese scientific focus areas, single‐photon‐counting or ultra‐low‐noise detectors are a priority.Furthermore,theserecommendationssetasagoal,very‐large‐format(100MegapixeltoGigapixel),high‐QE, UV‐sensitive detectors. These recommendations reflect the new understanding of thescientificopportunitiesenabledbytechnologicalbreakthroughsinlarge‐scaledetectorfabrication.

OurobjectivesaretiedtotheneedsofandrecommendationsforfuturemissionsmentionedinNWNH,guidedbymid‐decadalstudiesbeingcarriedoutbyCOPAGandtheAssociationofUniversitiesforResearchinAstronomy(AURA),andupdatedpertheneedsofflagshipconceptssuchasLUVOIRandHabEx.

19

Frontierastrophysicalinvestigationsarenecessarilyconductedatthelimitsofresolution,etendue,andsensitivity. A future 10‐m UV/Optical telescopemissionwill require significant detector advancesbeyondHST,GALEX,andFUSEdetectortechnologies,particularlyinQE,spectralresponsivityintheUV,resolution,andpixelcount.Ourprimaryperformancemetric,detectorQEintheUV,representsadramaticincrease(5‐10x)overpreviousmissions.DramaticallyincreasingdetectorefficiencycouldallowExplorer‐classorProbe‐classmissionstoperformflagship‐missionscience.

Asolid‐statedetectorwithhighefficiencyandphotoncountingoffersscalabilityandreliabilitythatarenecessary and attractive features for reliable, high‐performance, and cost‐effective instruments.LUVOIR’srequirementsaredirectlyapplicabletotheobjectivesof thiseffort.Additionally,detectorsdevelopedunderthisSATbutoptimizedforvisiblelightwouldalsohaveahighimpactonHabEx.

ObjectivesandMilestonesTable1showstheproject’smilestonesandschedule.

Milestone Year1 Year2 Year3

Q1 Q2 Q3 Q4

Demonstratesolar‐blind,SLAREMCCD

Procurewafersofstandardlarger‐formatEMCCDs

Thin,bond,andSL‐dopewafers

IncorporateadvancedALDfilters(200‐240nm)

Demonstratesolar‐blind,SLAR,low‐noiseCMOS

Selectlow‐noiseCMOSdesign

Procurewafersoflow‐noiseCMOS(e.g.,sCMOS)

Thin,bond,andSL‐dopesCMOS

IncorporateadvancedALDfilters(200‐240nm)

Extendvisible‐blindfiltertoFUV

Extendmultilayerdesigntocenterat140or150nm

IntegratewithSLAREMCCD

IntegratewithSLARsCMOS

Validationandenvironmentaltesting

TestnoiseandQEwithtemperatureandillumination.Lifetimetesting.

Table1.Projectmilestonesandschedule.

ProgressandAccomplishmentsProgressinallaspectsofdevelopmentmilestonesarediscussedhere.Theseincludefabricatingandproducing delta/superlattice‐doped EMCCDs, superlattice‐doped CMOS, out‐of‐band‐rejectionintegratedfilters,environmentaltestingforTRLadvancement,andsuborbitalandorbitalflight‐pathplanningandprogress isdiscussed. In addition to technicalprogressdetailedbelow,publication,

20

programmaticactivities,professionalsocietyactivities,andprofessionalhonorsandawardsarealsoincludedinthisreport.

EMCCDwaferswere procured and received. Thesewaferswere used to produce the superlattice‐doped,coatedanduncoatedEMCCDsusingtheprocessoutlined inFig.3.This includedbondingto“handle”or “support”wafer sprior to thinning—final thicknessdecisionsdependson theextentofdepletionregionandpassivationcapabilityfordark‐noisereduction.Inthefirstbatchofwafers,weimplementedononewafersuperlatticedopingandsingle‐layerAl2O3coatingdesignedfortheFaintIntergalacticRedshiftedEmissionBalloon(FIREBall)balloonexperiment,fundedbytheAstrophysicsResearchandAnalysis (APRA)Program.Severaldevicesweresubsequentlycoatedwithmultilayerdielectrics after they were diced. All devices were packaged, and detailed characterization wasperformedtodetermineoptimalperformance.TestingandcharacterizationswereperformedbothatJPL and Caltech (in photon‐counting mode) and as a team, cross‐checking results and verifyingperformancemeasuredpreviouslyonsuperlattice‐dopeddevicesate2v.Aspartofvalidation,deviceswere taken to Palomar for on‐sky data validation using the Cosmic Web Imager (CWI). Photon‐counting,low‐noise,andQEperformancewereverified.

Fig.3.End‐to‐endpost‐fabricationprocessing shown forEMCCDdevicewafer.Resultsof someof thesewerepublished[12,13].

Uponreceiptoffullfundingfor2017,discussionsresumedwithAndor,Sarnoff,andpotentiallye2vregardingacquiringwaferscontainingsCMOSorotherlow‐noiseCMOSdevices.Wehadcontactedtopgroupsinultra‐low‐noiseCMOSdevelopmentsuchasFairchildforsCMOS,Celeaste,e2V,Sarnoff,AMS‐CMOSIS, and Rutherford Appleton Laboratory (RAL) regarding wafer procurement andcollaboration.NDAwereexecutedwithSarnoff,Andor,AMS‐CMOSIS,andothersaspotentialCMOSsuppliers.Wafersandwaferfractionsfromtwovendorshavebeenreceived.Becausesomeofthese

21

waferswerenotfabricatedforbackillumination(BI),extraprocessingstepsandadditionalNDAswithfoundrieswererequired.Theseactivitiesarewellunderway.

Metal dielectric filters for enhancement in the 140‐150 nmwavelength rangewere designed. Incollaborationwithendusers,the150‐175nmbandwasidentifiedaspotentiallyuseful.Thelatterdesignwas implemented in a siliconwafer. The designed layerswere also implemented in non‐functionalsuperlattice‐dopedEMCCDstocheckforandaddressmechanicalissues.Thesedesignsarereadyforintegrationintofunctionalsuperlattice‐dopedEMCCDs.Prototype devices have been fabricatedwith five‐layer Al/AlF3model dielectric function (MDF).Reflectance measurements show good agreement with the model (Fig. 4), as do relative QEmeasurements(Fig.5).

Fig.4.Left:Prototypedevicewithanintegrated,five‐layer,Al/AlF3MDF.Right:Reflectancemeasurementfromthesampleontheleftshowinggoodagreementwiththemodel.

Fig.5.Left:Prototypedevicewithanintegrated,five‐layer,Al/AlF3MDF.Right:RelativeQEmeasurementsshowout‐of‐bandrejectionthatagreeswiththemodel.

Our TRL advancement plan uses environmental testing, and suborbital and small‐sat flights.Superlattice‐dopedandAR‐coateddevicesareduetoflyonFIREBall‐2(Fig.6)inSeptember2018,and superlattice‐doped devices with out‐of‐band‐rejection integrated filters are due to fly onSHIELDS(Fig.7),aheliophysicssoundingrocketinthefallof2019.

100 150 200 250 300 350 4000.0

0.2

0.4

0.6

0.8

1.0

Relative QE Measured Reflectance Model R Model T

Ref

lect

ance

/Tra

nsm

ittan

ce

Wavelength (nm)

22

Fig.6.Left:FIREBallundergoingon‐skycalibrationinFt.Sumner,NM.Left:FIREBalldetectorwiththree‐layerdielectricdeliveringQE>60%.

Fig. 7. SHIELDS rocket atUniversity ofArizona. The superlattice‐doped devicewith integrated out‐of‐band‐rejectionfilterisunderdevelopmentandisplannedtoflyonthisrocketinthefallof2019.

SHIELDSisafundedsoundingrocketledbyProf.WalterHarrisattheUniversityofArizona.ProfessorHarrisinpartoftheLUVOIRScienceandTechnologyDefinitionTeam(STDT).AsuperlatticeEMCCDwith detector‐integrated filters optimized for 120‐140 nm (10‐3 out‐of‐band rejection) is underfabricationandoptimization.

Environmental testing ispartof thiseffort.Thermalcyclinghasbeenanongoingpartofdevicecharacterizationasdevicesarecycledfromroomtemperaturetocold(downto‐100°C).Thisyear,wealsofocusedonradiationtesting.WechosethetestprotocoloftheWide‐FieldInfraredSurveyTelescope (WFIRST). In early June 2017, one superlattice‐doped EMCCDwas taken alongwithseveralWFIRSTEMCCDstoaLomaLindalocalhospital’sradiationfacilityforroom‐temperaturetesting. Several deviceswere selected and are under characterization for FIREBall‐2, anAPRA‐funded balloon‐borne experiment to study emissions from intergalactic medium and circum‐galacticmedium.FIREBall‐2dependsonhigh‐efficiencyphoton‐countingdetectors.Asuperlattice‐dopeddevicewasdeliveredinthepreviousreportingperiod,custom‐coatedwiththreedielectriclayersoptimizedforFIREBall’snarrowatmosphericUVwindowcenteredon205nm.ThedetectorwastestedatProf.Martin’sCaltechlab,atPalomarforphotoncountingcapability,andatsystemlevelbytheFIREBall‐2teamusingtheentireFIREBall‐2spectrographsystem.Thedetectorwastested to show photon‐counting, and nearly 60% QE (5× better than FIREBall’s MCP detectorresponse). A device with three‐layer coatings was selected for flight for good cosmetics andexcellentperformance.TheFIREBallteamwasreadytoflyinSeptember2017butunfortunatelyduetoweatherandsafetyconcernsunrelatedtoFIREBall,onlyoneballoonflewthatmonth.This

23

year,theFIREBallteambeganmeasurementsandcalibrationsinpreparationforaSeptember2018flight.Accordingtotheprojectleadandprojectscientist, thedetectorhasbeenunchangedovernearlyayearwithnoadjustmentnecessary.

Additionally, Star‐Planet ActivityResearch CubeSat (SPARCS) has baselined two2D‐dopedCCDs.SPARCSisa6Ucube‐satdesignedtomonitorM‐typestarsintwophotometricbandsinthenear‐UVand far‐UV (S‐NUV, 260‐300 nm; S‐FUV, 150‐170 nm), and is planned for 2021 flight. The FUVdetectorwill also have a detector‐integrated filter. SPARCS has baselined a dichroic design thatallowsforsimultaneousS‐NUVandS‐FUVobservation.SPARCSwouldadvance2D‐dopeddetectorsanddetector‐integratedout‐of‐band‐rejectionfiltertechnologiesforpotentialapplicationinfuturemissionconceptssuchasLUVOIRandHabEx.

PathForwardWe will complete further wafer processing and produce several 2‐megapixel arrays in part inpreparation for radiation testing.Wewillanalyzedata fromroom‐temperature,unbiasedradiationtesting of coated and uncoated superlattice‐doped 2‐megapixel EMCCDs. Informed by the room‐temperature results, wewill optimize our low‐temperature biased radiation testing plan.Wewillcomplete these tests in the fall of2018.Wewill evaluate andcharacterize theQE,darknoise, anduniformityofthesedevicesbeforeandafterradiationfollowingWFIRSTprotocol.Inparallel,wewillimplementanotherdesignofmetal‐dielectricfilmsintosuperlatticeEMCCDs.WewillcontinueworkingwiththeSHIELDSteam(WaltHarris;PI,U.ofArizona).Furtheroptimization(designandprocessing)of out‐of‐band‐rejection filters might be necessary based on lab feedback and for testing by theSHIELDSteam.InordertogetclosertotheabsolutevalueoftheQE,weareverifyingthemeasuredQEin other systems. The Caltech group also characterizes devices in photon‐countingmode.WewillresumeourdiscussionswithCMOSvendorsandworkonprocuringwafersfromourselectedvendors.

24

References[1] “GrowthofaDelta‐DopedSiliconLayerbyMolecular‐BeamEpitaxyonaCharge‐CoupledDevice

for Reflection‐Limited Ultraviolet Quantum Efficiency,” M.E. Hoenk, P.J. Grunthaner, F.J.Grunthaner,M.Fattahi,H‐F.TsengandR.W.Terhune,Appl.Phys.Lett.,61,1084(1992).

[2] “Delta‐dopedCCDs:HighQEwithLong‐termStabilityatUVandVisibleWavelengths,”S.Nikzad,M.E.Hoenk,P.J.Grunthaner,R.W.Terhune,F.J.Grunthaner,R.Winzenread,M.Fattahi,andH‐F.Tseng,Proc.ofSPIE,2198,907(1994).

[3] “Near‐100% Quantum Efficiency of Delta Doped Large‐Format UV‐NIR Silicon Imagers,” J.Blacksberg,S.Nikzad,M.E.Hoenk,S.E.Holland,andW.Kolbe,IEEETrans.OnElectronDevices553402,(2008).

[4] “Delta‐dopedback‐illuminatedCMOSimagingarrays:Progressandprospects,”M.E.Hoenk,T.J.Jones,M.R.Dickie,F.Greer,T.J.Cunningham,E.R.BlazejewskiandShoulehNikzad,ProceedingsoftheSPIE7419074190‐74115(2009).

[5] “SiliconDetectorArrayswithAbsoluteQuantumEfficiencyover50%intheFarUltravioletforSinglePhotonCountingApplications,”S.Nikzad,M.E.Hoenk,F.Greer,E.Hamden,J.Blacksberg,B.Jacquot,S.Monacos,ChrisMartin,D.Schiminovich,P.Morrissey,AppliedOptics,51,365(2012)

[6] “Delta‐dopedCCDsAsStable,HighSensitivity,HighResolutionUVImagingArrays,”S.Nikzad,M.E.Hoenk,P.J.Grunthaner,R.W.Terhune,R.Winzenread,M.Fattahi,H‐F.Tseng,andF.J.Grunthaner,Proc.ofSPIE,2217,355(1994).

[7] “Antireflection Coatings Designs for use in UV Imagers,” E. Hamden, F. Greer, M.E. Hoenk, J.Blacksberg,T.J.Jones,M.Dickie,B.Jacquot,S.Monacos,C.Martin,P.Morrissey,D.Schiminovich,S.Nikzad,AppliedOptics,50,4180‐4188(2011).

[8] “LLLCCD—LowLightLevelImagingwithouttheneedforanintensifier,”P.Jarram,P.Pool,R.Bell,D.Burt,S.Bowring,S.Spencer,Proc.SPIE4306,178(2001).

[9] “Impactron –ANew Solid State Image Intensifier,” J. Hynecek, IEEE Transaction on ElectronDevices,48No.10,2238‐2241(2001).

[10] “NewWorlds,NewHorizonsinAstronomyandAstrophysics,”Blandfordetal.,NationalAcademyofSciences,(2010).

[11] “Parametric Cost Models for Space Telescopes”, H.P. Stahl, T. Henrichs, C. Dollinger,InternationalConferenceonSpaceOptics,Rhodes,Greece,October2010.

[12] “UVphoton‐countingCCDdetectorsthatenablethenextgenerationofUVspectroscopymissions:ARcoatingsthatcanachieve80‐90%QE,”E.Hamden,D.Schiminovich,S.Nikzad,andC.Martin,ProceedingsoftheSPIE,8453,HighEnergy,Optical,andInfraredDetectorsforAstronomyV,845309(2012).

[13] “DeltadopedElectronMultipliedCCDwithAbsoluteQuantumEfficiencyover50%intheneartofarUltravioletRangeforSinglePhotonCountingApplications”,ShoulehNikzad,MichaelE.Hoenk,Frank Greer, Blake Jacquot, SteveMonacos, Todd Jones, J. Blacksberg, ErikaHamden, DavidSchiminovich,ChrisMartin,andPatrickMorrissey,AppliedOptics,51,365(2012).

25

Publications,Presentations,andProfessionalSocietyLeadershipRolesNikzadwashonoredbyelectionandsubsequentinductionintotheNationalAcademyofInventorsasafellow.

ShoulehNikzadintheinductionceremonyattheNationalAcademyofInventorsflankedbythePresidentofNAIandtheCommissionerforPatentsfortheU.S.PatentandTrademarkOffice

Twopaperswerepresentedasoralpresentationatthe2017InternationalImageSensorWorkshop(2017 IISW)byS.NikzadandM.E.Hoenk.Thesewerepublishedaspeer‐reviewedpapers in theproceedingsofIISW.S.NikzadalsoservedasthetechnicalprogramcommitteememberonIISW.

TwopaperswerepresentedattheScientificDetectorWorkshop(SDW),BaltimoreMD,inSeptember2017byS.NikzadandM.E.Hoenk.S.Nikzadservedasapanelmemberonthedetectorneedsforthefutureastrophysicsflagshipmissions,aswellasontheScientificOrganizingCommittee.

S. Nikzad served as conference co‐chair and proceedings co‐editor for the SPIE AstronomicalTelescopesandInstrumentations:GammaRaytoUltravioletmeeting.

M.E.HoenkservedontheprogramcommitteeoftheSPIEHighEnergy,OpticalandInfraredDetectorsforAstronomyVIIImeeting.

Sevenpaperswerepresentedorco‐authoredby the teamat theSPIEAstronomicalTelescopes+Instrumentation,Austin,Texas,June2018AstronomicalTelescopesandInstrumentations:GammaRaytoUltravioletandHighEnergy,OpticalandInfraredDetectorsforAstronomyVIII.

Photographs of participants and presenters at the SPIE conference co chaired by Nikzad on AstronomicalTelescopes+Instrumentations:GammaRaytoUltraviolet

1. S.Nikzad,A.D.Jewell,M.E.Hoenk,T.J.Jones,J.Hennessy,T.Goodsall,A.G.Carver,C.Shapiro,S.R.Cheng,E.T.Hamden,G.Kyne,D.C.Martin,D.Schiminovich,P.Scowen,K.France,S.McCandliss,andR.E.Lupu,“High‐efficiencyUV/optical/NIRdetectorsforlargeaperturetelescopesandUVexplorermissions:

26

developmentofandfieldobservationswithdelta‐dopedarrays,"JATIS3(3),036002(2017),doi:10.1117/1.JATIS.3.3.036002.

2. S.Nikzadetal.,“OverviewofHighEfficiencyandLowNoiseSolidStateDetectorsforFutureMissionsincludingtheFlagshipConcepts:TheHabitableExoplanetImagingMission,”SDW,BaltimoreMD(September2017)

3. M.E.Hoenk et al., “EfficiencyandStabilityof2D‐dopedSiliconDetectors,”SDW,BaltimoreMD(September2017)

4. S.Nikzad,J.Hennessy,M.E.Hoenk,A.Kiessling,M.R.Bolcar,D.F.Figer,S.Martin,andR.Morgan,“Solid state detectors for the Habitable Exoplanet imaging mission (HabEx) and the largeUV/optical/infrared (LUVOIR) surveyor mission concepts,” SPIE Astronomical Telescopes +Instrumentation,Austin,Texas(June2018)

5. Y.Yatsu,T.Ozawa,H.Mamiya,N.Kawai,Y.Kikuya,M.Matsushita,S.Matunaga,P.Bilgi,S.Nikzad,S.R.Kulkarni,N.Tominaga,M.Tanaka,T.Morokuma,N.Takeyama,A.Enokuchi,andK.Sasaki,“Conceptualdesignofawide‐fieldnearUVtransientsurveyina6UCubeSat,”SPIEAstronomicalTelescopes+Instrumentation,Austin,Texas(June2018)

6. A.D.Jewell,J.Hennessy,T.Jones,S.Cheng,A.Carver,D.Ardila,E.Shkolnik,M.E.Hoenk,andS.Nikzad,“Ultravioletdetectors forastrophysicsmissions:a case studywith SPARCS,” SPIE AstronomicalTelescopes+Instrumentation,Austin,Texas(June2018)

7. J. Hennessy, C.S.Moore, K. Balasubramanian, A.D. Jewell, K. France, and S. Nikzad, “Ultrathinprotective coatings by atomic layer engineering for far ultraviolet aluminum mirrors,” SPIEAstronomicalTelescopes+Instrumentation,Austin,Texas(June2018)

8. P.A.Scowen,E.Shkolnik,D.Ardila,J.Bowman,D.Jacobs,A.Jewell,M.Beasley,T.Berman,V.Gorjian,J.Llama,V.Meadows,S.Nikzad,M.Swain,andR.Zellem,“Monitoringthehigh‐energyradiationenvironment of exoplanets around low‐mass stars with SPARCS (star‐planet activity researchCubeSat),”SPIEAstronomicalTelescopes+Instrumentation,Austin,Texas(June2018)

9. R.M.Morgan,K.Warfield,H.P.Stahl,B.Mennesson,S.Nikzad,J.Nissen,K.Balasubramanian,S. Shaklan, J. Krist, D. Mawet, K. Stapelfeldt, and S. Warwick, “Technologymaturity for thehabitable‐zone exoplanet imagingmission (HabEx) concept,” SPIE Astronomical Telescopes +Instrumentation,Austin,Texas(June2018)

10. M.E.Hoenk,A.D.Jewell,S.Nikzad,Q.Looker,B.D.Tierney,andM.O.Sanchez,“Superlattice‐dopeddetectorswithhighstablequantumefficiencyinhighradiationenvironments,”SPIEAstronomicalTelescopes+Instrumentation,Austin,Texas(June2018)

Foradditionalinformation,contactShoulehNikzad:Shouleh.Nikzad@jpl.nasa.gov

27

DevelopmentofDMDArraysforUseinFutureSpaceMissionsPrepared by: Zoran Ninkov (PI; Rochester Institute of Technology, RIT); Sally Heap andManuel Quijada (NASA/GSFC); Massimo Robberto (STScI); and Alan Raisanen, DmitryVorobiev,andAntonTravinsky(RIT)

SummaryThisNASAStrategicAstrophysicsTechnology(SAT)projectbeganinMay2014.Theprojectseekstoinvestigatethefeasibilityofusingadigitalmicro‐mirrordevice(DMD)astheslitmaskforamulti‐objectspectrograph(MOS)systemforavarietyoffutureNASAspacemissions.Inparticular,weareinvestigatinganumberofkeyoperatingparametersforTexasInstruments(TI)commercial‐off‐the‐shelfDMDsincluding:replacingtheborosilicatewindowwithwindowstransmissiveatultraviolet(UV) and infrared (IR) wavelengths, tolerance to particle‐radiation effects, ability to survive thevibrational conditions of launch, and non‐specular scattering properties of the DMDs. The teamincludes Sally Heap at NASA/GSFC, who provides us with insight into the connection betweenastronomicalmeasurementrequirementsandourlaboratorytesting;MassimoRobbertoattheSpaceTelescopeScienceInstitute(STScI),whoprovidestestdesignguidanceandhaspreviousexperiencewith proposed DMD used by the European Space Agency (ESA); Manuel Quijada at GSFC, whoprovidesconsiderableopticsexperienceandtheuseoftheCarey5000SpectrometeratGSFC;andAlanRaisanenatRIT,whoprovidesthenecessarymicrosystemsexperiencetoallowforreplacementofDMDwindowsintheRITcleanroom.Additionally,wehaveworkedwithJonnyPellish(GSFCCode561)toconductheavy‐ionradiationtestingofDMDsatTexasA&MandTimSchwartz(GSFCCode549)forvibrationandshocktesting.Theprojecthasmadesignificantprogressthisyear,includinganimprovedmeasurementoflightscatteringfromtheTIDMD,asecondroundofheavy‐iontestingontheDMDsattheTexasA&MUniversityCyclotron,vibrationalandshocktestingofDMDsinre‐windowedandstandardpackages,extensivelow‐temperaturetestingwithourcollaboratorsatJohnsHopkinsUniversity(JHU),andre‐coatingaDMDwithhigh‐purityaluminumatGSFC.

BackgroundOurultimateobjectiveistoaddresstwokeyquestionsofNASA’sCosmicOrigins(COR)Program:

1. Howdidgalaxiesevolvefromtheveryfirstsystemstotheellipticalandspiraltypesweobservetoday?2. Howdidsuper‐massiveblackholesaffectthelivesofgalaxiesinwhichtheyreside,andviceversa?

Ground‐based telescopes and the Hubble Space Telescope (HST) have shown us that the Hubblesequence of elliptical and spiral galaxies was in place by redshift z = 1. However, what physicalprocessesdrovez>1galaxiestojointheHubblesequence?Tounderstandgalaxyevolution,weneedtocarryoutalargespectroscopicsurveyoftheskywithaparticularfocusongalaxiesatredshiftsofz = 1~2. Experience with the Sloan Digital Sky Survey [1] indicates that several hundred thousandgalaxiesneedbeobservedinordertodistinguishamongthemanypossibledriversofgalaxyevolution(e.g.,accretion,mergers,starformation,stellarevolutionandfeedback,growthofblackholes,etc.).

AlargespectroscopicsurveyrequiresaMOSabletorecordthespectraofhundredsofgalaxiesinasingleexposure.TheMOSmusthaveadjustableslitstoeliminateconfusionwithnearbysourcesandtoblockoutunwantedzodiacalbackground,whichwouldotherwiseswampthelightfromthesefaintgalaxies.TheMOSshouldhaveaccesstothefarUV(1200 ‐ 2000Å)radiationemittedbyaz~1galaxybecausethisspectralregionhasarichsetofdiagnosticsofstars,gas,anddustinthegalaxy.Accesstotheblue‐redspectralregions(2000 – 8000Å)isalsoessentialfordeterminingthepreciseredshift

28

ofagalaxy, itsstellarmass,and itselementalabundances;and forcharacterizingdustextinction.Becausethelightfromaz~1galaxyisredshiftedbeforereachingus,alargespectroscopicsurveyshouldbesensitiveoverthespectralinterval2000 ‐ 16000Å.

TheProblem:NoexistingMOShassuchawidespectralrange,letaloneaccesstotheUV.TI’sDMDwouldmake an excellent slit selector for a spectrograph if itwere sensitive in theUV.However,commercialDMDwindowsblockUVlight.

ScientificImpact:AUV‐transmittingDMDwindowenablesabreakthroughinobservationalpowersufficient to address two key COR science issues. No other telescope, ground‐ or space‐based,present or planned, can accomplish this investigation, because it can’t observe all the spectraldiagnosticsfromLy(~1200Å)toH+[NII](~6600Å)inthesamehigh‐redshiftgalaxy.

OurprojectintendstoinvestigatetheapplicabilityofDMDstothisandotherspace‐basedapplicationsby testing the radiation hardness and light‐scattering properties of these devices. In addition, ourprojectwilllookatapproachestoreplacingthecommerciallyprovidedwindowsonDMDs.

TheSolution:WethereforeproposetooptimizetheperformanceofDMDsfortheUVregion.Thisrequires replacing the DMD window with a UV‐transmitting window (> 2000 Å) with an anti‐reflectioncoatingoneachside,optimizedfortheUV,optical,andIR.Becausethetargetgalaxiesareataredshiftofz~1,theobservedspectrumofagalaxyover0.2 – 1.6μmrecordsthelightemittedbythe galaxy in the spectral range 0.1 – 0.8 μm. This wavelength region contains virtually all theimportantspectraldiagnosticsofstars,gas,anddustinthegalaxy.

ObjectivesandMilestonesTable1providesthemajormilestonesofthisproject.Ourprojectstartedmoreslowlythanexpected,principallybecauseittooklongerthanexpectedtoidentifyvendorsfortheneededcomponents(e.g.,windows,DMDs)andservices(L‐1Standards&Technology).TherewasalsoalongdelayingettingpurchaseordersthroughtheGSFCsystem.Finally,althoughwereceivedaquotefromaUS‐baseddistributortoacquireTICinemaDMDswhenwesubmittedtheproposal,itturnedouttheycouldnotresellthoseintheUS.TheonlywaytogettheseDMDswasthroughTI’sEuropeandistributor,atahigher‐than‐budgetedcost.STScIprovidedtheadditionalfunds,buttherewasadelayinplacingthesub‐contractandthustheorder.

Milestone Vendor / Work Location Dates Comments

Proton-testing of DMD and data analysis

LBNL 88” Cyclotron Completed 2014 Results published

Receipt of MgF2 and Heat Exchanger Method (HEM) Sapphire windows

Photonics Solutions Group, Blue Ridge Optics, GT Crystal Systems

Received Sep 2015 After many delays from vendors, we received all of the windows; all windows met our specifications

Replacement of TI DMDs window

L-1 Standards & Technology

Received Dec 2015 Devices all accepted

Receipt of Cinema DMDs + drive electronics

VISITECH, Germany Received Jan 2016 Final quote for these items was higher than budgeted; supplemental funding allowed order to proceed

Replacement of TI DMD windows with custom windows

Semiconductor & Microsystem Fabrication Laboratory, RIT

Aug 2015 & Jan 2016 DMDs with Sapphire and MgF2 windows have been fabricated as have devices with Kapton and Mica windows for heavy-ion testing

29

Vibration and shock testing

NASA GSFC Code 549 Testing completed May 2016; Analysis completed Feb 2017

Results accepted by Journal of Astronomical Telescopes, Instruments, and Systems

Measurement of light scattering from eXtended Graphics Array (XGA) DMD

RIT, Carey 5000 at GSFC Code 551

May 2016 Mar 2017 High signal-to-noise ratio (SNR) measurements of standard DMDs; Final measurements of bare DMDs with original coating and new Al coating

Heavy-ion testing of DMD and data analysis

Texas A&M Radiation Effects Facility

Tests Aug 2015 & Apr 2016; Analysis completed Feb 2017

On-orbit event rates estimated; Results accepted by Journal of Astronomical Telescopes, Instruments, and Systems

Recoating of DMDs with high-reflectivity aluminum

NASA GSFC Code 551 Feb 2017 Coated a section of a functioning XGA DMD with high reflectivity aluminum; DMD remained operational

Low-temperature testing of DMDs

Instrument Development Group at JHU

May 2017 Supplied DMDs with re-windowed & original packages for testing at 77 K

Gamma-ray testing NASA GSFC June 2018 Tests of multiple DMD devices

Extended-time low-temperature testing

JHU/APL November 2017 Hold DMD patterns for typical astronomical observation times at low temperature

Optical modeling module

RIT June 2018 Permit inclusion of DMD scattering in standard software packages

Table1.MilestonesofthisSATproject.

ProgressandAccomplishmentsForDMDarraystobesuitableforfutureNASAmissions,anumberofperformanceissuesmustbeaddressed.Thisprojectattemptstoinvestigatethesequestions,andtoimproveDMDstomakethemmoresuitableforsuchinstrumentationrequirements.ThetwocommerciallyavailableDMDswewillbeevaluatingarethe0.7XGA1024 × 76813.6‐µmpixelpitchandtheCinema2048 × 108013.6‐µmpixelpitchDMD.

RadiationTesting:Theheavy‐ion radiation testingwasperformedat theCyclotron Instituteof theTexasA&MUniversity.Thefacilityisequippedwiththreebeamtypes:15,25,and40MeV/amu.Overthetworoundsoftesting,weusedthe25MeV/amubeamwithfourdifferentions–neon,argon,krypton,andxenon.DMDswerere‐windowedwith2‐µm‐thickpellicleandtestedunderacceleratedheavy‐ionradiation (control electronics shielded from radiation), focusing ondetection of single‐event effects(SEEs)includinglatch‐upevents.TestingshowedthatwhileDMDsaresensitivetonon‐destructiveion‐inducedstatechanges,allSEEswereclearedwithasoftreset(that is,sendinganewpatterntothedevice).TheDMDsdidnotexperiencesingle‐event‐inducedpermanentdamageorfunctionalchangesthatrequiredahardreset(powercycle),evenathighionfluence.Theprotonandheavy‐iontestingsuggeststhattheSEE‐rateburdenwillbemanageableforaDMD‐basedinstrumentwhenexposedtosolar‐particlefluxesandcosmicraysinorbit.Usingthe95%confidenceboundsobtainedfromthefittedmodel(Fig.1)wecalculateloweranduppervaluesforthepredictedworst‐casein‐orbitsingle‐eventupset(SEU)ratetobe3.42and11.7micromirrorsin24hours,foranXGADMDwith1024 768mirrors.

30

Fig.1.MeasuredSEUcross‐sectionasafunctionoflinearenergytransfer(LET)ofincidentionsfrombothruns.TheWeibullcumulativedistributionfitand95%‐confidence‐levelboundsarerepresentedbysolidanddottedlines,respectively.

Low‐TemperatureTesting:WehavecollaboratedwithcolleaguesattheInstrumentDevelopmentGroupatJHUtobuildonourpreviouseffortstotestDMDperformanceatcryogenictemperatures.AtRIT,weusedamodifiedInfraredLaboratoriesliquidnitrogendewartotesttheDMDattemperaturesas low as 130K [2]. This is the lowest temperaturewe could achieve, because theDMD controlelectronicshadtobehousedinthedewar.Thecommercialformatterboarddissipatedtoomuchheattoallowustoreach77K.

AtJHU,togetherwithourcollaborators,wehaveusedamuchlargerdewarwithacryo‐cooler(Fig.2),whichallowedus to reacha temperatureof77K,evenwhileusing thesame formatterboard.SeveralXGADMDswithHEMSapphire,fusedsilica,andoriginalTIwindowswereusedinextensivetesting.Amongthetestsperformedwerehigh‐duty‐cycletests(wheretheDMDmirrorsareflippedmanytimes,rapidlyatlowtemperature)and“operationalconditions”tests(wheretheDMDmirrorsarelatchedinthesamestateforabout20minutesatatime,aswouldbedoneinaspectrograph).Thesetestshavebeencompletedandtheresultshavebeenpublished. TheDMDdevicesworkedwithoutproblematthelowtemperatures.Theonlyissuesnotedwerewithcontaminantsintroducedduringthere‐windowing.

Fig.2.Left:TestchamberandopticalsetupusedforcryogenicDMDoperabilitytesting.Right:DMDandformatterboardmountedtothechamberinterfacebracket(FPC,FlexiblePrintedCircuit).

Re‐Coating DMDs with Al: Replacing the standard borosilicate window with a UV‐transmissivematerial(MgF2,HEMSapphire,fusedsilica)extendstheoperationalrangeoftheDMDtoapproximately200 nm.TheUVreflectanceoftheDMD(ignoringlossesduetofillfactoranddiffraction)islessthan

31

thatofpurealuminum,becausetheDMDmirrorsaremadewithanaluminumalloy.Belowabout200 nm,the DMD reflectance drops rapidly (Fig. 3, Right). This decrease is seen for all aluminum (andaluminum‐alloy)mirrorsiftheyarenotprotectedagainsttheformationofanaluminumoxidelayer.Inthelastyear,weattemptedtoextendtheuseoftheDMDtoapproximately100nm,byrecoatingtheDMDwithhigh‐purityaluminumandprotectingitwithathinfilmofLiForAlF3.

Fig.3.Left:Were‐coatedasectionofaTIXGADMDwithhigh‐purityaluminumatNASAGSFC(lighter‐colorsquare).TheDMD remainedoperationalafter the coatingprocess.Right:Reflectanceof the re‐coated regionincreasedacrossthevisiblerange,withsignificantimprovementsintheUV.

The re‐coatedDMD remainedoperational (Fig. 3, Left),with noobvious differences between thecoatedandoriginalregions,exceptimprovedreflectance.There‐coatedregionshowedareflectanceimprovementofseveralpercentacrossthevisiblerange,withmoresignificantgainsintheUV;at200nm,thereflectanceincreasedfrom48%to65%.Atthiswavelength,thereflectancedifferencebetweenthere‐coatedDMDandastandardaluminummirrorisduetothe92%fill‐factoroftheDMD.Atlongerwavelengths,diffractioncreatesadditionallosses,aswellasthe“ringing”structureseenintheblue curve.Neither the re‐coatedDMDnor thewitnessmirror samplewasprotectedagainstoxidationinthisexperiment,sothereflectancedrop‐offnear200nmpersists.However,ourinitialtestsshowthatDMDscansurvivethistypeofre‐coating,suggestingthatDMDscanbemadeusableinthe100 – 400nmrange,ifthecoatingisprotectedwithafluoridefilm.LithiumfluoridewindowscanbeusedtoprotecttheDMDusingthere‐windowingtechniqueswedevelopedaspartofthisSATprogram,toallowoperationinthe108 – 3000nmrange.Furthermore,weareinvestigatingtheuseofDMDswithout protectivewindows, to extend the usable range to 91.2 nm (rest‐frameLymanlimit),whichisnearthecurrentstate‐of‐the‐artaluminumcoatings[3,4].

Inaddition,scatteringmeasurementsonthere‐windowedXGAdevicesatwavelengthsdownto200nmhavebeencompletedandpublished.Anewprocesshasbeendevelopedandprototypedforre‐windowingDMDswithoutepoxybutratherusinglaserwelding,anddeviceshavebeenmanufacturedthatareashermeticastheoriginaldevicesfromTI.Thegamma‐rayradiationtestingofDMDsatGSFCwillbedone19‐21June2018,andthoseresultswillbeusedtodeterminetotalin‐orbitupsetrates.

PathForwardGiventherobustnessoftheDMDsinalltestsperformedwithinthescopeofthisASTprogram,webelievethedevicesarereadyforactualdeployment.Ourgoalistoproposetheiruseinasub‐orbitalprogramproposalforfinalvindicationastothirsurvivabilityandutility.

References[1] J.E.Gunnetal.,“The2.5‐mTelescopeoftheSloanDigitalSkySurvey,”TheAstronomicalJournal,

131,Issue4,2332‐2359(2006)

32

[2] K.Fourspring,Z.Ninkov,S.Heap,M.Robberto,andA.Kim,“Testingofdigitalmicromirrordevicesfor space‐based applications,” Proc. SPIE, 8618, Emerging Digital Micromirror Device BasedSystemsandApplicationsV,id.86180B(2013)

[3]M.A.Quijada,J.DelHoyo,andS.Rice,“Enhancedfar‐ultravioletreflectanceofMgF2andLiFover‐coatedAlmirrors,”Proc.SPIE,9144,91444G(2014)

[4] K.Balasubramanian, J.Hennessy,N.Raouf, S.Nikzad,M.Ayala, S. Shaklan, P. Scowen, J.DelHoyo,andM.Quijada,“AluminummirrorcoatingsforUVOIRtelescopeopticsincludingthefarUV,”Proc.SPIE,9602,96020I(2015)

Publications1. A.Travinsky,D.Vorobiev,Z.Ninkov,A.D.Raisanen,J.Pellish,M.Robberto,andS.Heap,“Effectsof

heavy ion radiation on digitalmicromirror device performance,” Optical Engineering, 55 (9),094107(2016)

2. A.Travinsky,D.Vorobiev,Z.Ninkov,A.D.Raisanen,M.Quijada,S.Smee,J.Pellish,T.Schwartz,M.Robberto,S.Heap,D.Conley,C.Benavides,N.Garcia,Z.Bredl,andS.Yllanes,“EvaluationofDigitalMicromirrorDevicesforuseinspace‐basedMulti‐ObjectSpectrometerapplication,”J.Astron.Telesc.Syst.,3(3),035003,doi:10.1117/1.JATIS.3.3.035003(2017)

3. D.Vorobiev,A.Travinsky,Z.Ninkov,andM.Quijada,“Developmentofdigitalmicromirrordevicesforusinthefar‐ultravioletregime,“Proc.SPIE10706‐211(2018)

4. A.Travinsky,D.Vorobiev,K.Oram,G.M.Nero,andZ.Ninkov“On‐skyperformanceevaluationofRITMOS,amicromirror‐basedmulti‐objectspectrometer,”Proc.SPIE10702‐59(2018)

5. R.Content,AustralianAstronomicalObservatory(Australia);Y.Wang,IPAC,Caltech(UnitedStates);M.Robberto,SpaceTelescopeScienceInstitute(UnitedStates);R.Barkhouser,JohnsHopkinsUniv.(UnitedStates);Z.Ninkov,RochesterInstituteofTechnology(UnitedStates);S.Smee,JohnsHopkinsUniv.(UnitedStates);M.Dickinson,NationalOpticalAstronomyObservatory(UnitedStates);H.C.Ferguson,SpaceTelescopeScienceInstitute(UnitedStates);L.Hillenbrand,Caltech(UnitedStates);C.Hirata,TheOhioStateUniv.(UnitedStates);W.Fraser,Queen'sUniv.Belfast(UnitedKingdom);J.Bartlett,JetPropulsionLab.(UnitedStates);R.Benjamin,Univ.ofWisconsin‐Whitewater(UnitedStates);J.Brinchmann,LeidenObservatory(Netherlands),LeidenUniv.(Netherlands);R.Chary,Caltech(UnitedStates),IPAC(UnitedStates);A.Cimatti,Univ.degliStudidiBologna(Italy);C.Conroy,Harvard‐SmithsonianCtr.forAstrophysics(UnitedStates);E.Daddi,CEA‐IRFU(France),Univ.ParisDiderot–Paris7(France);M.Donahue,MichiganStateUniv.(UnitedStates);O.Doré,PeterEisenhardt,JetPropulsionLab.(UnitedStates);G.Helou,IPAC,Caltech(UnitedStates);J.D.Kirkpatrick,Caltech(UnitedStates),IPAC(UnitedStates);S.Malhotra,NASAGoddardSpaceFlightCtr.(UnitedStates);L.Moscardini,Univ.degliStudidiBologna(Italy);M.Ressler,JetPropulsionLab.(UnitedStates);J.Rhoads,NASA/GSFC(UnitedStates);J.Rhodes,JetPropulsionLab.(UnitedStates);andA.Shapley,Univ.ofCalifornia,LosAngeles(UnitedStates),“ATLASprobeforthestudyofgalaxyevolutionwith300,000,000galaxyspectra,”Proc.SPIE10698‐17(2018)

Foradditionalinformation,contactZoranNinkov:ninkov@cis.rit.edu

33

Ultra‐StableStructures:DevelopmentandCharacterizationUsingSpatialDynamicMetrologyPreparedby:BabakN.Saif(PI;NASA/GSFC)andLeeFeinberg(NASA/GSFC)

SummaryOnepossiblesuccessortotheJamesWebbSpaceTelescope(JWST)isanobservatorythatcombinesgeneralultraviolet‐optical‐infrared(UVOIR)astrophysicswiththesearchforlifeonhabitableEarth‐likeexoplanetsusingalarge‐aperturesegmentedtelescope.Workonthisproblembeganin2009asa potential Advanced Technology Large‐Aperture Space Telescope (ATLAST) architecture. Earlyworkfocusedonascalable9.2‐msegmentedtelescopethatcouldbelaunchedonaDeltaIVHeavyvehicle.Morerecently,thishasprogressedtoa16‐msegmentedtelescopearchitecture.Themostsignificantarchitecturaldriverbeyondtheaperturesizeisthe10‐10contrastrequiredtoblockoutthe bright stars sufficiently to detect dimEarth‐like planets orbiting starswithin their habitablezones. Achieving this requires a combination of a high‐throughput coronagraph with sufficientbandpassandwavelengthrangetoperformspectroscopicsurveys,andanultra‐stabletelescopethatmaintains better‐than‐10‐picometer stability for most observations. Achieving few‐picometersstability, given thermal and dynamic disturbances, requires both passive and active means in asystemwithmulti‐levelhierarchies.

BackgroundPicometerInterferometryofReflectiveSurfaces

Animportantfirststeptoachievethislevelofstabilityisachievingpicometer‐levelmetrologythatcancharacterize the thermal and dynamic behavior of an optical system, starting from the smallestcomponents, through subsystems, up to the system as awhole. This requires ametrology systemcapableofmeasuringthermalanddynamicchangesofbothdiffuseandreflectivesurfacesofsystemelementstopicometeraccuracy.Onecannotassumethatsystemstabilityscaleslinearlywithlevelsofstimulusoverordersofmagnitude.Moreprecisely,thetransferfunctionofasystemisnotconstantoverordersofmagnitudeinstimuluslevel.

Atwhat level of accuracy can onemeasure dynamics components? Our principal approach is tocomparetheamplitudeofmeasuredZerniketermswhenthestructureisstimulatedmechanicallytomeasurements without stimulus. Different Zernike terms have different phases and theircontribution to surfacevariancevaryover a stimulus cycle.However, since theseareorthogonalfunctions, the time‐averaged total surface variance is the sum of individual‐surface Zernikevariances.Therelativecontributionsofeachtermcanbeillustratedbyshowinghowtheresidualrootmeansquare(rms)varieswiththeZerniketerm,bysuccessivelyremovingZernikecomponentsfromthenetdynamicfigureasafunctionofZerniketerm,andcomputingtheresultingrms.

ObjectivesandMilestonesOurobjectiveistodeveloppicometersurfacemetrologyofmirrorsandstructures.Tothisend,weworkwithavendortodevelopadynamicaldigitalspeckle‐patterninterferometerwithpicometerprecision. In parallel, we develop an isolated tabletop setup to measure dynamics and drift ofmaterialandsmallstructures,includingastimulisystemabletoexertpicometer‐levelexcitations.

34

Wethendevelopstructuresandmirrorswithdynamicscontrolledatpicometerlevels.Finally,weredesignandmodelthedynamicsofthemeasuredmaterial/structures.

ProgressandAccomplishmentsThe two major components of our picometer‐metrology test bed are picometer‐level dynamicsmeasurementsandathermal‐vacuumchambercapableoffraction‐of‐a‐milliKelvinstabilityovermanyhours.Interferometrytopicometerlevelsofdiffuselight,suchasthatreflectedbynon‐specularsurfaces,ischallengingbutrequiredtoenabledesignandfabricationsystemswithpicometerstability.Toachievethis,wepurchasedahigh‐speedspeckleinterferometer.

High‐SpeedSpeckleInterferometer

Attheendof2017, theteamcompletedtestingontheMadCityactivelycontrolledpiezoactuatorusinga4DHSIsystemonasmallreflectivetestmirror.Thiswasdonebyattachingaconvexmirrortotheactuatorandmeasuringthemirrormotionatitscenterofcurvature.TheactuatormotionperMadCitywasabout12‐picometer‐amplitudesinewave,andourmeasurementshoweda13‐picometer‐amplitudesinemotion.Thisgaveustheconfidenceneededtomoveontorepeatingthetestwithacompositetestarticle,usingournewhigh‐speedspeckleinterferometer.

Thehigh‐speedspeckleinterferometerwasdeliveredby4DTechnologytoGSFCinFebruaryof2018.Extensivetestinguncoveredsomeissues.High‐frequencynoiseappearedonthecamera,alongwithaberrationsduetophaseshiftingandlow‐powercouplinginthereferencearmoftheinterferometer.These problems were discussed with 4D Technology and solutions were identified. Theinterferometerwassentbackforrepair.

Ofthreeproblemsidentified,thehighspatialnoiseonthecameraneededimmediateattentionasitpreventedusfromdoingourpicometer‐levelmeasurements.Thiswasisolatedtothecameracablingandwasresolved.Phase‐shiftingaberrationsisbeingaddressedbyorderingnewWollastonprism,whilelow‐powercouplinginthereferencearmisbeingaddressedbyorderinganewfibercoupler,with both parts expected to be delivered in twomonths.With the cabling repair completed, theinterferometerreturnedtoGSFC,allowingustobeginourspecklemeasurements.

Forthismeasurementwemeasuredadiffusesurface,apieceofcarbonfiberattachedtothefrontface of the mirror using super glue, using the speckle mode of the new interferometer. A 20‐picometer‐amplitude sine wave was applied to the actuator as was done for the reflectivemeasurements lastyear.Themeasurement (Fig.1) showeda16‐picometer‐amplitudesinewave.These data show that ourmeasurements do not suffer biases and systematics, and thatwe canachieveadequateresolution.Thelargesignaltonoiseratio(SNR)showsthattheamplitudecaneasilybereducedbyafactoroftwoandstillbedetectablewithgoodSNR.

35

Fig.1.CarbonFibermotion

TemperatureStabilitytoMilliKelvins

TheSmithsonianAstrophysicalObservatory (SAO)LargeUVOIR (LUVOIR)ultra‐stable aluminumthermal‐vacuumchambersystem(Fig.2)is30”indiameterand30”inlength.Theassemblyincludestwo bolt‐on aluminum doors (35” outer diameter, OD) with Viton O‐ring seals and handles tofacilitatedoormounting.Onedoorincludesa10”‐diameterfused‐silicaviewport,andtheotherdoorhasprovisionformountingasecond10”viewport.Threeheaterzonesmaintainhighstabilityforthetestarticlewithinthechamber.Theheaterpadsareadheredtotheexternalsurfacesofthecylinder.Thermal and acoustical barrier foam panels surround the chamber to dampen local ambientacousticalnoiseandambienttemperature.Thethermalcontrolelectronicsrack(Fig.2)includesacontrollaptopwithSAO’snested‐heatercontrol‐loop‐logicsoftware,aheater‐powerdrivemodule,ahigh‐precisionthermometrysystem,andapowersupply.

Fig.2.Left:Thermal‐vacuumchamber.Right:Ultra‐Stablethermometrysenseandcontrolsystemelectronicscontrolrackforthechamber,currentlylocatedintheAIMlabinBuilding5GSFC.

Functionaltestingwithanominalset‐pointof23.5°Cshowedanaveragetest‐articlethermalstabilityof+0.4/‐0.2milliKelvin(23.5+0.0004/‐0.0002°C)over80hours(Fig.3).

36

Fig.3.Testresults:23.5+0.0004/‐0.0002°Cthermalstabilityachievedforover80hours.

Ultra‐StablePicometer‐ScaleMirrorAssembly

Preparationsforopticaltestsusingathermallycontrolledultra‐stableULEmirrorassemblyinsidethe SAO chamber are underway for picometer‐scale speckle interferometry characterizations infiscalyear(FY)2019.Tosupporttheseactivities,anUltra‐StablePicometer‐ScaleMirrorAssembly(USPS‐MA)opticaltestarticlewasdevelopedforintegrationinsidetheSAOchamber(Fig.4).ThechamberiscurrentlylocatedattheGSFCAIMLabinBuilding5.Asofthiswriting,theUSPS‐MAisfullyassembledandundergoingfunctionaltestingforaplanneddeliveryon18July2018.

Fig.4Cross‐sectionalviewoftheUSPS‐MAtestconfigurationintheSAOultra‐stablethermal‐vacuumchamberwith6‐dof(degreeoffreedom)opticalstagesforproperorientationwithrespecttotheopticalwindow.

TheUSPS‐MAwillbeusedinthreeconfigurations:(1)witha100‐mm‐diameterULEtestsubstratewitha1‐µmrmsdiffusesurface(Fig.5), (2)witha150‐mmULEtestsubstratewitha1‐µmrmsdiffusesurface(Fig.6),and(3)withasurrogatemirrorassemblyusinga150‐mmfused‐Sisubstrate.Thethirdconfigurationisforcharacterizingthethermalcontrolbehavior,helpingusunderstandanythermalgradientswithinthemirrorsubstrate,andasareferenceforcontrol‐temperature‐sensorlocations.Duetothelimitednumberofchannelsinthecurrentthermometrysystem,onlyonesensorpersubstrateisplannedasacontrolsensorontheULEsubstrates.Thesurrogatefused‐Sisubstratewas‘instrumented’withfivehigh‐precisionthermistors,allowingustocharacterizeitandprovidethenecessaryinformationforthetwoULEsubstrates.

37

Fig.5.Left:USPS‐MAwith100‐mm‐diameterassembly.Right:Explodedviewforclarityofassembly.

Fig.6.Left:USPS‐MAwith150‐mm‐diameterassembly.Right:Explodedviewforclarityofassembly.

TheUSPS‐MAconsistsofthermalcontrolhardwareasshowninFig.7below.Thethermalcontrolhardwareincludes,fromlefttoright:

1. A thermaldiffuserplate to ‘blend’possible thermal gradients that couldbe ‘seen’by the testarticlesubstrate.

2. AheaterplatewithfourcircularKaptonheaters.3. Acoldplateinconjunctionwithare‐circulatingchillertoprovideacold‐biasenvironmentforthe

heaterplate.4. Two(2×)insulatingthermalbafflestopreventthermalinstabilitiestothethermal‐vacchamber

thermalcontrolanditsperformance.5. Anoutercoverto‘house’theinternalthermalcomponents.

Fig.7.Explodedviewofthethermalhardwareassociatedwiththethermalcontrolsystemthatwillmaintainthemirrorsubstratesatthemilli‐Kelvinstabilityrequiredforpicometer‐scalemetrology.

TheULEsubstrateswerebondedtoholdingframesusinganSAO‐developedminiatureflexuredesignsimilar to that shown in Figs. 8 and 9. This substrate edge‐bonding techniqueminimizes epoxy‐induced surface figure errors (SFE). Also, bonding the substrate to ametal frameminimizes thelikelihood of potential micro‐lurching associated with ‘dry’ mirror interfaces such as knife‐edgetripodstands.

38

Fig.8.Mirrorsubstratemounting implementedtosecureacircularULE flattodemonstratethepicometer‐levelmetrologysystemalongwiththemilli‐Kelvinthermalcontrolsystem.Afused‐Sisquareflatsubstrateismountedina titanium frame usingmicro‐flexures bonded between the thinmirror substrate and the housing frame. Theassemblywasvibratedtoverifyitsstructuralintegrity.

Fig.9.Close‐upillustrationofthemicro‐flexuredesign.Theedgeshownisthe1‐mm‐thickULEsubstrate.

PathForwardTheULE‐substrateUSPS‐MAarticleshavebeenfullymanufacturedandassembledandarethermallycontrolled using the same high‐precision thermometry sense and control system (Fig. 2, right)developedfortheultra‐stablechamber.Functionaltestingisunderwayatthetimeofthiswriting(Fig.10).Thetestarticleassemblywillthenbeplacedintheultra‐stablethermal‐vacuumchamberforfurthertestingofthepicometer‐classmetrologysystem.

Fig.10.USPS‐MAfunctionaltestingwithsurrogatefused‐SisubstrateinsidetheSAOSESchamberpriortodeliverytoGSFC.

Foradditionalinformation,contactBabakN.Saif:babak.n.saif@nasa.gov

39

BuildingaBetterALD‐useofPlasmaEnhancedALDtoConstructEfficientInterferenceFiltersfortheFUVPreparedby:PaulScowen(PI;ArizonaStateUniversity,ASU);BriannaEller,DanielMessina,ZhengJu,FranzKoeck,RobertJ.Nemanich,andHongbinYu(ASU);TomMooney(Materion);andMattBeasley(SouthwestResearchInstitute,Boulder)

SummaryThe goal of thiswork is to use plasma‐enhanced atomic layer deposition (PEALD) to synthesizemirrorsandfilterscompatiblewithnear‐ultraviolet(UV)andfar‐UVoptics.Thedevelopmentofthistechnologywillultimatelyprovidediagnostictoolstoaccessarangeoftopicsforstudy, includingprotostellar and protoplanetary systems, intergalactic‐medium (IGM) gas from galactic starformation,andthemostdistantofobjectsintheearlyuniverse.Sincethebeginningoftheyear,ourteam,fromtheSchoolofEarthandScienceExplorationandthePhysicsDepartmentatASU,Materion,andPlanetaryResources,hasbeenworkingtoinitiatethisresearch.Ourmostsignificantprogresstodate has been the design, development, and the partial assembly of the equipment necessary tocompletethisresearch.Inthisperiod,theassemblyofthePEALDhasadvancedwithdemonstrationoftheplasmasystemandnearcompletionofthegasdeliveryandcontrolsystem.

1.1 BackgroundAtomiclayerdeposition(ALD)isalayer‐by‐layerdepositiontechniquethatsynthesizesultra‐thin,uniform,andconformal filmsasshowninFig.1.Thehighqualityofthesefilmshasconsequentlyresultedinaugmentedcoatingsandopticalelements.Atthesametime,majoradvanceshavebeenmadeinopticaldesignsanddetectortechnologies.Asaresult,measurementoffar‐UVandnear‐UVbandshasimproveddramatically.Thedevelopmentofthistechnologyultimatelyallowsaccesstoemission and absorption lines in the UV, which are emitted from a range of targets, includingprotostellarandprotoplanetarysystems,IGMgasfromgalacticstarformation,andthemostdistantofobjectsintheearlyuniverse.Thesediagnostictoolsrequiretheimplementationofstableopticallayers,includinghigh‐UV‐reflectivitycoatingsandUV‐transparentfilms[1].

Fig.1.Schematicrepresentationofthelayer‐by‐layerdepositionprocessofthethermalandplasma‐enhancedALD.Duringthereactantorplasmaexposure,thesurfaceisexposedtoareactantgasorplasma.

Inthiswork,wewillusearangeofmaterialstoimplementstableprotectiveovercoatswithhighUVreflectivityandunprecedenteduniformityandusethatcapabilitytoleverageinnovativeUV/opticalfilterconstructiontoenablethesciencementionedabove.Thematerialswewilluseincludealuminum

40

oxideandsiliconoxide(asanintermediarystepfordevelopmentonly)andarangeoffluoride‐basedcompounds(forproduction).Thesematerialswillbedepositedinamultilayerformatoverametalbasetoproduceastableconstruct.Specifically,wewillusePEALDfordepositionandconstructionofreflectivelayersthatprotectbarealuminumformirroruseintheUV.Ourdesignsindicatethatbyusing PEALD we can reduce adsorption and scattering in the optical films because of the lowerconcentrationof impurities and increased controlover the stoichiometry, yielding vastly superiorqualityandperformanceovercomparabletraditionalthermalALDtechniques[2‐17]currentlybeingdevelopedbyotherNASA‐fundedgroups[18].ThesecapabilitieswillallowustopushtheblueedgeinusableUVreflectivityformagnesium‐fluoride‐protectedaluminumbelowthecurrent115‐nmlimit.

Thisworkwilldemonstrateforthefirsttimewhetherloss‐freeoxidesofmaterialssuchasAl,Hf,andSicanbedepositedusingALDtolowercutoffreflectivitiesintheUVtoaslowas92nm.WewillalsodemonstratetheuseofPEALDtodepositlow‐lossthinfilmsoffluoride‐basedmaterials,andaluminummetal.Usingthesetechniques,wewillthendemonstrateourproof‐of‐conceptofusingthesetechniquestogether to construct thin‐film,multilayermetal‐dielectric cavitieswith a reflective surface as thefoundation, that can be tuned to isolate specific emission lines of astronomical importance. Theresultingoptical technologieswill advance thecoating stability, thickness, andperformanceof thinfilms in the far‐UVsoughtbyNASA, tomatchrecentUVdetectoradvances.Such improvementwillenable next‐generation space‐based far‐UV missions, opening access to the wealth of diagnosticinformationthefar‐UVoffersforexoplanet,starformation,andcosmological/IGMscience.

1.2 ObjectivesandMilestonesOurresearchseekstodemonstrateseveralobjectives:

Films ofmaterial can be deposited to demonstrate the approach using PEALD techniques toproduceloss‐freefilmsofe.g.silicon,hafnium,andaluminumoxide;theresultingcoatingswillbeofathicknessandapurityfarhigherthancanbedeliveredbycurrenttechniquesthatinvolvesputteringdeposition;

Usingthesamedepositiontechniques,PEALDcandepositthin(tensofnm)low‐lossfilmsoffluoridesofaluminumandmagnesiumaswellase.g.lithium,lanthanum‐calcium,andberyllium,thatcanserveasprotectiveovercoatsformaterialswhichwouldotherwisebeeasilyoxidizedbyexposuretoair;

Aluminumdeposition,protective‐layerdeposition,andcharacterizationcanbecompletedin‐situin a controlled environment that minimizes contamination, improving the reflectivity of theresultingfilmsandtheirinterfacesbyreducingscatteringandadsorption;

DepositionofsuchprotectiveovercoatsoveraluminummetalcanbeachievedwithPEALDtoprovideasufficientlycrystalline,uniform,andstablestructure,pushingbluewardthecurrentlyobserved 115‐nm cutoff in efficient reflectivity from atomic sputtering deposition ofmagnesiumfluoride,therebyextendingtherangeofdiagnosticemissionandabsorptionlinesavailableforscience;

Extendthemetal‐dielectricovercoatprocesstoconcavemirrorstodemonstratetheperformanceofthereflectivesurfacesinanopticaltestbedenvironment;

UseourPEALDapproachtoapplyalternating layersofmetalsanddielectrics,producingmulti‐cavitystructuresexhibitingveryhighperformance;thisgoaliscurrentlylimitedbytheinabilitytodepositverythinlayerswithgreataccuracy,whiledemonstratingfilmtoughnessand‘bulk’thin‐filmmateriallosses;

Apply themultilayer approach to the construction ofmulti‐layer dielectricmirrors to act asreflectionfiltersorhighreflectorsinnarrowbandsystems;and

Similarlyconstructmulti‐layerbroadbandmirrors,thoughttoexhibithigherperformancethanmetal‐basedmirrors(usingashort‐waveextensiontoprototypedichroicsourgroupisalreadydevelopingforspaceapplications).

41

Table1summarizesthetimelinetoachievetheseobjectives,modifiedtoaccommodatefabricationofequipment.

Activity Name Start Date Finish Date Upgrade in-situ reflectivity to 120 nm Completed

PEALD of oxides on evaporated aluminum Completed

PEALD of Al2O3 and SiO2 on aluminum Completed

Upgrade PEALD for aluminum deposition 1/1/17 8/31/18 PEALD of aluminum 7/1/17 8/31/18 Install precursors for PEALD of fluorides 8/31/17 8/31/18 Oxides on PEALD aluminum 10/3/17 9/30/19 Upgrade in-situ reflectivity to 90 nm 8/2/17 9/30/19 PEALD of aluminum fluoride on aluminum 10/23/17 10/31/18 PEALD of magnesium fluoride on aluminum 10/23/17 10/31/18 PEALD of aluminum and magnesium fluoride 12/2/17 10/31/18 Magnesium fluoride/aluminum filters 2/1/18 11/30/18 Oxide-fluoride multilayers and protective layers 4/1/18 11/30/18 Multi-layer broadband mirrors 4/1/18 11/30/18

Table1.Timelineforobjectivesandmilestones.

1.3 ProgressandAccomplishmentsThe initial state of this research has largely focused on developing the equipment necessary tosynthesize and characterize the oxygen‐free structures required to achieve the aforementionedobjectives. Assembly of the systems is underway, where the chambers added to the ultra‐highvacuumsystemarehighlightedinFig.2.Specifically,twosystemshavebeenaddedtothesetup:

AfluoridePEALDsystemforboththealuminummetalandmetalfluoridesneededforthiswork;and AvisibleandUV(VUV)opticalsystem,whichwillbeusedtocharacterizereflectanceofthefilms

depositedwithoutatmosphericcontamination.

Fig.2.Photo(top)andschematic(bottom)ofASUin‐situultra‐highvacuumsystem;blueshowsnewsystemsinprogress (UPS, UV Photoelectron Spectroscopy; AES, Auger electron Spectroscopy; XPS, X‐ray Photoelectron

42

Spectroscopy;MBE,MolecularBeamEpitaxy;iPlas,InnovativePlasmaChemicalVaporDeposition;ECR,ElectronCyclotronResonance).

Akeycomponentofthisworkisrelatedtothein‐situnatureofthedepositionandcharacterizationsoontobeenabled.Sinceoxidationofaluminumhaspresentedasignificantchallengeinpreviousresearch [18], this work will allow for the deposition of aluminum andmetal fluorides withoutexposuretoatmosphere.

1.3.1 1.Plasma‐EnhancedAtomicLayerDeposition

ThenewPEALDsystemisintheinitialstagesofassemblyandbasedonthepreviousoxidesystemavailableinthelab,withonlyafewalterations.Specifically,theplasmawasdesignedwitharemoteconfiguration, helping reduce ion bombardment of the sample, mitigating potential ion damage.Plasmawill be ignitedwith 13.56‐MHzRF‐excitation applied at ~200W to a helical copper coilwrappedarounda32‐mm‐diameterquartztubeandmaintainedatapressureof~100mTorrwithaflow rate of ~35 sccm (standard cubic centimeters per minute). This system must achieve abackground pressure of < 5×10‐8 Torr and processing pressures of ~10 mTorr. The pumpingrequirements,therefore,varyduringtransferanddeposition.Toassistwithtransfer,thesystemisequipped with a Pfeiffer turbo with pumping speed of 300 liters/sec and a dry backing pump,enabling lower pressures; however, the chemicals used during deposition are often too harsh,reducingthelifetimeoftheturbopumps.Therefore,whenoperating,theturboisisolatedwithagatevalve, and anEbara two‐stagedrypumpwith apumping speedof~7000 liters/sec isused.Thepumpingstageisventedwithnitrogengasduringoperationtofurtherensuresystemlongevity.Inaddition,thegas‐flowmechanismsaredesignedtodelivertheprecursorstothechamberwiththecorrecttimingsequenceusingmassflowcontrollers(MFCs),pneumaticallyactuatedvalves,andacustomLabViewprogram.Meteringvalveswerealsoaddedtothegaslinestofurthercontroltheamountofprecursorreleasedintothechamber.Lastly,abutterflyvalvebeforethetwo‐stagedrypumpisusedtomaintaintherequiredpressuresduringprocessing.

The system is in the final stages of completion, as shown in Fig 3. To facilitate assembly, thefabricationhasbeencompletedinstageswherethefirststagesareasfollows:

43

Fig.3.Left:FrontviewofPEALDsystem.Thisrackhousesallcontrolequipment.Center:Sideviewofsystem.Mountedon the frame is thechamberandequipmentessential for thegasdelivery system.The stainless‐steelcanisterisfortrappinglargerreactorbiproductswhilethewhitecanisterisforHFabatement.Thepumpinthebackisthetwo‐stagedrypump.Right:Sampleholdermountedonarotaryflangeinsidethechamber. Stage1.VacuumandTransferMechanisms.Thisincludesthedesignofthesystemandchamber

aswellastheinstallationoftheequipmentneededtoachieveandmaintainvacuum,aswellastransferthesamplesfromthein‐situvacuumsystem.

Stage2.Plasmagenerationandgasdelivery.Theplasmageneratingsystemhasbeenaddedandsuccessfullytestedwithargonat35sccmand400mTorr.Asimilartestwillbecarriedoutoncethebutterflyvalveandtwo‐stagedrypumpareoperational,inordertoobtainthedesired100mTorr operating pressure.The rack to support the equipment has been fabricated andmounted.Inaddition,theplumbingfortheargonandhydrogenlinesisinprogress.

Stage3.AluminumPEALDprocessandgasdelivery.Fortunately,therehasbeensomeworkdoneon PEALD of aluminum, using trimethylaluminum and hydrogen plasma [19]; therefore, thisprocesswillbeused.Theremainingsub‐systemsare:installationandweldingofgaslines,fillingandmountingprecursorcontainers,andcreatingacustomLabViewprogramforsystemoperation.

Stage4.Modification for fluorides.Unfortunately, thecorrosivenessand toxicityofHFraisesafetyconcernsthatmustbecarefullyconsideredforthisprocess.Toaddresstheseconcerns,thefollowingprecautionswillbeimplemented:

StoringtheHFinabubblerasanHF‐pyridinemixture,whichislessvolatilethanHF; Electroplatingthebubblerwithgoldtopreventcorrosion; Replacingthequartztubewithsapphiretopreventetching; Introducinganaciddrybedabatementsystemfollowingthetwo‐stagedrypumptoremove

wasteby‐products; AddingatoxicgasanalyzertoensureallresidualHFiswithinacceptablelimits;and Coatingthechamberwithathickaluminumlayer,priortousingHF,topreventetchingof

thechamber.

Stage5.Fluorideprecursors.Fluorideprocesseshavenotbeendevelopedextensively.Infact,todate,therearenopublishedprocessesforfluoridePEALD.However,therehasbeensomeworkon thermalALDprocesses,whichuseHF [20,21],TiF4 [22‐25],andTaF4 [26]as the fluorinesource.WhileTiF4andTaF4comewithsignificantlylesssafetyhazards,thefilmsdepositedwiththesematerialsaretypicallycharacterizedbysomemetalcontamination,whichwilllikelycauseabsorption in the desiredwavelengths. A fluoride‐based plasmawas also considered for thiswork but dismissed due to concerns that it might lead to etching rather than deposition.Therefore,anHF‐pyridinemixtureseemstobethebestoptionatthispoint.TotakeadvantageofthedesirablepropertiesofPEALD,theHFstepwillbecoupledwithahydrogenplasmasteptoensure film purity. For aluminum and magnesium fluoride, trimethylaluminum and bis‐ethylcyclopentadienylmagnesiumhavebeenchosenastherespectivemetalprecursors.

1.3.2 2.VisibleandUltravioletOpticalSystem

Wehavealsoestablishedin‐situreflectivityto120nmtocharacterizePEALDoxidesandaluminum.Inaddition,theVUVopticalsystemwillneedtobeupgradedforlowerwavelengths(>90nm)inthefuture, whichwill be accomplished by replacing the deuterium lampwith a windowless hollowcathode lamp using inert gases or mixtures. This system (Fig. 4) has also been designed andfabricated,andisundergoingcalibration.

44

Fig.4.Imagesofthecustom‐designedUVspectrometerunit,whichcurrentlycandeliverreflectivitymeasurementsfrom254to120nm.

2.1ThermallyevaporatedaluminumcoatedinPEALDoxides.SometestingofPEALDAl2O3hasbeenexploredaswellasameanstotestandestablishprotocolsforthesystem.ThistestingwillcontinuewhilethealuminumandfluoridePEALDsystemisinprogress.Someinitialresultsshow,asexpected,thatthethicknessoftheAl2O3layersinfluencesthereflectivityatlowerwavelengthsasshowninFig.5.

Fig.5.UVreflectanceofPEALDAl2O3onthermallyevaporatedAl.Someadditionalworkisneededtoensure

consistency.

2.2UVreflectanceofadditionalmirrors.Additionalmirrorshavebeenpurchased,includinggold,silver, aluminum, UV aluminum, and MgF2‐coated aluminum; to test, develop, and implementrigorousprotocolstoensurereliableandconsistentmeasurementsinthissystem.Inparticular,smallfluctuationsinpositionandpressureinfluencethemeasuredintensity.Therefore,additionalworkisneededtoensurethesystemprovidesreliablemeasurements.Usingsampleswithknownreflectanceishelpingtoisolatetheseissues.

1.4 3.FilterDesign–theKeyContributionsofTomMooneyandMatthewBeasley

Matthew Beasley inspected the facilities in progress andmade specific suggestions as to whichmaterialsshouldbeprioritizedforALDdeposition.

ThechamberwouldbecapableofAl,AlF3bydefaultandDr.BeasleysuggestedMgF2andLaF3.

TomMooneycreatedatrialrunofafilterusingconstantsintheliteratureandanalyzedthetolerancetovaryingfilmthickness.Thisprovidedthedirectionthatmeasuringtheopticalconstantsoftheas‐depositedfilmsisthenextkeystep.Thatwayafilterwithdesiredperformancecanbedesignedforfabricationandtesting.

TomMooneyretiredandwenowhaveDr.RobertSprague(alsoofMaterion)takinghisplace.

150 200 250 300 350 4000.0

0.2

0.4

0.6

0.8

1.0

inte

nsi

ty

Wavelength

thermally evaporated aluminum + 4.5 nm ALD Al2O3 + H2 plasma + 4.5 nm ALD Al2O3

45

1.5 4.Accomplishmentsthispastyear

Thefollowingaresomeofourteam’saccomplishmentsthispastyear.

Our filter design and fabrication technology was included in a suborbital rocket proposalsubmittedinresponsetoROSESAPRA2018byco‐IMatthewBeasleyasPI;

We developed alternative funding for diamond high‐temperature electronics on Venus withNASA’sTheVenusExplorationAnalysisGroup(VEXAG);

WeexpectadditionalfundingthroughARPA‐Eforatomic‐layeretchingofGaN,enabledbyHFtechnologysimilartothatofthisSATproject;and

We leveraged our expertise to begin developing ultra‐wide bandgap semiconductors, wherefluorideswilllikelyplayacentralrole.

1.6 PathForwardTheplanofworkisstructuredaroundthegoalsandmilestonesstatedabove.Ingeneral,theplanisto focusonassembling thePEALDassoonaspossiblewhilecontinuing thecalibrationof theUVspectrometer. Thiswill enable the fabrication and use of bandpass filters. More specifically, thefollowingwillbenecessary:

1. DemonstratetheuseofPEALDtodepositloss‐freeoxidesofsiliconandaluminumonevaporatedande‐beam‐depositedaluminumsurfaces.Theseoxideswereselectedbasedontheirpromiseforstable,high‐performanceoxides.Thisismotivatedbecausecurrentmethodsproducelossyfilms,whicharenotusablebelow190nm.

2. Establish thecapability forPEALDaluminum filmsandcharacterize theUVreflectivityof thefilms and the surface contamination using in‐situ characterization tools. These aluminumsurfaceswillbethebasisofthestructurespursuedinthisprogram.

3. DemonstratethefeasibilityofusingPEALDtodepositlow‐lossfilmsoffluoridecompoundsonALDaluminum.Thefluoridesof interest includemostsignificantlyaluminumandmagnesiumfluoride,butotherswillbeconsideredaswell.Wewillneedtodemonstratestability,uniformity,andperformancebeforeadvancing.

4. Extend the reflectivity characterization to 90 nm and characterize the fluoride‐aluminumstructures. This is a critical step, as the aluminum‐oxide/fluoride layer will serve as thefoundationnotonlyforsimplereflectivesurfaces.Forfilters,weintendtoshowproof‐of‐conceptinthenextsegment.Thequalityofaluminumandinterfacewilldefinetherequiredperformance,whichwillbeverifiedbeforeadvancingtothenextstages.

Inthissystemdevelopmentphase,workwillbeconductedbyDr.Nemanichandhisteam,focusingondevelopingprocessesthatresultinoptimizedfilms.Anumberofcomplexitiesinthesystemdesignandconstructionhavedelayedthedevelopmentphase,whichisnowanticipatedtobecompletedinlate2018.

Beforemovingontothenextsequence,theopticsdesignteamofDrs.Scowen,Mooney,andBeasleywillrevisitthedemonstratedperformanceofthethreeclassesofproduct(above)andcomparethemtothemodelsthatinitiatedthework.Stabilitywillbemeasuredusingvacuumexposureandspectralretest,andsimilarlyhumidityexposurefollowedbyspectralretest.Iftherearedifferencesinstabilityandreflectivity,weneedtounderstandtheiroriginsbeforewecanbuildprototypefilters,sincetheseparametersarecriticalfortuninganindividualfilterforaparticularbandpass.Oncethishasbeendone,wewillimplementthetechniquesusedaboveforthedepositionofmultilayerconstructions.

Thenextphasewillthenaddressthefollowinggoals:

5. Optimize far‐UV reflectivity of fluorides on aluminum, to deliver a wide‐bandpass far‐UV‐optimized mirror. The optimization will focus on minimizing interface contamination and

46

optimizing material‐growth parameters and film thickness for far‐UV reflectivity. We willdemonstrateperformanceasfaraspossibleintothefar‐UVwithournewVUVsystem.

6. Construct metal‐dielectric Fabry‐Perot band‐pass filters using aluminum and magnesiumfluoride,leadingtomulti‐cavitystructuresthatwillexhibitveryhighperformance—basedonourmodels.Thisdesignapproachiscurrentlylimitedbytheabilitytodepositverythinlayerswithgreataccuracy,filmtoughness,and‘bulk’thin‐filmmateriallosses.Webelieveonlytwotofivelayerswillbeneededtodemonstrateproof‐of‐concept,witheachlayer~10nmthick.

7. Fine‐tune approach and models to produce designs for multi‐layer dielectric (narrow‐band)mirrorstoactasreflectionfiltersorhighreflectorsinnarrow‐bandsystems.Wewilldemonstrateconstructionandperformance.

8. Demonstratetheconstructionofmulti‐layerbroad‐bandmirrors,whichwebelieve—basedonmodels—willexhibithigherperformancethanmetal‐basedmirrors.

9. Weexpectthisdesignandconstructionphasetooccupymostoftheremainderof2018.Toachieveallthiswork,wewilldemonstrateourapproach,designs,andmethodologiesdeliverthenecessaryimprovementsinthicknesscontrol,lower‘bulk’losses,absenceofcolorcenters,smootherfilms(lowerscatter),abilitytodepositverythinfilms(withbulk‐likeopticalproperties),durability,andlowerstress.Theaboveplan,executedinthismanner,willdemonstrateallthesegoals.

1.7 References[1] J.M. Shull and C.W.Danforth, “Identifying theBaryons inaMultiphase IntergalacticMedium,”

arXiv1208.3249S(2012)

[2] O.K.Kwon,S.H.Kwon,H.S.Park,andS.W.Kang,“PEALDofarutheniumadhesionlayerforcopperinterconnects,”J.Electrochem.Soc.151,C753(2004)

[3] O.K.Kwon,S.H.Kwon,H.S.Park,andS.W.Kang, “Plasma‐enhancedatomic layerdepositionofrutheniumthinfilms,”Electrochem.Solid‐StateLett.7,C46(2004)

[4] J.W.Lim,S.J.Yun,andJ.H.Lee,“LowtemperaturegrowthofSiO2filmsbyplasma‐enhancedatomiclayerdeposition,”ETRIJ.27,118(2005)

[5] W.J.Maeng,S.J.Park,andH.Kim,“Atomic layerdepositionofTa‐basedthin films:Reactionsofalkylamideprecursorwithvariousreactants,”J.Vac.Sci.Technol.B24,2276(2006)

[6] M.K. Song and S.W.Rhee, “Phase formation in the tantalum carbo‐nitride filmdepositedwithatomiclayerdepositionusingammonia,”J.Electrochem.Soc.155,H823(2008)

[7] J.S.Park,M.J.Lee,C.S.Lee,andS.W.Kang,“Plasma‐enhancedatomiclayerdepositionoftantalumnitridesusinghydrogenradicalsasareducingagent,”Electrochem.Solid‐StateLett.4,C17(2001)

[8] J.S.Park,H.S.Park,andS.W.Kang,“Plasma‐enhancedatomiclayerdepositionofTa‐Nthinfilms,”J.Electrochem.Soc.149,C28(2002)

[9] J.Y.Kim,K.W.Lee,H.O.Park,Y.D.Kim,H.Jeon,andY.Kim,“BarrierCharacteristicsofTaNFilmsDepositedbyUsing theRemotePlasmaEnhancedAtomicLayerDepositionMethod,” J. KoreanPhys.Soc.45,1069(2004)

[10] J.Y.Kim,Y.Kim,andH.Jeon,“CharacteristicsofTiNFilmsDepositedbyRemotePlasma‐EnhancedAtomicLayerDepositionMethod,”Jpn.J.Appl.Phys.,42Part2,No.4B,L414(2003)

[11] Y.Kim,J.Koo,J.W.Han,S.Choi,H.Jeon,andC.G.Park,“CharacteristicsofZrO2gatedielectricdeposited using Zr t–butoxide and Zr(NEt2)4 precursors by plasma enhanced atomic layerdepositionmethod,”J.Appl.Phys.92,5443(2002)

47

[12] J.Koo,Y.Kim,andH. Jeon, “ZrO2GateDielectricDepositedbyPlasma‐EnhancedAtomicLayerDepositionMethod,”Jpn.J.Appl.Phys.,41Part1,No.5A,3043(2002)

[13] B. Hoex, J. Schmidt, P. Pohl, M.C.M. van de Sanden, and W.M.M. Kessels, “Silicon surfacepassivationbyatomiclayerdepositedAl2O3,”J.Appl.Phys.104,044903(2008)

[14] J.Koo,S.Kim,S.Jeon,H.Jeon,Y.Kim,andY.Won,“CharacteristicsofAl2O3ThinFilmsDepositedUsing Dimethylaluminum Isopropoxide and Trimethylaluminum Precursors by the Plasma‐EnhancedAtomic‐LayerDepositionMethod,”J.KoreanPhys.Soc.48,131(2006)

[15] P.K.ParkandS.W.Kang,“EnhancementofdielectricconstantinHfO2thinfilmsbytheadditionofAl2O3,”Appl.Phys.Lett.89,192905(2006)

[16] P.K.Park,E.S.Cha,andS.W.Kang,“InterfaceeffectondielectricconstantofHfO2∕Al2O3nanolaminatefilmsdepositedbyplasma‐enhancedatomiclayerdeposition,”Appl.Phys.Lett.90,232906(2007)

[17] P.K.Park,J.S.Roh,B.H.Choi,andS.W.Kang,“InterfacialLayerPropertiesofHfO2FilmsFormedbyPlasma‐EnhancedAtomicLayerDepositiononSilicon,”Electrochem.Solid‐StateLett.9,F34(2006)

[18] J.Hennessy,A.D.Jewell,F.Greer,M.C.Lee,andS.Nikzad,“Atomiclayerdepositionofmagnesiumfluorideviabis(ethylcyclopentadienyl)magnesiumandanhydroushydrogenfluoride,”J.Vac.Sci.Technol.A33,01A125(2015)

[19] J.‐H. Park, D.‐S. Han, Y.‐J. Kang, S.‐R. Shin, and J.‐W. Park, “Self‐forming Al oxide barrier fornanoscaleCuinterconnectscreatedbyhybridatomiclayerdepositionofCu–Alalloy,”J.Vac.Sci.Technol.A32,01A131(2014)

[20] Y.LeeandS.M.George,“AtomicLayerDepositionofMetalFluoridesusingVariousPrecursorsandHydrogenFluorides,”ALDconference(2015)

[21] J.Hennessey,B.K.Balasubramanian,A.Jewell,S.Nikzad,C.S.Morre,andK.France,“ThinALDfluoridefilmstoprotectandenhanceAlmirrorsinFarUV,”ALDconference(2015)

[22] T.Pilvi,T.Hatanpää,E.Puukilainen,K.Arstila,M.Bischoff,U.Kaiser,N.Kaiser,M.Leskelä,andM.Ritala,“StudyofnovelALDprocessfordepositingMgF2thinfilms,”J.Mater.Chem.17,5077(2007)

[23] M.Mäntymäki,J.Hämäläinen,E.Puukilainen,F.Munnik,M.Ritala,andM.Leskelä,“AtomicLayerDepositionofLiFThinFilmsfromLithdandTiF4Precursors,”Chem.Vap.Dep.19,111(2013)

[24] M.Mäntymäki,M.J.Heikkilä,E.Puukilainen,K.Mizohata,B.Marchand,J.Räisänen,M.Ritala,andM.Leskelä,“AtomicLayerDepositionofAlF3ThinFilmsUsingHalidePrecursors,”Chem.Mater.27,604(2015)

[25] T.Pilvi,K.Arstila,M.Leskelä,andM.Ritala,“NovelALDprocessfordepositingCaF2thinfilms,”Chem.Mater.19,3387(2007)

[26] T.Pilvi,E.Puukilainen,U.Kreissig,M.Leskelä,andM.Ritala,“AtomicLayerDepositionofMgF2ThinFilmsUsingTaF5asaNovelFluorineSource,”Chem.Mater.20,5023(2008)

Foradditionalinformation,contactPaulScowen:paul.scowen@asu.edu

48

High‐PerformanceSealed‐TubeCross‐StripPhoton‐CountingSensorsforUV‐VisAstrophysicsInstrumentsPreparedby:OswaldSiegmund(SpaceSciencesLaboratory,UCBerkeley)

SummaryMicrochannel‐plate (MCP) detectors have been the detector of choice for many ultraviolet (UV)astronomymissionsandinstrumentsoperatedsuccessfullyoverthelasttwodecades,suchasEUVE,FUSE,GALEX,HST‐STISandHST‐COS[1‐4].TheyarealsoimplementedonthesoontobelaunchedNASAICONSmallExplorer[5],GOLDMissionofOpportunity[6],andarewidelyusedonNASAUVsub‐orbitalsoundingrocketinvestigations.MCPdetectorshavethepotentialtocombinehighspatialresolution(<20µmfullwidthathalfmaximum,FWHM),photon‐counting(noiseless)imaginginarobust,radiation‐hardpackagethatisscalabletoverylargeformats(>10cmand>5k×5kresels)andevencurvedfocalplanes(JUNO‐UVS[7]).TheyoperateatambienttemperatureswithverylowdarkcountrateandnewreadouttechnologiesallowoperationatMCPgainsfactorsof~20lowerthanbefore. The objective of this Strategic Astrophysics Technology (SAT) program is to exploit thedevelopmentsinatomic‐layer‐deposited(ALD)MCPs,photocathodesandcross‐strip(XS)readouttechniques to provide a new generation of enhanced‐performance, sealed‐tube, photon‐countingsensors that span the 115‐nm to 400‐nm regime. Efforts in subcomponent areas have achievedconsiderable technical advancement, but putting them into a robust, integrated package andadvancing the Technology Readiness Level (TRL) from 4 to 6 for the next UV‐Vis Astrophysicsinstrumentshasyettobeattempted.InthisSATprogram,weareemployingtheserecentadvancesin MCP detector technology to gain performance improvements in spatial resolution, quantumefficiency(QE),long‐termstability,environmentalstability,noisereductionandlarge‐areaformatseconomicallybyincreasingtheTRLby:

1. Implementing sealed‐tube, XS‐readout, 5‐cm format MCP detectors based on redesign of acommercialsealed‐tubepackage(Planacon)byincorporatingaceramicXSanode,ALDMCPs,andphotocathodesandwindowstoaccommodatetheUVregime.

2. IncorporatingALDMCPswithlarge‐areaformats(5‐cm×5‐cmactiveareawith10‐µmpores)intothetubes,andshowingthattheycanundergotheprocessingstepswhilemaintainingtheirperformanceandstabilitycharacteristics.

3. Integrating ceramic XS anodes into the tubes and showing that ultra‐high vacuum (UHV)processingcanbeestablishedsothatthelongevityofthefinaltubeismaximizedwhileprovidingtheperformancecharacteristicsoftheXSscheme.

4. EstablishingUV‐VisphotocathodesinthesealedtubebyadaptingthedesigntoaccommodateopaquephotocathodesontotheMCPs,andincorporatingsemitransparentUVphotocathodesonMgF2inputwindows.

5. Showingthatthesealedtubecanoperatein,andwithstand,fullflight‐likeenvironmentaltests;andthatitcanbeintegratedwithelectronicssystemsbeingdevelopedinparallelforspaceflight‐compatible,low‐power‐mass‐volume,ASIC‐readoutelectronics.

Webeganthisprojectinearly2018andhavemadeprogressoninitiatingthemajordevelopmenttasks.ThefirstsemitransparentcathodetestsareabouttocommenceaftercommissioningtheUHVfacilityandloadingMgF2windows.ALDsampleMCPshavebeentestedandanopaquealkali‐halidecathodehasbeendepositedandisreadytobeevaluated.AmeetingwiththeMCPvendorhasallowedustoestablishtherequiredfabricationspecificationsandtheinitialMCPorderforthesealedtube

49

MCPsis inplace. TheXSanodedesignwasreviewedanddiscussedwiththesealed‐tubevendor.Subsequentlyafirstbatchofanodeswasorderedandreceived,andis intest.Ameetingwiththesealed‐tubevendor facilitatedadesignreview thatrefined themechanicaldesignandprocessingsteps. The initial tube fabrication order is now in place. Meanwhile, we have commissioned anXS/MCPtestfacility(detector/electronics)andtesteditwithsample50‐mm‐format20‐µm‐poreALDMCPsandwereabletoresolvetheporesacrosstheentiredetector.

BackgroundThe2010AstrophysicsDecadal Survey recommended thatNASAprepare for aUVmission tobeconsidered by the forthcoming 2020 Astrophysics Decadal Survey. In response to thisrecommendation, the Astrophysics Visionary Roadmap of 2015 identified a large UV/Optical/IR(LUVOIR) Surveyor observatory with improvements in sensitivity, spectroscopy, high‐contrastimaging,astrometry,angularresolution,and/orwavelengthcoverageasaprimemissioncandidate.Additionally,theCosmicOrigins(COR)ProgramAnnualTechnologyReport(PATR;NASA440‐RPT‐0017)listedasaPriority1technologyrequirementtheneedtoroutinelyproducelarge‐formatUV‐sensitivedetectorarraysthatcanbeusedinavarietyofExplorer,medium,andstrategicmissionssuchasaLUVOIRtelescope.Inresponse,theStudyonApplicationsofLargeSpaceOptics(SALSO)workshophighlighted[8]possiblefuture(large‐aperture)UVastronomymissionssuchas(i)LUVOIR[9],(ii)theHabitableExoplanetObservatory(HabEx)[10]and(iii)CosmicEvolutionThroughUVSpectroscopy(CETUS)[11].Theseon‐orbitUVfacilitieswouldbeabletoprobetheverylimitsoftheuniverseasoutlinedinthescienceinvestigationsofstrategicgoal3.4oftheNASAScienceMissionDirectorateplanfor2007–2016.FutureUVdetectorsmustpossesslargerformatswithhigherspatialresolutiontotakeadvantageofthisincreasedcapabilityofthelargephoton‐collectingareaaswellaslowinherentread‐noise/ordarkrateinordertofullyexploittheverylowskybackgroundatfar‐UVwavelengths,thusenablingtheobservationofveryfaintobjects.

Inthisproject,weaimtodemonstratethatXSsealedtubeswithALDMCPscanbeimplementedasrobust, high‐performance detectors for 115 nm to 400 nm, addressing needs described in theLUVOIR/HABEX/CETUS study requirements, and for potential application to many future NASAmissions. In addition, we aim raise the TRL byminimizing themass/volume and qualifying thedetectorforspaceuse.

MCPswithimagingreadoutsinsealedtubesarenotrecentinventions.However,readouts,electronics,MCPs,photocathodes,andsealed‐tubetechniqueshaveadvancedconsiderablyinthelastfewyearsandcannowbeimplementedinahighTRL(5‐6)XSsealed‐tubesensorwithsuperiorperformancecharacteristics.Across‐strip‐sensoranodescheme(Fig.1)derivesphoton‐eventcentroidpositionsfromthechargedistributionaneventproducesacrossasetofstripsinanMCP‐amplifieddetector.TheincomingphotonproducesaprimaryelectronfromaphotocathodeontheMCPinput,orfromaphotocathode on the input window. The photoelectron is then multiplied within the pores of amicrochannel‐platepairandtheresultingelectroncloudiscollectedontwoorthogonalsetsofmetalstripsthatformtheXSanode.Toensureanaccurateeventcentroidpositionisachieved,thesizeoftheelectroncloudisoptimized

sothechargeimpingesonseveralneighboringstrips.

50

Fig.1.AnXS‐readoutMCPsealed‐tubeimaging‐sensorscheme.AphotocathodeisdepositedonawindowfacingapairofMCPs.EmittedphotoelectronsaredetectedbytheMCPs,multiplied,andcollectedbyseveralstripsineachaxisoftheanodetoenableeventcentroidpositionstobecalculated.

RecentworkonthesubcomponentsofthisdevicethroughAstrophysicsResearchandAnalysis(APRA)fundinghasresultedinsignificantadvances[12].ALDMCPswithlarge‐areaformats(12cm×12cmwith 10‐µm pores, up to 20 cm × 20 cm with 20‐µm pores), very low background rates (<0.05events/cm2/sec), extended lifetimes (> 4×1013 events cm‐2) without degradation. XS anodes andelectronicswithspatialresolutionsof<20µmFWHMover formatsof100mm×100mmandeventhandlingratesof5MHzat<15%deadtime.PhotocathodesintheUV‐Visprovide50%QEat~115nmto180nmandrangecutoffsintherange300nmto400nm(solarblindness).

Combiningthesedevelopmentshasasignificant impactonpotential futureNASAsub‐orbitalandsatelliteinstruments.Aneffectivewaytodothisisbyadoptionofacommercialdevice,thePlanacon(Fig.2),anddeveloping/integratingallthenecessaryelements(MgF2window,UVphotocathodes,ALDMCPs,andXSreadout).Thesmallerporesizes(~10µm)andhigh‐resolutionXSreadoutswillfacilitate higher spatial resolutions over the large formats. At the same time, the reduced (÷3)detectionefficiencyforhigh‐energybackgroundeventsdemonstratedbyuseofALDMCPswillalsoimproveobservationalsensitivities.Thechemicalcompatibilityof thenewMCPborosilicateglassandtheALDmaterialshasthepotentialtoprovidefurtherimprovementsinthestabilityandlifetimeof these detectors, due to the rigorous pre‐conditioning steps for sealed tubes. In addition,improvementsinfabricationprocessesprovidetheopportunitytoreducetheimagingfixedpatternmodulation.XS readouts integrated into sealed‐tubepackages canalso take full advantageof theefforts currently taking place [13] to establish high‐performance, spaceflight‐compatible, low‐power‐mass‐volume ASIC readout electronics. Together, these developments will provide asignificantstepinrealizinghigh‐performance,robustMCPdetectorsforallclassesofastrophysicsinstrumentsfromsub‐orbitaltomajorfacilityinstruments.

51

Fig.2:Top:Sealed‐tube50‐mmPlanaconwitha10‐µm‐pore ALDMCP pair and a 32×32 pad anode readout.Bottom: Cross‐section of proposed Planacon sensorscheme.MgF2 isused for inputwindow,withaUV/Vissemitransparentmultialkalicathode.ApairofALDMCPscoatedwithanopaquephoto‐cathodeprovidesenhancedUVdetectionand signalamplification,andaXSanodeaccomplishesphoton‐countingimagingreadout.

ObjectivesandMilestonesThe “commercial” Planacon (Fig. 2) [14] has a 50‐mm format that uses a square sealed‐tubeconfiguration.Thephotocathodeisasemitransparentmultialkalionastandardglasswindowwithaproximitygaptoapairof10‐µmporestandardMCPs.Thereadoutisanarrayofeither8×8or32×32pads.OurobjectivetoimplementaPlanaconscheme(Fig.2)withaUVMgF2entrancewindowandaUV‐optimizedbialkalisemitransparentphotocathodewithanarrow(~200‐µm)proximitygap.TwoALDMCPsreplacethestandardMCPs,whileanopaqueUVphotocathodewouldbedepositedontothesurfaceoftheinputMCP.ThepadarrayanodewouldalsobereplacedwithaXSanodereadout.Specificareasofdevelopmentrequiredare:

1. Obtaininghigh‐performanceALDMCPS in10‐µmporesizecompatiblewith introduction intoPlanacon devices. This entails establishment of gain value (106), uniformity, and stability asdescribedinthebackgroundinformationabove;alongwithdemonstrationofspatialresolutionat~20‐µmFWHM.

2. Implementing high‐resolution XS anodes in a refractory ceramic compatible with Planaconintegration.ThisincludesfabricatinganodescapableofwithstandingPlanacon‐tubeintegrationandprocessing.Inaddition,demonstratingthehighspatial‐resolutionandimagingperformanceat~20‐µmFWHM.

3. Optimizing and incorporating UV photocathodes into the Planacon, both semitransparent onMgF2 entrancewindows and opaque photocathodes on the ALDMCPS. Initialmilestones aredemonstration of semitransparent multialkali photocathodes on MgF2 windows and opaquealkali halides onALDMCPswithhigh efficiency and sharp long‐wavelength cutoffs (350nm).ThesemustthenbeimplementedintothePlanacondevices.

4. Demonstrating the fully implementedPlanaconXS‐detectorschemewithexistingelectronics toshow subcomponent performance characteristics; then testing them in appropriate flight‐likeenvironments(thermalandvibration),alongwitheventualradiationtesting.

ProgressandAccomplishments Sealed‐TubePlanaconImplementation

Initialworkinthedesignandtestingofthephotocathodes,MCPs,andXSreadoutsisbeingundertakenbyUCB‐SSL.MeanwhileweareworkingwithPhotonis(thePlanaconvendor)inastep‐wisemannertoincludeallof these into final sealed‐tube test articles.WeheldakickoffmeetingwithPhotonisand

52

reviewed the design, processing steps, and packaging to incorporate all the new hardware. ThedimensionsfortheanodeshavebeenestablishedandweareprocuringceramicanodeblanksformakingtestsealsofceramicanodesubstratestoPlanaconbodiesandthenverifyingthehigh‐vacuumseal.ThedistancebetweentheanodeandtheMCPstackinthedesignissetat2.5mmandisoptimizedforthechargecloudspreadoftheMCPconfiguration.ThenaninitialtubeconstructioneffortwillincorporateaceramicXSanodeintoaPlanaconwitha10‐µmMCPpairandabialkaliphotocathodewithasapphireplanar UV window. The objective is to get an initial process run that establishes the ALD MCPenvironmental(UHV)processingregimeincludingahigh‐temperaturebakeoutandasubsequent“burn‐in”beforethephotocathodeisdeposited.Secondly,weaimtoshowleak‐tightsealingofthetubesothatwecanconfirmphotocathodeperformance,aswellasXSandALDMCPoperation, insupportof theseconditerationofsealedtubes.WeestablishedanorderfortheinitialsetuptestsatPhotonisandforprocessingthefirsttubeiterationthatisexpectedtobeaccomplishedbytheendof2018.

UV‐VisPhotocathodesfor115nmto≥400nm

The photocathode and photoelectron detection by the MCP determine the MCP’s QE [15]. Thephotocathode can be deposited on a window directly in front of the MCP (a proximity‐focusedsemitransparent cathode, Fig. 2)ordirectlyonto theMCP (opaquemode) [16‐19].Opaque alkalihalidephotocathodesarewidelyused forextreme‐UV (EUV)/UVsensors, asare semitransparentmultialkali photocathodes for visible detectors. Newmaterials exemplified by GaN [12] are alsocandidatesforfar‐UV/near‐UVphotocathodes.

Alkalihalide(CsI,KBr)opaquephotocathodesonMCPsobtainhighQE(50%@~110nm)andhavebroadbandsensitivityfrom10nmto160nm[20].WehavepreviouslyusedthesephotocathodesinsealedtubesandachievedgoodperformanceonearlyALDMCPs.ToestablishtheprocessforusingthemasopaquecathodesinthePlanacontubes,weobtainedsampleALDMCPsfortrialdepositionsandbeganbydepositingaCsIphotocathode(Fig.3),soontobetestedforefficiencyanduniformity.

Fig.3.CsIopaquephotocathodedepositedontoa20‐µmpore33‐mm,60:1L/d,ALDMCPwithMgOand13°porebias.

PhotonisiscontinuouslyproducingmultialkaliandbialkaliphotocathodesfortheirexistingPlanacondevices,soarefullyequippedtoimplementthem.However,tomeetLUVOIRgoalsforexample,wewishto optimize for UV response and long‐wavelength cutoff, and minimize the intrinsic dark rate.Traditionalvisible‐sensitivephotocathodeslikebialkali(Na2KSb)havea30%peakat350nmbutalsoshowhighQEintheUV[21]thathasnotbeenpreviouslyaccessedforUVsensors.InourexistingUCBfacilities,wehaveproducedmanysemitransparentbialkaliphotocathodes[22]upto20‐cmformatwithpeakQEsof25%at350nm.Wejustcompletedsettingupaprocesstankforbialkali‐photocathode(Na2KSb)depositionontoMgF2substrates,tosettheprocessandthicknessthatoptimizestheQEbelow300nmwhileobtainingasharpcutoffabove400nm.Thegeneralpracticeistomakethecathodesthin,becausethehighattenuationcoefficient forblue lightproducesphotoelectronsclose to thesurfacewheretheycanescape,butallowsredlighttopassthroughwithoutinteraction.Initialtestsareshown

53

inFig.4forthinbialkaliphotocathodesmadeonMgF2windows.AlthoughthepeakQEisnotoptimizedyet,wedemonstratedagoodcutoffat350nmwiththemostrecentdeposition.

Fig.4.QEforblue‐optimizedsemitransparentbialkaliphoto‐cathodesonaMgF2window.

WesubsequentlypreparedanewsetofthreeMgF2windowsandinstalledthemintotheprocesstanktopermitafurtheriterationoftheprocess,intendingtooptimizethe“in‐band”QE.Thesystemhasbeensuccessfullypreconditioned(high‐vacuumbake)and is ready forphotocathodedepositions.Followingtheestablishmentofabaseline,wewouldworkwithPhotonistoadjusttheirphotocathodedepositionprocesstosuit.

ALDMicrochannelPlates

A new type of MCP (Fig. 5) constructed by ALD on borosilicate microcapillary arrays, usinginexpensive hollow borosilicate tubes, has recently been developed. The fabrication method issomewhatsimilartothatofconventionalMCPs[23],butwithouttheneedtoremoveacoreglassinsidethetubes.ResistivelayersandsecondaryemissivelayersaredepositedusingALD,replacingthehydrogenreductionusedforconventionalMCPactivation.Borosilicatemicrocapillarysubstratesare very robust comparedwith standardMCP glass. MCPs are regularlymadewith 20‐µm‐poreborosilicateupto20cm,and10‐µmpore,10‐cmand5‐cmsquareformatswith60:1L/dratio.Thegainuniformityisgoodevenoverthelargestareas[24].ThebackgroundeventimagedistributionofALDMCPs isuniformand isvery low‐rate,0.03events/cm2/sec1 [25] (>5x lower thanstandardMCPs). Detection efficiency for MeV gamma rays is a potentially important factor in reducinginstrumental background rates in high‐radiation environments. For ALD MCPs, the efficiency is~0.7%[25]ascomparedwith~2%forconventionallead‐glassMCPs.

Fig.5.Borosilicateglassmicrocapillarysubstratesarefunctionalized with ALD. Separate sets of ALDdepositionssettheresistancesandsecondaryemission(Al2O3andMgO)layerstocreateaworkingMCP.

WemetwiththeMCPvendor,IncomInc.,toestablishMCPspecificationsandprocessing.TheALDMCPs will be 53.5 mm square using 10‐µm pores, 60:1 L/d, 13° pore bias and MgO secondaryemissivelayers,whichisastandardMCPconfigurationforthePhotonisPlanacon.borosilicateALDMCPsfromIncomalsocomestandardwith10‐µmporesubstrates.Theycanbemadeinavarietyof

54

sizes,howeverthemostconvenientschemeforthePlanaconMCPsistousethestandard‐sizeblockmaterial and thendice into the final 53‐mm format.Thenominalopenarea for the substrates is~75%, which maximizes photoelectron detection efficiency for semitransparent photocathodephotoelectrondetection.Thegoal forMCP resistance is in the range20 ‐30MΩ, inkeepingwithmodelsforacceptablepowerdissipation,aswellashighlocaleventratehandling.Wealsojusttestedsome10‐µmporesampleMCPstoestablishexpectationsforMCPsforthePlanacontrials.Theresults(Fig. 6) show good gain characteristics, allowing Incom to proceedwith construction of full‐sizeMCPs.Weplacedanorderforthefirstsetsof10‐µmporeMCPsforthefirstPlanaconbuild.

Fig. 6. Gain is high and repeatable for 33‐mm‐diameterborosilicateALDMCPswith10µmpore,74%openarea,60:1L/DwithaMgOemissivelayer.

XSReadoutsforSealedTubes

TheXSanodeitselfismadewithlayersofmetalsandinsulatorsonaninsulatingsubstrate,producingtwosetsofstripsinorthogonaldirections.Thetopandbottomlayershaveequalexposedareas,andthus collect the charge from the MCPs with equal charge sharing (Fig. 1). Each anode strip isconnectedtothebacksideoftheanodeusinghermeticallysealedvias,allowingmountingofallthedetectorelectronicsoutsidethevacuum.Inthiscase,weemployedceramicsforthesubstratesandinsulating layers using standard thick‐film technologies. Ceramic XS anodes are low‐outgassing,refractory,andaccommodatebeingputintosealed‐tubeultra‐high‐vacuumdevices.Weadopteda47‐mmXS,72×72strip‐anode(Fig.7)designthatiscapableofhighresolution[26]withgoodimagelinearity.Theupperandlowerorthogonalfingersoccupyapproximately50%ofthearearespectively(Fig. 8) and are separated by insulating glass layers. The first batch of anodes has already beenprocured and was recently delivered. These are currently undergoing tests, but initial benchmeasurements show that the finger isolation and the inter‐capacitances (< 10 pF) are allwithinexpectations. In parallel,we are obtaining ceramic blanks so that Photonis can seal these anodeblankstothecurrentproductionbodiesbeforeusingtheactualXSanodes.

Fig.7.Cross‐stripanodewith47‐mmactiveareaonceramic substrate for inclusion in a high‐vacuumPlanaconsealed‐tubepackage.

55

Fig.8.Sectionononecornerofthe47‐mmXSanodeshowingtheupperandlowerfingersofthecrossstripsthatareexposedforeventchargecollection.

ComponentTesting

Wehave established cross‐delay‐line and cross‐strip (Fig. 9) readout anode detectors in 50‐mmformat,andreadoutelectronicsforphoton‐countingtestsofMCPsandXSanodes.ThecurrentPXS‐II(Fig.10)encodingelectronicshasalsobeencommissionedforevaluationandtestingofXSanodes,andtoprovidethehigh‐resolutionimagingforALDMCPtesting.Weinstalledsample53‐mmALDMCPswith20‐µmporesintothedetectortoverifythesystemisreadytotestthefinal10‐µmMCPswhentheyaredeliveredtous.TheresultsareshowninFigs.11and12.TheimagingissufficienttoresolveeveryMCPpore,andthegainfortheseMCPsisveryuniformacrossthefieldofview.Theperformanceofthistest‐bedissufficientforallourevaluationsofthesealed‐tubesub‐componentsasweaccumulatethemandproceedtointegratethefinaldevices.

Fig. 9. 50‐mm active area, cross‐strip‐anode imagingdetector for50‐mmMCPandXSanode tests.Planaconstyledetectorbody.

Fig.10.PXS‐IImainelectronicssystem;providingphotonpositionto16k×16kaccuracy(achieved<20‐µmFWHMspatial resolution with XS detectors at MCP gains of~2×106atMHzrates.

Fig.11. Imagewith cross‐stripanodedetector fora50×50mm2 cross‐strip readout detector at ~2×106gain(pairof20‐µm‐poreMCPs).

56

Fig.12.MagnifiedareaofFig.11showsitiscapableofveryhighresolution(~5µm/pixel)with thecross‐strip detector. Individual 20‐µm pores are clearlyvisible.

PathForwardThenext steps include receipt anddispatchof theanodeblanks toPhotonis for vacuum‐sealing‐processverification.Meanwhile,wewillcompletetheverificationtests,bothfunctionalandphysical,of the 47‐mm XS anodes we received, and make any adjustments needed to ensure they arecompatiblewiththePlanaconstructure.ThispavesthewaytosubsequentlysendthemtoPhotonisforsealingtoPlanacontubebodies.ALDMCPsin10‐µmpore,13°‐biasmaterialwithMgOemissivelayersareonorder,butwillnotarrivebeforefall2018.Inthemeantime,wewillbetestingalltheALDMCPperformancecharacteristicsof33‐mmsampleMCPstoverifyALDprocessingconditionsandmakerecommendationsforthefinalprocessingsequencefortheMCPstobeincorporatedintothesealedPlanacontubes.Whenthefull‐size,10‐µmMCPsdoarrive,wewilltestthemandthendicethemtotherequiredsize.Then,wewilltesttheperformanceofthe53‐mmMCPsinpairs,including,

57

gain,pulse‐heightdistribution(PHD),background,uniformity,etc.inourhigh‐resolutionopen‐faceXSdetectorsystem,anddispatchtoPhotonis forPlanacontubeintegration.Inconcertwiththeseefforts, we are poised to test both opaque‐alkali‐halide and semitransparent‐photocathodeprocesses.ThesewilltaketheformofanumberofprocessiterationswherewemeasuretheQEandbackground,aswellastheeffectonMCPimagingcharacteristics.ItisnotanticipatedthatthefirstPlanacondevicewillusethemodifiedphotocathodeprocess.Rather,theinitialPlanaconsealedtubewillbedirectedat includingtheXSanodeelementandALDMCPs, formeasuringitsperformancecomparedtostandarddevices.AsignificantmilestoneistocompletethefirstPlanacondeviceandcarryoutthebasicfunctionaltests(gain,QE,background,imaging,etc.).

References[1] O.H.W.Siegmund,M.A.Gummin,J.M.Stock,G.Naletto,G.A.Gaines,R.Raffanti,J.Hull,R.Abiad,T.

Rodriguez‐Bell,T.Magoncelli,P.N.Jelinsky,W.Donakowski,andK.E.Kromer,“PerformanceofthedoubledelaylinemicrochannelplatedetectorsfortheFar‐UltravioletSpectroscopicExplorer,”Proc.SPIE,vol.3114,pp.283–294(Oct.1997)

[2] J.Vallerga,J.B.McPhate,A.P.Martin,G.A.Gaines,O.H.W.Siegmund,E.Wilkinson,S.Penton,andS.Beland,“HST‐COS far‐ultravioletdetector: finalgroundcalibration,”Proc.SPIE,vol.4498,pp.141–151(Dec.2001)

[3] P.N. Jelinsky, P.F. Morrissey, J.M. Malloy, S.R. Jelinsky, O.H.W. Siegmund, C. Martin, D.Schiminovich,K.Forster,T.Wyder,andP.G.Friedman,“PerformanceresultsoftheGALEXcrossdelaylinedetectors,”Proc.SPIE,vol.4854,pp.233–240(Feb.2003)

[4] J.Green,J.Bally,R.Brown,D.Ebbets,W.Freedman,J.Grunsfeld,J.Huchra,S.Kilston,R.Kimble,J. Morse, R. O'Connell, K. Sembach, K. Shull, O.H.W. Siegmund, and E. Wilkinson, “ScientificRationale fora10meterUV‐OpticalTelescope,” presented at theAstronomical Society of thePacificConferenceSeries,vol.348,p.559(2006)

[5] T.J.Immel,S.B.Mende,R.A.Heelis,C.R.Englert,J.Edelstein,J.M.Forbes,S.England,A.I.Maute,J.J.Makela,F.Kamalabadi,G.Crowley,A.W.Stephan,J.D.Huba,J.Harlander,G.R.Swenson,H.U.Frey,G.S.Bust,J.M.Gerard,B.A.Hubert,D.E.Rowland,D.L.Hysell,A.Saito,S.Frey,M.Bester,andC.E. Valladares, “ICON:The IonosphericConnectionExplorer ‐NASA'sNext SpacePhysicsandAeronomyMission,”AGU,vol.42,p.03(Dec.2013)

[6] R.Eastes,M.Codrescu,W.McClintock,A.Aksnes,D.Anderson,L.Andersson,A.Burns,S.Budzien,R.Daniell,K.Dymond,F.Eparvier,J.Harvey,T.Immel,A.Krywonos,M.Lankton,J.Lumpe,G. Prolss,A.Richmond,D.Rusch,O.H.W.Siegmund,S.Solomon,D.Strickland,andT.Woods,“Global‐scaleObservationsoftheLimbandDisk(GOLD) ‐NewObservingCapabilities forSpaceWeatherSpecificationandForecasting,”AGU,vol.13,p.1085(Dec.2007)

[7] R.Gladstone,M.Versteeg,T.K.Greathouse,V.Hue,M.Davis,J.‐C.Gerard,D.Grodent,andB.Bonfond,“JunoUltravioletSpectrograph(Juno‐UVS)ObservationsofJupiterduringApproach,”AAS,vol.48,p.#402.02(Oct.2016)

[8] P.A.Scowen,T.Tripp,M.Beasley,D.Ardila,B.G.Andersson,J.M.Apellaniz,M.Barstow,L.Bianchi,D.Calzettietal.,“FindingtheUV–VisiblePathForward:ProceedingsoftheCommunityWorkshoptoPlantheFutureofUV/VisibleSpaceAstrophysics,”PASP129(977),076001,IOPPublishing(2017)

[9] K.France,B.Fleming,G.West,S.R.McCandliss,M.R.Bolcar,W.Harris,L.Moustakas,J.M.O'Meara,I. Pascucci et al., “The LUVOIR Ultraviolet Multi‐Object Spectrograph (LUMOS): instrumentdefinitionanddesign,”Proc.SPIE10397,1039713,(2017)

[10]P.A.Scowen,D.Stern,R.Somerville,M.Rud,S.Martin,andM.Beasley,“ScienceandarchitecturedriversfortheHabExUltravioletSpectrograph(UVS),”Proc.SPIE0398,1039807(2017)

[11]S.HeapandtheCETUSTeam,"TheUVSurveyMissionConcept,CETUS,"AmericanAstronomicalSociety,AASMeeting#231,id.#355.30(2018)

[12]C.Ertley,O.H.W.Siegmund,A.Tremsin,andJ.Hull,“MicrochannelPlateImagingDetectorsforHighDynamicRangeApplications,”IEEETrans.Nucl.Sci.,vol.PP,no.99,pp.1–1(2017)

58

[13]J. Vallerga, R. Raffanti, M. Cooney, H. Cumming, G. Varner, and A. Seljak, “Cross strip anodereadouts for large format, photon counting microchannel plate detectors: developing flightqualifiedprototypesofthedetectorandelectronics,”TWEPP,pp.91443J–91443J–12(Jul.2014)

[14]Photonisonlinecatalog,"Planacon"[15]O.H.W.Siegmund,J.B.McPhate,J.Vallerga,A.Tremsin,S.R.Jelinsky,andH.J.Frisch,“Novellarge

formatsealedtubemicrochannelplatedetectorsforCherenkovtimingandimaging,”vol.639,no.1,pp.165–168(May2011)

[16]S.R.Jelinsky,O.H.W.Siegmund,andJ.A.Mir,“Progressinsoftx‐rayandUVphotocathodes,”Proc.SPIE,vol.2808,pp.617–625(Oct.1996)

[17]A.TremsinandO.H.W.Siegmund, “UVradiationresistanceandsolarblindnessofCsIandKBrphotocathodes,”IEEETrans.Nucl.Sci.,vol.48,no.3,pp.421–425(2001)

[18]J.I.Larruquert,J.A.Méndez,J.A.Aznárez,A.Tremsin,andO.H.W.Siegmund,“Opticalpropertiesandquantumefficiencyofthin‐filmalkalihalidesinthefarultraviolet,”Appl.Opt.,vol.41,no.13,pp.2532–2540(May2002)

[19]A. Tremsin andO.H.W. Siegmund, “Thequantum efficiencyand stabilityofUVand softX‐rayphotocathodes,”Proc.SPIE,vol.5920,pp.I–13,Aug.2005.

[20]O.H.W.Siegmund,J.Vallerga,andA.Tremsin,“Characterizationsofmicrochannelplatequantumefficiency,”Proc.SPIE,vol.5898,pp.113–123(Aug.2005)

[21]P.Dorenbos,J.T.M.deHaas,R.Visser,C.W.E.vanEijk,andR.W.Hollander,“QuantumefficienciesofseveralVUV‐sensitivephotomultipliertubes,”Nucl.Instr.Meth.Phys.Res.A,vol.325,no.1,pp.367–369(Feb.1993)

[22]O.H.W.Siegmund,J.B.McPhate,S.R.Jelinsky, J.Vallerga,A.Tremsin,R.Hemphill,H.J.Frisch,R.G. Wagner, J. Elam, and A. Mane, “LargeAreaMicrochannel Plate ImagingEvent CountingDetectorswithSub‐NanosecondTiming,”IEEETrans.Nucl.Sci.,V60(2),pp.923–931(Apr.2013)

[23]M.Lampton,“TheMicrochannelImageIntensifier,”ScientificAmerican,vol.245(5),pp.62(1981)[24]O.H.W.Siegmund,J.Vallerga,A.Tremsin,J.McPhate,H.Frisch,J.Elam,A.Mane,R.Wagner,and

G.Varner,“LargeAreaandHighEfficiencyPhotonCountingImagingDetectorswithHighTimeandSpatialResolutionforNightTimeSensingandAstronomy,”AMOSConf.p.92(2012)

[25]O.H.W.Siegmund,C.Ertley,C.A.Craven, J.Vallerga,M.Popecki,A.O'Mahony, andM.J.Minot,“HighSpeedLargeFormatPhotonCountingMicrochannelPlate ImagingSensors,”ProcAMOS,vol.1,p.94(2015)

[26]O.H.W.Siegmund,A.Tremsin,andJ.Vallerga,“DevelopmentofcrossstripMCPdetectorsforUVandopticalinstruments,”Proc.SPIE,vol.7435,pp.74350L–74350L–10(Aug.2009)

Foradditionalinformation,contactOswaldSiegmund:ossy@berkeley.edu

59

AdvancedUVOIRMirrorTechnologyDevelopmentforVeryLargeSpaceTelescopesPreparedby:H.PhilipStahl(NASAMSFC)

SummaryThe Advanced Mirror Technology Development (AMTD) project is in Phase 2 of a multiyear effort initiated in Fiscal Year (FY) 2012, to mature the Technology Readiness Level (TRL) of critical technologies required to enable 4-m-or-larger monolithic or segmented ultraviolet, optical, and infrared (UVOIR) space telescope primary-mirror assemblies for general astrophysics, ultra-high-contrast observations of exoplanets and National Interest missions. AMTD-2 is jointly funded by NASA and the Air Force.

In 2012, AMTD-2 defined three major milestones. During 2016/17 AMTD-2 successfully accomplished half of these:

Harris Corp fabricated a ~150 Hz 1.5-meter Ultra-Low Expansion (ULE®) mirror (1/3rd scale of a 4-m mirror) substrate using stacked core method, demonstrating lateral scalability of the stacked core technology.

o NASA MSFC characterized the mechanical and thermal performance of the Schott owned 1.2-meter Zerodur® mirror (polished and integrated into an assembly by Arizona Optical Systems).

o NASA MSFC validated by test its integrated STOP model prediction of the 1.2m Zerodur® mirror’s mechanical and thermal performance; including verification of CTE homogeneity.

The balance is scheduled for completion in the fourth quarter of FY17:

o Characterizing the mechanical and thermal performance of the Harris 1.5-m ULR® mirror. o Validating by test MSFC’s integrated STOP model prediction of the 1.5-m ULE® mirror’s

mechanical and thermal performance; this testing also includes verification of CTE homogeneity. Additionally, lessons learned about low-temperature-slumping (LTS) of large stiff mirror substrates from fabricating the 1.5m ULE® mirror are being incorporated into various current and planned NASA and National Interest missions. An AMTD-3 proposal to resolve the issues was submitted to the ROSES 2016 SAT. Finally, NASA MSFC is using integrated design tools developed on AMTD-2 to design candidate optical telescope assemblies for the potential Habitable Exoplanet Imager mission.

The foundation of AMTD’s success continues to be its integrated team of Government/Industry scientists, systems engineers, and technologists executing a science-driven systems engineering approach to technology development. Additionally, AMTD continues mentoring the next generation of scientists and engineers via interns, co-ops, and volunteers. In this cycle AMTD hosted two undergraduate student interns: Samantha Hansen of Rutgers University and Mary Elizabeth Cobb of University of Alabama in Huntsville.

AMTD results were presented at Mirror Tech Days 2016 and published in proceedings of the 2016 SPIE Astronomy Conference1 and 2017 SPIE Optics and Photonics Conference2,3 (see list at report end).

Background“Are we alone in the Universe?” is probably the most compelling science question of our generation.

Per the 2010 New Worlds, New Horizons Decadal Report4: “One of the fastest growing and most exciting fields in astrophysics is the study of planets beyond our solar system. The ultimate goal is to image rocky

60

planets that lie in the habitable zone of nearby stars.” The Survey recommended, as its highest priority, medium-scale activity such as a “New Worlds Technology Development (NWTD) Program” to “lay the technical and scientific foundations for a future space imaging and spectroscopy mission.” The National Research Council (NRC) report, NASA Space Technology Roadmaps & Prioroties5, states that the second highest technical challenge for NASA regarding expanding our understanding of Earth and the universe in which we live is to “Develop a new generation of astronomical telescopes that enable discovery of habitable planets, facilitate advances in solar physics, and enable the study of faint structures around bright objects by developing high-contrast imaging and spectroscopic technologies to provide unprecedented sensitivity, field of view, and spectroscopy of faint objects.” NASA’s Enduring Quests Daring Visions6 called for a LUVOIR surveyor mission to “enable ultra-high-contrast spectroscopic studies to directly measure oxygen, water vapor, and other molecules in the atmospheres of exoEarths,” and “decode the galaxy assembly histories through detailed archeology of their present structure.” As a result, NASA will study in detail a LUVOIR surveyor and a HabEx Imager concept for the 2020 Decadal Survey.7,8 Additionally, AURA’s From Cosmic Birth to Living Earths9 details the potential revolutionary science that could be accomplished from “directly finding habitable planets showing signs of life.”

Directly imaging and characterizing habitable planets requires a large-aperture telescope with extreme wavefront stability. For an internal coronagraph, this requires correcting wavefront errors (WFEs) and keeping that correction stable to a few picometers rms for the duration of the science observation. This places severe specification constraints upon the performance of the observatory, telescope, and primary mirror. One important problem is dynamic WFE. For either large monolithic mirrors (for HabEx) or smaller mirror segments (for LUVOIR), mechanical disturbances create rigid-body motion of the mirror on its mount. These inertial motions introduce dynamic WFE when the mirror distorts (or bends) as it reacts against its mount. Achieving wavefront stability is a systems engineering trade between mirror stiffness and vibration isolation. Per Lake,10 WFE is proportional to rms magnitude of the applied inertial acceleration divided by the square of the structure’s first mode frequency. Therefore, to achieve <10 pm rms requires either a very stiff system or very low acceleration loads. It is easier to improve performance by increasing stiffness rather than increasing isolation. For a given stiffness mirror, a 10× reduction in acceleration results in a 10× WFE reduction. For a given acceleration level, a 10× increase in stiffness results in a 100× WFE reduction. While systems engineering analysis is still preliminary, the HabEx program estimates that it needs a 4-m monolithic primary mirror with a first mode frequency greater than 100 Hz and maybe as high as 200 Hz. The LUVOIR program estimates that it needs 1.5-m segments with a first mode frequency greater than 220 Hz (JWST) and maybe as high as 500 Hz.

Before AMTD, the state of art for lightweight stiff glass mirror substrates was defined by AMSD and ATT. For their sizes (AMSD 1.5m × 50mm; ATT 2.5m × 150mm) both mirrors have >150 Hz first mode required for stable on-orbit mechanical performance. But neither has sufficient structural depth for >100 Hz at 4m. The easiest way to increase stiffness is to increase thickness. AMTD-1 advanced TRL by successfully demonstrating a new process (stacked-core low-temperature fusion) that extended the previous state of art (SOA) for deep-core substrates from <300 mm to >400 mm. This was done by making a 43-cm diameter full-scale ‘cut-out’ of a 4-m mirror. AMTD-2 continued this advance by making a 450 Hz, 1.5-m x 165-mm mirror.

ObjectivesandMilestone:AMTD’s objective is to mature towards TRL6 technologies to enable large monolithic or segmented UVOIR space telescopes. Phase 1 advanced technology readiness of six key technologies required to make an integrated primary mirror assembly (PMA) for a large aperture UVOIR space telescope.

Large-Aperture, Low-Areal Density, High-Stiffness Mirror Substrates Support System Mid/High-Spatial Frequency Figure Error Segment Edges

61

Segment-to-Segment Gap Phasing Integrated Model Validation

Phase 2 is continuing the efforts in High-Stiffness Substrates, Support Systems, Segment-to-Segment Gap Phasing and Integrated Model Validation with three clearly defined milestones:

Fabricate a ⅓-scale model of 4-m class 400-mm thick deep-core ULE® mirror substrate to demonstrate lateral scaling of the deep-core process. (Successfully Completed in 2016.)

Characterize two candidate primary mirrors (the ⅓-scale mirror and a 1.2-m Extreme Lightweight Zerodur® Mirror owned by Schott) by measuring their modal performance and optical performance from 250 K to ambient. (Schott 1.2-m Zerodur® mirror successfully characterized in 2016; 1.5-m ULE® Harris mirror is schedule for characterization in 2017.)

Add capabilities and validate integrated design and modeling tools to predict the mechanical and thermal behavior of the candidate mirrors, validate models, generate Pre-Phase-A point designs, and predict on-orbit optical performance. (Schott 1.2-m Zerodur® mirror mechanical and thermal performance models were successfully validated by test in 2016; 1.5-m ULE® Harris mirror models are schedule for validation by test in 2017; Additionally, NASA MSFC is using integrated design and modeling tools on the potential Habitable Exoplanet Imager mission.).

ProgressandAccomplishments:Large-Aperture, Low-Areal-Density, High-Stiffness Mirror Substrates

Need: To achieve ultra-stable mechanical and thermal performance required for high-contrast imaging, either (4-m to 8-m) monolithic or (8-m to 16-m) segmented mirrors require thicker, stiffer mirrors.

Milestone #1: Fabricate a ⅓-scale model of 4-m class 400-mm thick deep-core ULE® mirror substrate to demonstrate lateral scaling of the deep-core process.

Accomplishment: During FY 2016/17, AMTD-2 progressed this technical area by fabricating a 450Hz 1.5m substrate to accomplish its defined milestone and learning lessons from turning that substrate into a mirror.

High-Stiffness Mirror Substrates: Previously, AMTD-1 advanced TRL by successfully demonstrating the ability to make a 40-cm thick subscale mirror substrate via the stacked-core low-temperature fusion (LTF) process. This extended the previous SOA for deep-core substrates from <300 mm to >400 mm. This was done by making a 43-cm diameter 40-cm thick full-scale ‘cut-out’ of a 4-m mirror. In 2016, AMTD-2 achieved its major milestone for this technology when Harris Corp successfully LTF’ed a 1.5-m diameter by 165 mm thick 5-layer ULE® mirror substrate with a 450 Hz first mode frequency. In 2016/17 Harris Corp low-temperature-slumped (LTS) the mirror substrate to a 3.5 meter radius of curvature then ground, polished, coated it with protective aluminum and integrated it into a flight-like mount (Figure 1). In 2017 the mechanical and thermal performance of this mirror assembly will be characterized at NASA MSFC.

62

Figure 1: (left) Low-Temperature-Fused Mirror Substrate; (top right) 5-layer design; (bottom right) completed mirror assembly.

During AMTD-1 when the 43-cm deep-core mirror was slumped from 5.0 to 2.5-meter radius of curvature, there was noticeable deformation in the core walls. To quantify the magnitude of this bending MSFC imaged the mirror’s internal structure via x-ray tomography. A small amount of bending was expected because slumping places the concave surface in compression and stretches the convex surface; this places the core elements in shear stress. The measured deformation exceeded that expectation. Fortunately, analysis indicated that such core-wall bending had a limited effect on the mirror’s strength.

In designing the 1.5-meter 1/3rd scale model of a 4-meter mirror, Harris Corp used proprietary modeling tools to predict the visco-elastic performance of the mirror (Figure 2). The spacing between the wedge-shaped core elements was specifically increased to prevent adjacent core-walls from touching. While the core walls never touched, they did get within <0.25 mm at four locations. (Figure 2). In 2017 MSFC plans to image the internal structure of the mirror via X-Ray computed tomography, use that data to create an as-built 3D model of the mirror to predict its mechanical and thermal performance, and characterize its performance.

Figure 2: (left) predicted visco-elastic deformation used to design mirror substrate; (right 3 images) actual visco-elastic deformation; 4-locations had gaps of less than 0.25 mm.

Explaining core-wall bending is complicated. Previous to AMTD, the only mirrors fabricated via LTS replication were AMSD/MMSD. Neither of these mirrors exhibited core-wall bending. Preliminary analysis indicates that the effect depends upon the shear stress in the radial core-walls. The greater the amount of shear stress, the greater the amount of viscous flow of the glass during LTS replication and the greater the core-wall bending. Preliminary analysis indicates that this shear stress is proportional to the unsupported radial core wall length divided by the radius of curvature (independent of core thickness and independent of whether the core is composed of a single layer or multiple layers). The AMTD-1 2.5-m ROC 0.43-m diameter x 400-mm thick mirror had significantly larger core cells than the AMTD-2 3.5-m ROC 1.5-m diameter x 165-mm thick mirror; and thus, less bending. An AMTD-3 proposal to resolve this issue was submitted to the ROSES 2016 SAT.

Arnold Mirror Modeler: the Arnold Mirror Modeler was developed under AMTD to rapidly create and analyze detailed mirror designs. The AMM creates a complete analysis stream, including model, loads [static and dynamic], plots and a summary file of input variables and results suitable for optimization or trade studies. Values of all settings can be archived and recalled to continue or redo any configuration. In FY16/17 an updated AMM was released and is being used to perform point design trade studies for HabEx closed-back ULE® and open-back Zerodur® mirror systems.

Support System

Need: Large-aperture mirrors require large support systems to ensure they survive launch and deploy on-orbit, stress-free, and undistorted. Additionally, segmented mirrors require large structure systems that establish and maintain the mirror’s shape.

63

Accomplishment: During FY 2016/17, AMTD-2 progressed this technical area by continuing to develop mount capabilities in the Arnold Mirror Modeler and use it to perform trade studies of candidate mirror mount systems for HabEx. The emphasis of these trade studies has been on understanding and specifying dynamic primary mirror wavefront error.

Dynamic WFE is produced when a mirror is placed in inertial acceleration by a mechanical disturbance, causing it to react (i.e. bend) against its mounts. A ‘static’ example is gravity sag. The acceleration of gravity causes a mirror to bend on its mount. Assuming that no resonant mode is excited, a mirror’s dynamic WFE has the same shape as its gravity sag with an amplitude proportional to the disturbance’s ‘G-acceleration’. Assuming that the mirror substrate’s first mode stiffness is higher than the mechanical disturbance frequencies, the biggest accelerations occur when the mechanical disturbance excites a mount resonance mode. AMTD-2 has studies dynamic WFE for various mirror substrates on both 3-point and 6-point mounts attached to the substrate at the edge, 80% and 50% radial points (Figure 3).

Figure 3 Static gravity sag and dynamic deformation of 180-Hz 4-m diameter closed-back mirror on 3-point mount attached at edge, 80% and 50% radial locations.

Mid/High-Spatial Frequency Error:

Need: High-contrast imaging requires mirrors with very smooth surfaces (< 10 nm rms). While deformable mirrors can correct low-order errors, they cannot correct mid and high-spatial frequency errors. Such errors can arise from the fabrication process or CTE inhomogeneity and can introduce artifacts into the dark hole.

Milestone #2: Characterize thermal induced surface figure error of two candidate primary mirrors (1.5-m ULE® mirror and 1.2-m Zerodur® mirror owned by Schott) from 250 K to ambient.

Accomplishment: During FY 2016/17, AMTD-2 progressed this technical area by characterizing the thermal performance of the 1.2-m Extreme Lightweighted Zerodur® Mirror (ELZM) owned by Schott. This test accomplishes half of this major milestone. Anticipated testing of the 1.5-m ULE® mirror in 2017 will complete this milestone. Previously, AMTD-1 had demonstrated the ability of the Harris Corp’s ion polishing process to produce a 5.4-nm rms surface.

Thermal Characterization: In 2016/17 NASA MSFC enhanced its 23m x 7m thermal vacuum test chamber by adding a pressure tight enclosure that allows test equipment to be placed at the center of curvature of short radius of curvature mirrors (Figure 4). This new capability enabled testing of the

64

Schott 1.2-m ELZM mirror from ambient to 250K. No thermal deformation induced high-spatial quilting associated with the lightweight was measured. The test did measure 9.4 nm rms of mid-spatial error associated with coefficient of thermal expansion (CTE) inhomogeneity (Figure 4).

Figure 4: (left) Short radius of curvature mirror test setup with pressure tight enclosure. (center) Schott ELZM Mirror in test setup. (right) Measured thermal surface figure error from 250K to ambient.

Segment-to-Segment Gap Phasing

Need: To avoid speckle noise which can interfere with exoplanet observation, internal coronagraphs require an ultra-stable wavefront.

Accomplishment: During FY 2015/16, AMTD Phase 2 progressed this technical area by continuing the systems engineering effort to understand the interaction between optical telescope wavefront stability and coronagraph contrast leakage.

Contrast Leakage WFE Tolerances: In our previous studies, we evaluated the contrast leakage over specific regions of interest (ROI). These studies showed a correlation between numbers of segment rings and noise in a dark-hole ROI. These studies also provided preliminary tolerance values for the allowed amounts of different wavefront errors. Because of the asymmetric nature of the ROI, the tolerance results were misleading. Therefore, in FY2016/17 we implemented a new method that decomposes contrast leakage into average radial (photometric noise) and azimuthal (systematic noise) components as defined by Shaklan11. The new method confirms our previous conclusions that segment-to-segment co-phasing (piston and tip/tilt) errors must be stable in the 10 to 20 picometer PV range. The new method was used to evaluate the contrast leakage in annular ROI for 50 random trials global Seidel and Zernike aberration produced by rigid body motion of the primary and/or secondary mirror assemblies. The maximum allowance for static aberration and contrast leakage for aberration exhibiting sinusoidal variation was also studied. Table 1 lists the maximum amount of random WFE that a 4th order radial coronagraph can tolerate while keeping the photometric noise less than 10-10 and the systematic noise less then 5x10-11 over an annular ROI from 1.5 to 2.5 λ/D.

Table 1: PV Aberration Amplitude Tolerance for Contrast Leakage over an annular ROI from 1.5 to 2.5 λ/D

Aberration (Random)

WFE (pm) for 1x10-10 of Photometric Noise

WFE (pm) for 5x10-11 of Systematic Noise

Tip/Tilt 9,600 35,000 Seidel Power 1,100 22,000

Zernike Astigmatism 6,800 49,000 Zernike Trefoil 6,800 44,000

Zernike Hexafoil 9,600 78,000 Seidel Spherical 300 11,000

Seidel Coma 6,800 840

65

Clearly, photometric noise (radial average) is more sensitive to Seidel spherical than Seidel power. It is equally apparent that systematic noise is more sensitive to rotationally asymmetric aberrations, such as Seidel coma, than to rotationally symmetric aberrations like Zernike astigmatism.

Integrated Model Validation

Need: On-orbit performance is driven by mechanical stability (both thermal and dynamic). As future systems become larger, compliance cannot be fully tested; performance verification will rely on results from a combination of sub-scale tests and high-fidelity models. It is necessary to generate and validate as-built models of representative prototype components to predict on-orbit performance for transmitted wavefront, point spread function (PSF), pointing stability, jitter, thermal stability, vibro-acoustic, and launch loads.

Milestone #3A: Add capabilities to integrated design and modeling tools to predict the mechanical and thermal behavior of the candidate mirrors; use these tools to generate Pre-Phase-A point designs and predict on-orbit optical performance.

Milestone #3B: Validate by test integrated design and modeling tools to predict the mechanical and thermal behavior of the candidate mirrors.

Accomplishment: During FY 2016/17, AMTD Phase 2 accomplished Milestone 3A by continuing to develop its integrated design and modeling tools and by using these tools to conduct point design trade studies for the HabEx primary mirror, including on-orbit performance predictions. Additionally during FY2016/17, AMTD-2 accomplished half of major Milestone 3B by using its integrated modeling tools to predict the mechanical and thermal behavior of the Schott 1.2-m ELZM mirror assembly then validating those predictions by test. AMTD-2 expects to complete major Milestone B in 2017 by validating by test its performance predictions for the 1.5-m ULE® mirror assembly.

Mechanical and thermal models were made of the Schott 1.2-m ELZM mirror. The mechanical model predicted a gravity sag deformation of 125 nm rms and a first free-free resonant bending mode of 207 Hz. The thermal model predicted a 21 nm rms total surface figure error (SFE) consisting of contributions from its athermal mount, through thickness thermal gradient and bulk CTE homogeneity. The largest contributor of this error is from an assumed CTE homogeneity of 10 ppb (based on Schott catalog data). (Figure 5)

Figure 5: (left) Schott 1.2-m ELZM assembly. (center) Predicted first mode frequency. (right) Predicted thermal surface figure error.

66

The models were validated by test in the MSFC XRCF. First mode frequency was measured via tap test on foam blocks to be 196 Hz (5% agreement). Gravity sag was measured by rotation test to be 142 nm rms; with a 31 nm rms difference between predicted and measured. (Figure 6) This difference could be caused by a 2-mm error between the model and the ‘as-built’ mount pad locations.

Thermal model predictions are validated by measuring how the mirror shape deforms from 294K to 250K. For input into the thermal model, the mirror assembly is fully instrumented with thermal sensors to measure its bulk temperature and thermal gradients. Using the Schott catalog specification of 10 ppb CTE homogeneity the predicted thermal SFE is 21 nm rms. But, the measured SFE is 9.4 nm rms. After consulting with Schott, we were informed that the 1.2-m ELZM mirror has a CTE homogeneity of 5 ppb. With this new CTE specification, the predicted SFE is 9.55 nm rms. (Figure 7)

Figure 7: (left) Instrumented Thermal Test. (center) Measured Thermal SFE. (right) Predicted Thermal SFE with corrected CTE Homogeneity.

Milestone Schedule and Future Plans

AMTD has quantifiable milestones for each technology. Figure 8 shows the Phase 2 schedule. The primary tasks for FY 2017 are to predict the mechanical and thermal performance of the 1.5-m ULE® mirror and validate those predictions by test at the MSFC XRCF.

Figure 8. AMTD Phase 2 schedule.

67

AMTDPublications 1. Stahl, H. Philip, “Advanced mirror technology development (AMTD) project: year five status,” Proc.

SPIE 10401, 2017. 2. Brooks, Thomas, Tony Hull, Ron Eng and H. Philip Stahl, “Modeling the extremely lightweight

Zerodur mirror (ELZM) thermal soak test”, Proc. SPIE 10374, 2017. 3. Stahl, H. Philip, “Advanced mirror technology development (AMTD) project: overview and year four

accomplishments,” Proc. SPIE 9912, Advances in Optical and Mechanical Technology for Telescopes and Instrumentation II, (22 July 2016); doi:10.1117/12.2234082.

4. Committee for a Decadal Survey of Astronomy and Astrophysics; National Research Council, New Worlds, New Horizons in Astronomy and Astrophysics, The National Academies Press, Washington, D.C., 2010.

5. NASA Space Technology Roadmaps and Priorities: Restoring NASA’s Technological Edge and Paving the Way for a New Era in Space, NRC Report, 2012.

6. Kouveliotou, Centrella, Peterson, et al, Enduring Quests, Daring Visions: NASA Astrophysics in the Next Three Decades, 2014, arXiv:1401.3741, 2014.

7. Hertz, Paul, “Planning for the 2020 Decadal Survey: An Astrophysics Division White Paper”, January 4, 2015, available at science.nasa.gov/astrophysics/documents/.

8. NASA Town Hall, AAS Winter Meeting, Kissimmee, FL, 2016. 9. Dalcanton, Seager, et al, From Cosmic Birth to Living Earths: The Future of UVOIR Space

Astronomy, Association of Universities for Research in Astronomy, 2015, www.hdstvision.org/report/.

10. Lake, Peterson, and Levine, Rationale for defining Structural Requirements for Large Space Telescopes, AIAA Journal of Spacecraft and Rockets, Vol. 39, No. 5, 2002.

11. Stuart B. Shaklan, Luis Marchen, John Krist and Mayer Rud, “Stability error budget for an aggressive coronagraph on a 3.8m telescope”, SPIE Proceedings 8151, 2011.

Foradditionalinformation,contactH.PhilipStahl:h.philip.stahl@nasa.gov

68

PredictiveThermalControl(PTC)TechnologytoEnableThermallyStableTelescopesPreparedby:H.PhilipStahl(NASA/MSFC)

SummaryThePredictiveThermalControl(PTC)technologydevelopmentprojectisamultiyeareffortinitiatedinFiscalYear(FY)2017,tomaturetheTechnologyReadinessLevel(TRL)ofcriticaltechnologiesrequiredtoenableultra‐thermallystabletelescopesforexoplanetscience.

During2017/2018,PTChassuccessfullyprogresseditsthreedefinedobjectives:

1. Validatethermalopticalperformancemodels.2. Derivethermal‐systemstabilityspecifications.3. DemonstratePredictiveThermalControl.

byaccomplishingoradvancingthreeofitsfivequantifiablemilestones:

Milestone#1 (Complete): Created a high‐fidelity STOP model of the1.5‐m ULE® AdvancedMirrorTechnologyDevelopment(AMTD‐2)mirror,including3Dcoefficientofthermalexpansion(CTE) distribution and reflective coating, that predicts its optical performance response tosteady‐state and dynamic thermal gradients. Structural model was created using 3D X‐RayComputedTomography;

Milestone#2(Complete):Derivedspecifications for thermalcontrolsystemasa functionofwavefrontstabilityforaVectorVortexCoronagraph;

Milestone #3 (In Process): Designed and started fabricating a predictive Thermal ControlSystem for the AMTD‐2 1.5‐m ULE® mirror that senses temperature changes and activelycontrolsthemirror’sthermalenvironment;

Milestone#4 (InProcess): Performed preliminary Structural‐Thermal‐Optical‐Performance(STOP) model validation tests of the 1.5‐m ULE® mirror in a relevant thermal vacuumenvironmentattheMSFCX‐rayandCryogenicFacility(XRCF)testfacility;and

o Milestone#5:Usevalidatedmodeltoperformtradestudiestodeterminehowthermo‐opticalperformancecanbeoptimizedasafunctionofmirrordesign,materialselection,mass,etc.

The foundationofMSFC’s technologydevelopmentsuccess isourapproachofusingscience‐drivensystemsengineeringand teamingwith thebest scientists, engineers, and technologists in industry,academia, and government. PTC is a joint effort between NASA MSFC and Harris Corporation.Additionally,MSFCcontinuesmentoringthenextgenerationofscientistsandengineersasinterns,co‐ops, and volunteers. In this cycle, MSFC hosted three undergraduate student interns: JonathanMcCready,NorthCarolinaStateUniversity(Summer2018);HaoTang,UniversityofMichigan(Spring2018);andJonathanGaskins,UniversityofNorthCarolinaCharlotte(Fall2017,Summer2018).

PTCresultswerepublished inSPIEproceedingsandpresentedatMirrorTechnologyDays in theGovernmentWorkshop[1‐4].

69

Background“Arewealoneintheuniverse?”isprobablythemostcompellingsciencequestionofourgeneration.

Perthe2010NewWorlds,NewHorizonsDecadalReport[5]:“Oneofthe fastestgrowingandmostexcitingfieldsinastrophysicsisthestudyofplanetsbeyondoursolarsystem.Theultimategoalistoimagerockyplanetsthat lieinthehabitablezoneofnearbystars.”TheSurveyrecommendedasitshighestprioritymedium‐scale activity suchas a “NewWorldsTechnologyDevelopment (NWTD)Program”to“laythetechnicalandscientificfoundationsforafuturespaceimagingandspectroscopymission.” The National Research Council (NRC) report, “NASA Space Technology Roadmaps andPriorities”[6],statesthatthesecond‐highesttechnicalchallengeforNASAregardingexpandingourunderstanding of Earth and the universe in which we live is to “develop a new generation ofastronomicaltelescopesthatenablediscoveryofhabitableplanets,facilitateadvancesinsolarphysics,andenablethestudyoffaintstructuresaroundbrightobjectsbydevelopinghigh‐contrastimagingandspectroscopictechnologiestoprovideunprecedentedsensitivity,fieldofview,andspectroscopyoffaintobjects.”NASA’s“EnduringQuests,DaringVisions”[7]calledforaLargeUltraviolet/Optical/Infrared(LUVOIR)Surveyormissionto“enableultra‐high‐contrastspectroscopicstudiestodirectlymeasureoxygen,watervapor,andothermolecules intheatmospheresofexoEarths,”and“decodethegalaxyassemblyhistoriesthroughdetailedarcheologyoftheirpresentstructure.”Asaresult,NASAisstudyingin detail the LUVOIR and HabEx concepts for the 2020 Decadal Survey [8, 9]. Additionally, theAssociationofUniversitiesforResearchinAstronomy(AURA)report“FromCosmicBirthtoLivingEarths” [10]details thepotential revolutionaryscience thatcouldbeaccomplished from“directlyfindinghabitableplanetsshowingsignsoflife.”

Directly imaging and characterizing habitable planets requires a large‐aperture telescope withextremewavefrontstability.Foraninternalcoronagraph,thisrequirescorrectingwavefronterrors(WFEs) and keeping that correction stable to a few picometers rootmean square (rms) for thedurationofthescienceobservation.Thisplacesseverespecificationconstraintsontheperformanceof theobservatory, telescope, andprimarymirror.Per the2015CosmicOriginsProgramAnnualTechnology Report (PATR) [11], a “Thermally Stable Telescope” is critical, highly desirabletechnologyforastrategicmission.“Wavefrontstabilityisthemostimportanttechnicalcapabilitythatenables 10‐10 contrast exoplanet science with an internal coronagraph. State of art for internalcoronagraphyrequiresthatthetelescopemustprovideawavefrontthatisstableatlevelslessthan10pmfor10minutes(stabilityperiodrangesfromafewminutesto10sofminutesdependingonthebrightnessofthestarbeingobservedandthewavefront‐sensingtechnologybeingused).”

Thermal wavefront error occurs because of coefficient of thermal expansion (CTE); slewing thetelescope relative to the sun causes its structure ormirrors to change temperature. Changes inthermal heat load cause the structure holding the mirrors to expand/contract and the mirrorsthemselvestochangeshape.Fortunately,thermaldrifttendstobeslow,i.e.,manyminutestohours.Nopreviousmissionhaseverrequiredatelescopewithawavefrontthatisstableatlevelsoflessthan10pmper10minutes.State‐of‐the‐art(SOTA)forambienttemperaturespacetelescopesare‘cold‐biased’withheaters.Thetelescopeisinsulatedfromsolarloadsuchthat,forallorientationsrelativetothesun,itisalwaysata ‘cold’temperature(forexample,250K).Thetelescopeisthenwarmedtoanambienttemperatureviaheaterpanelsontheforwardstray‐lightbaffletubeaswellasbehindandbesidethemirror.SOTAthermalcontrolsystemshavedemonstrated~1‐Kstabilityfora1‐mclasstelescopeusinga‘bang/bang’controller.Analysisindicatesthat~100mKshouldbepossiblewithaproportionalintegraldifferential(PID)controller.

PTC plans to advance the SOTA in thermal control by demonstrating a control logic calledModelPredictiveControl(MPC)[12,13].MPCplacesaphysics‐basedmodelintothecontrollooptodeterminecontrolvariables(heaterpowerlevels)basedonstatevariables(temperaturemeasurements).MPC

70

determines heater power levels using a completely different logic than proportional control.Proportional control adjusts heater power in proportion to the difference betweenmeasured anddesired temperatures at one location.MPCusesmultiple control zones and takes into account theinterdependency between all control zones’ temperatures and heater power. Preliminary analysisindicatesthat(assumingthatthermalperformanceis linear) it ispossibletoachievepmwavefrontstabilitybyeithercontrollingtheshroudtoasmalltemperature(10mK)orbyrapidlycorrectingthetemperature.Giventhatmirrorsandtelescopehaveathermalresponsetime,thebestwaytoachievepm‐level stability is to sense and correct for changes in the thermal environment faster than thetelescopecanrespond.Additionalstabilitycanbeachievedbyincreasingthesystem’sthermalmass.Basedonthisanalysis,wecurrentlyassesssuchasystem’sTRLat3.PTCwilladvancethisTRLbytestusingthe1.5‐mUltraLowExpansion(ULE®)mirrorfabricatedbyAMTD‐2.

ObjectivesandMilestonePTChasdefinedthreeobjectivestomaturethetechnologyneededforanexoplanetsciencethermallystabletelescopebydeveloping“thermal‐designtechniquesvalidatedbytraceablecharacterizationtestingofcomponents.”

1. Validatingmodelsthatpredictthermal‐opticalperformanceofrealmirrorsandstructurebasedon their structuraldesignsandconstituentmaterialproperties, i.e.,CTEdistribution, thermalconductivity,thermalmass,etc.

2. Derivingthermalsystemstabilityspecificationsfromwavefront‐stabilityrequirement.3. DemonstratingutilityofaPredictiveControlthermalsystemforachievingthermalstability.

Toachieveourobjectives,wedefinedadetailedtechnicalplanwithfivequantifiablemilestones:

Milestone#1:Developahigh‐fidelitytraceablemodelofthe1.5‐mULE®AMTD‐2mirror,including3DCTEdistributionandreflectivecoating,thatpredictsitsopticalperformanceresponsetosteady‐stateanddynamicthermalgradients.

Milestone#2:Derivespecificationsforthermalcontrolsystemasafunctionofwavefrontstability.

Milestone#3:DesignandbuildapredictiveThermalControlSystemfora1.5‐mULE®mirrorthatsensestemperaturechangesandactivelycontrolsthemirror’sthermalenvironment.

Milestone #4: Validate model by testing the 1.5‐m ULE® mirror in a relevant thermal‐vacuumenvironmentattheXRCFtestfacility.

Milestone #5: Use validated model to perform trade studies to determine how thermo‐opticalperformancecanbeoptimizedasafunctionofmirrordesign,materialselection,mass,etc.

ProgressandAccomplishmentsObjective#1:ValidatedHigh‐FidelitySTOPModel

Need:Designinga telescopetohave10‐pm‐per‐10‐minutesWFEstabilityusingmodelpredictivecontrolrequiresavalidatedhigh‐fidelitySTOPmodel.

Milestone#1:Developahigh‐fidelitySTOPmodelofthe1.5‐mULE®AMTD‐2mirror,including3DCTEdistributionandreflectivecoating,thatpredictsitsopticalperformanceresponsetosteady‐stateanddynamicthermalgradients.

Milestone#4: Validate high‐fidelity STOPmodel by testing the 1.5‐mULE®mirror in a relevantthermal‐vacuumenvironmentattheXRCFtestfacility.

Accomplishment:DuringFY2018,PTCcompletedMilestone#1andadvancedMilestone#4.Ahigh‐fidelitymodel of 1.5‐mULE® AMTD‐2mirrorwas developed and used to correlate a 231‐K static

71

thermalsoakdeformationofapproximately20nmrmstoanuncertaintyof13.4nmrms.Additionally,itwasusedtocorrelateadeformationofapproximately80‐nmrmsfora115‐Klateralthermalgradient.

ToaccomplishMilestone#1,thePTCTeamcreatedahigh‐fidelitymodeloftheAMTD‐21.5‐mULE®mirror in NASTRAN that accurately models its ‘as‐built’ mechanical dimensions and 3D CTEdistribution. The ‘as‐built’ mechanical dimensions were accurately modeled using 3D X‐raycomputedtomographytomeasuretheinternalstructureofthemirror(Fig.1).Thesemeasurementswerethenportedintoamechanicalmodel(Fig.2).Toadda3DmappingofCTEdistribution,HarrisCorporationprovidedMSFCwithCorningCTEdatamapsforeachofthe18coreelementsandthelocationofeachelementinthecore(Fig.3).

Fig.1.1.5‐mAMTD‐2mirrorbeingX‐RayCTscanned.

Fig.2.X‐RayCTdatawasconvertedintoMechanicalmodel.

Fig.3.HarrisCorpprovidedCorningCTEdataforfromwhereeachcoreelementwascutfromitsboule(left)andthelocationofthatcoreelementintheAMTD‐2mirror(right).

TheprocessbywhichtheAMTD‐2mirrorwasfabricatedresultedinbowingofthecorewalls.TheeffectwasdiscoveredbecausetheATMD‐2mirrorwasdeliberatelyfabricatedtoaveryfastF/#whichplacedthemirrorinanunusuallyhighstressstateduringfabrication.Suchastressstatewouldnotbeimposedonanactualflightmirror.The3DX‐rayCTscanproduceda3Dmappingofthecorewallbowing.AcustomalgorithmwaswrittentoconverttheX‐rayCT3Dmappingintoafiniteelementmodel.

ToadvanceMilestone#4,the1.5‐mULE®AMTD‐2mirror’sresponsetostaticthermalloadsandlateralthermal gradientswas tested in theXRCF. This testwas conducted jointlywith theAMTD‐2 staticthermalsoaktest.Thistestwasabare‐mirror‐onlytest,i.e.mirroronlywithnothermalcontrolsystem.

First, the high‐fidelitymodelwas correlated to cryo‐deformation of the AMTD‐2mirrormeasuredduringthestaticthermalloadtest.Thisdeformationconsistsoftwocomponents:theopto‐mechanical‐thermaldeformationofthemirrormountsystemandthemirrorsubstrate’sCTEdistribution.Critical

72

to model correlation is understanding the temperature distribution on the mirror and its mountsystem.Toachievethisdistribution,themirrorisfullyinstrumentedinthetestsetup(Fig.4).

Fig.4.PTCTestSetupwithThermalSensors

Asthetemperatureofthemirrorandmountchangesfrom293to231K,thealuminumbackplaneexpands, and the mount struts apply a prying force to the mirror. The prying signature is notsymmetriceventhoughthedesignissymmetric,whichmeansthattheas‐builtmounthasunintendedasymmetries.Themodelappliespryingforcesdirectlytothebondpadandthecombinationofforcesthatmostcloselymatchesthetestdatawasfound.Figure5showsan18.9‐nmrmssurfaceshapethatbestfitsthetestdataproducedentirelybyapplyingforcesontothebondpads.CTEinhomogeneityalsoproducecryo‐deformation.Figure6showsa16.6‐nmrmssurfaceshapethatbestfitsthetestdataproducedentirelybythemirror’s3DCTEdistribution.

Fig.5.PredictedAMTD‐218.9‐nmrmsMountCryo‐Deformation.

Fig.6.PredictedAMTD‐216.6‐nmrmsCTECryo‐

Deformation.

CombiningmountandCTEeffects,thePTChigh‐fidelitymodeloftheAMTD‐2mirrorpredicts24.7‐nmrmsofthemeasured28.8‐nmrmswitha13.4‐nmrmsuncertainty(Fig.7).

73

Fig.7. Left:Thermal soak of themirror andmount going from 293 to 231K resulted in ameasured cryo‐deformation of 28.8‐nm rms.Center:High‐fidelityAMTD‐2model predicts 24.7‐nm rms of deformation frommount effects and CTE inhomogeneity. Right: Difference between measured deformation and predicteddeformationis13.4‐nmrms.

To furtheradvanceMilestone#4, the1.5‐mULE®AMTD‐2mirror’s response toa lateral thermalgradient was tested in the XRCF. Again, this test was conducted jointly with the AMTD‐2 staticthermalsoaktestandwasabare‐mirror‐onlytest.

PTChasmodifiedMSFC’sXRCFfacilitytointroducethermalgradientsintomirrorsystemsusingsolarlamps(Fig.8).Thiscapabilityhasbeentestedonthe1.5‐mAMTDULE®mirrorusingasinglelamparray.Athermalgradientof115Kwasintroducedintothemirrorcausinga78.7‐nmrmssurfacedeformation(Fig.9).Ournextstepistoaddanactivethermalcontrolsystemtothismirror.

Fig.8.SolarLampArrayaddedtoMSFCXRCF.

Fig.9.AMTD‐2mirrortestedwithoneLampArrayElement;ThermalGradientproduced;andmeasuredSurfaceFigureError.

Objective#2:TraceableSpecificationsforanActiveThermalControlSystem

Need:Designingatelescopetohave10‐pm‐per‐10‐minutesWFEstabilityviaactivethermalcontrolrequiresavalidatedSTOPmodel.Suchamodelisrequiredtoderivetheactivethermalcontrolsystem’s

74

performancespecifications,suchas:sensingresolution(1mKor10mK),controlaccuracy(10mKor50mK),controlperiod(1minor5min),numberanddistributionofsenseandcontrolzones,etc.

Milestone#2:Derivespecificationsforthermalcontrolsystemasafunctionofwavefrontstability.

Milestone #5: Use validated model to perform trade studies to determine how thermo‐opticalperformancecanbeoptimizedasafunctionofmirrordesign,materialselection,materialproperties(i.e.,CTE),mass,etc.

Accomplishment: During FY 2018, in collaboration with the HabEx engineering team, PTCcompletedMilestone#2.TheHabExengineeringteamhasdefinedandprovidedtoPTCasensitivityanalysis of Zernike polynomial terms for the Vector Vortex 6 coronagraph and flowed thesesensitivitiesinanerrorbudgetdowntothermalstabilityfortheprimarymirror.PTCmeasuredtheSchottZerodurmirror,showing11.3‐nmrmsofsurfacedeformationper62Kofbulktemperaturechange. Decomposing this error into Zernike polynomials indicates that if this mirror’s bulktemperaturehasathermalstabilityofapproximately1mK,itwillmeettheHabExVVC‐6wavefrontstabilityrequirement(Fig.10).Milestone#5isnotscheduledforcompletionuntiltheendofPTC.

Fig.10.1.2‐mSchottZerodurMirrorThermalSoakDeformationwillmeettheHabExVVC‐6ThermalStabilityRequirementifitstemperatureisstableto~1mK.

Objective#3:Demonstrateutilityofapredictive‐controlthermalsystemforachievingthermalstability.

Need:A telescopewith 10‐pm‐per‐10‐minutesWFE stability requires an active thermal controlsystemthatisbeyondthecurrentstateofart(i.e.,bang‐bangorproportionalcontrol).

Milestone#3:Design,build,andtestapredictivethermalcontrolsystemforthe1.5‐mULE®AMTD‐2mirrorthatdemonstratestheutilityofaphysics‐basedmodelinthecontrollooptodeterminecontrolvariables(heaterpowerlevels)basedonstatevariables(temperaturemeasurements).

Milestone#3hastwogoals.First,providedataneededbyMilestone#4tovalidateMilestone#1’sthermal model. Second, demonstrate Predictive Thermal Control. PTC will be considereddemonstratedifitcancorrectforexternallyimposedthermalgradients(i.e.,radial,lateral,andaxialgradients).OthergoalsofMilestone#3include:self‐tuningoflesswell‐knownthermalparametersinthethermalmodeltoimprovethePTC’sveracity,informingthedesignofhardwarelikeheatedbathtubsandthermalshroudstoenablecontrollability,anddirectlyimposingmeasurablethermallyinducedWFEintothemirror.

Accomplishment:DuringFY2018,PTCprogressedtechnologyforObjective#3bymodifyingtheXRCFtobeable to impose intomirrorsystemsaxialand lateral thermalgradients (Fig.8).Additionally,MSFCandHarrisCorpdesignedandstartedfabricationofathermalcontrolsystemfortheHarris1.5‐mULE®mirror(Fig.11).Thethermal‐controlhardwarehas26controlzonesthatwillbeusedtoactively

75

correctsurfacefigureerrorormitigatetheeffectofanexternalthermaldisturbance(Fig.12).Thermal,stress,andopticalanalysisofthisdesignhasbeendonetoevaluatethedesign’sutility.

Fig.11.ThermalControlSystemwitha26‐zonecontrolforAMTD‐2’s1.5‐mULE®mirror

Fig.12.ThermalControlSystemcanintroduceRadial,Lateral,andAxialthermalgradients.

STOPanalysiswasdonetoestimatehowmuchthesurfacemightchangeandhowwellthethermalcontrolsystemcancorrectthesurfacewhenheatlampsareturnedon.Withoutthermalcontrol,themirror’ssurfacefigureerrorisestimatedtobe7.5‐nmrms(Fig.13).Withthermalcontrol,thiserrorisexpectedtobereducedto1.5‐nmrms(Fig.14).Becausetheestimatederrorissosmall,PTCisinvestigatingfabricatinga1‐maluminumtestmirrorinFY2019.

Fig. 13. Estimated surface figure errorwithoutactivethermalcontrol.

Fig.14.Estimatedsurfacefigureerrorwithactivethermalcontrolcorrection.

76

PathForwardPTC has quantifiable milestones for each technology. Figure 15 shows PTC’s five quantifiablemilestones foreach technologyon the ‘new’baseline schedule.PTCwasoriginallyproposedas athree‐yearproject,butduetolackoffundingwasstretchedtofouryears.

Fig.15.PTCPhase‐1schedule.

77

References[1] T.E.Brooks,R.Eng,T.Hull,andH.P.Stahl,“ModelingtheExtremelyLightweightZerodurMirror

(ELZM)thermalsoaktest,”Proc.SPIE10374,OpticalModelingandPerformancePredictionsIX,103740E(6September2017),doi:10.1117/12.2274084

[2] T.E. Brooks, “Predictive thermal control applied toHabEx,” Proc. SPIE 10398, UV/Optical/IRSpaceTelescopesandInstruments:InnovativeTechnologiesandConceptsVIII,1039814(5Sept2017),doi:10.1117/12.2274338

[3] T.E.BrooksandR.Eng,“1.5‐mULE®Cryo&MechanicalTestResults,”MirrorTechnologyDaysintheGovernment(2017)

[4] T.E.Brooks,“ComputedTomographyandStictioninaLowTemperatureSlumpedMirror,”MirrorTechnologyDaysintheGovernment(2017)

[5] Committee for aDecadal SurveyofAstronomyandAstrophysics;NationalResearchCouncil,“NewWorlds,NewHorizons inAstronomy andAstrophysics,” The National Academies Press,Washington,D.C.(2010)

[6] NRC,“NASASpaceTechnologyRoadmapsandPriorities:RestoringNASA’sTechnologicalEdgeandPavingtheWayforaNewErainSpace,”(2012)

[7] C. Kouveliotou et al., “EnduringQuests,Daring Visions:NASA Astrophysics in theNext ThreeDecades,”(2014),arXiv:1401.3741

[8] P.Hertz,“Planningforthe2020DecadalSurvey:AnAstrophysicsDivisionWhitePaper,”(January4,2015),availableatscience.nasa.gov/astrophysics/documents/

[9] NASATownHall,AASWinterMeeting,Kissimmee,FL(2016)[10] J.Dalcantonetal.,“FromCosmicBirthtoLivingEarths:TheFutureofUVOIRSpaceAstronomy,”

AssociationofUniversitiesforResearchinAstronomy,(2015),www.hdstvision.org/report/.[11] “CosmicOriginsProgramAnnualTechnologyReport”(October2015)[12] T.E. Brooks, H.P. Stahl and W.R. Arnold, "Advanced Mirror Technology Development (AMTD)

thermal trade studies," Proc. SPIE 9577, Optical Modeling and Performance Predictions VII,957703(23September2015);doi:10.1117/12.2188371

[13] T.E. Brooks, "Predictive thermal controlapplied toHabEx," Proc. SPIE 10398, UV/Optical/IRSpaceTelescopesandInstruments:InnovativeTechnologiesandConceptsVIII,1039814(5Sept2017),doi:10.1117/12.2274338

Foradditionalinformation,contactH.PhilipStahl:h.philip.stahl@nasa.gov

78

DevelopmentofLarge‐Area(100cm2)Photon‐CountingUVDetectorsPreparedby:JohnVallerga(SpaceSciencesLaboratory,UCBerkeley)

SummaryMicro‐Channel‐Plate (MCP) detectors have been an essential imaging technology in space‐basedNASA ultraviolet (UV) missions for decades, and have been used in numerous orbital andinterplanetary instruments. The Experimental Astrophysics Group (EAG) at the University ofCalifornia(UC)Berkeley’sSpaceSciencesLaboratory(SSL)wasawardedanAstrophysicsResearchandAnalysis(APRA)grantin2008todevelopmassivelyparallelcross‐strip(XS)readoutelectronics.TheselaboratoryXSelectronicshavedemonstratedspatialresolutionsof15-μmfull‐widthathalf‐maximum(FWHM),globaloutputcountratesof2 MHz,andlocalcountratesof100kHz;allatgainsa factor of ~20 lower than existing delay‐line readouts [1]. They have even been deployed inbiomedical and neutron‐imaging labs but are presently too bulky and high‐powered for spaceapplications,thoughacurrentversionwassuccessfullyflownonfourrocketflightsince2014[2].

EAGhasbeenawardedtwoStrategicAstrophysicsTechnology(SAT)grantstodevelopXStechnologyforlarge,photon‐countingUVdetectors:thefirst in2012fora50‐mmsquaredetector,andmorerecently(2016)scalingthistechnologytoa100‐mmsquaredesign.ThegoaloftheseSATprogramsistoraisetheTechnologyReadinessLevel(TRL)ofthisXStechnologyby:

1. DevelopingnewApplicationSpecificIntegratedCircuits(ASICs)thatcombineoptimizedfasteramplifiersandassociatedAnalog‐to‐DigitalConverters(ADCs)inthesamechip(s).

2. Developing a Field‐Programmable‐Gate‐Array (FPGA) circuit that will control and read outgroupsoftheseASICssothatXSanodesofmanydifferentformatscanbesupported.

3. Developing a spaceflight‐compatible 100‐mm XS detector that can be integrated with theseelectronicsandtestedasasysteminflight‐likeenvironments.Thisdetectordesigncanbeuseddirectlyinmanyrocket,satellite,andinterplanetaryUVinstruments,andcouldbeeasilyadaptedto different sizes and shapes to match various mission requirements. New technologicaldevelopmentsinphotocathodes(e.g.,GaN)orMCPs(e.g.,low‐background,surface‐engineered,borosilicate‐glassMCPs)couldbeaccommodatedintothisdesignastheirTRLsadvance.

SincethestartofourprojectinApril2012,wehavedesignedandconstructeda50‐mmXSdetectorwith a new low‐noise anode and have demonstrated its excellent performance using our bestlaboratoryelectronics.Wehavescaledupandfabricatedthe50‐mmdetectordesignto100 mm × 100 mm(320channels),andconfirmeditsperformancebeforeandafterafull3‐axisvibrationtest.The new version of our ASIC design, fabricated in a 130‐nm CMOS process that combines thefunctions of the previous two versions to reduce board complexity, will soon be submitted to afoundry. Upon receipt of the mounted dies in ~4 months, we plan to demonstrate its dynamicperformanceandincorporateitintoafull100‐mmXSreadoutsystem.

BackgroundXSreadoutscollect thechargeexiting fromastackofMCPswith twosetsofcoarselyspacedandelectricallyisolatedorthogonalconductingstrips(Fig.1).Whenthechargecollectedoneachstripismeasured,acentroidcalculationdeterminestheincidentlocationoftheincomingevent(photonorparticle).Thisrequiresmany(e.g.,160,320)identicalamplifierswhoseindividualoutputsmustallbedigitizedandanalyzed.TheadvantageofthistechniqueoverexistingandpreviousMCPreadouttechniques(wedgeandstrip,delay‐line,intensifiers)isthattheanodecapacitanceperamplifieris

79

lower, resulting in lowernoise. This allows (factor of~20) lowerMCP gain operationwhile stillachieving better spatial resolution compared to the delay‐line MCP readouts of current spacemissions [3], thereby increasing the dynamic range of MCP detectors by up to two orders ofmagnitude.Thiscanalsobescaledreadilytolarge(> 100 mm × 100 mm)oruniqueformats(e.g.,circularforopticaltubes,rectangularforspectrographs,andevencurvedanodestomatchcurvedMCPfocalplanes).TheXSreadouttechnologyismatureenoughtobepresentlyusedinthefieldinmanylaboratoryenvironmentsproducingqualityscientificresults[4,5],andisreadyforthenextstepofdevelopment–preparingforanorbitalordeep‐space‐missionimplementation.

Fig.1.SchematicofaXSreadoutbehindanMCPstack.

Our laboratory XS readout electronics, called the Parallel XS (PXS) electronics, consist of apreamplifierboardplacedneartheMCPanodeandaboxedsetofelectronicscontainingADCsandFPGAs.ThePXSelectronicsperformancepresentlymeetsorexceedsallofthespecificationsofthepreviousflightsystemsmentionedabove.However,thePXSlaboratoryelectronicsaretoobulkyandmassive,anduserelativelyhighpowerandthereforearenotcurrentlysuitableforalong‐termspacemission.OneimportantgoalofthepresenteffortistoreplacethePXSelectronicswithASICsthatcombinethefunctionalityofthepreampboardandthedownstreamADCsintooneortwolow‐power,low‐masschip(s).WhenasetofthesechipsiscombinedwithanFPGAandXSanode,weexpecttheperformancetoexceedthehigher‐powerPXSelectronicsduetothenoiseimprovementexpectedfromthesmaller‐scalecomponents.

In addition to space‐flight‐appropriate ASIC development, our first SAT program planned toconstructaflightprototype50‐mmXSMCPdetectorwithaXSreadoutusingournewASICs.ThenewASICsandFPGAcontrolelectronicsarebeingintegratedintoacompactpackagesotheperformanceofthewholedetectorsystemcanbequalifiedinspace‐likeenvironments(e.g.,thermalandvibrationtests).Infact,weexpectthisdetectortobethebaselineofmanyExplorer‐classmissionproposalsinthefuture.ThisXSdesigncanalsobescaledeasilytootherusefulformatsrequiredbyspecializedinstruments.Forexample,doublingthelengthofonedetectordimensionentailsaddingmorestripstotheanodeandmoreASICstoreadthemout,butnotaredesignoftheASIC.

80

ObjectivesandMilestones1.DesignandfabricateASICstoamplifyanddigitizecross‐stripsignalcharges

Amajorthrustofour50‐mmprogramwasdevelopingnewASICsthatcanovercomethelimitationsonthefrontendofourlaboratoryelectronics.WedesignedandfabricatedinputASICsthathavethefollowingfeatures:

a. An optimized front‐end charge‐sensitive amplifier (CSA) matched to the anode‐strip loadcapacitancewithfastsignalriseandfalltimestominimizeevent“collision.”

b. Fast (~GHz) analog sampling to fully characterize both amplitude and arrival time of theintrinsicallyfastinputchargepulse.

c. DigitalconversionoftheanalogsamplesintheASIC,avoidingcomplex,bulky,andhigh‐powerdiscreteADCsdownstream.

d. ASICself‐triggeringcapabilitiestoselectandtransferonlyeventdataacrosslongcablestotheFPGA,wherethecentroidingandtimingcalculationstakeplace.

AfterdemonstratingtwoworkingASICs(CSAv3andHalfGRAPH2describedbelow),integratedintothe50‐mmreadoutelectronics,wereturnedtotheoriginalconceptofcombiningtheCSAwiththefastsamplingADC.Asthenewdigitizer/samplerwillnowbeinthesameCMOSprocessastheCSAv3,wecanrealizetheoriginal“GRAPH”ASICconcept,combiningbothonasingledie.Therearemanyreasonsfordoing this,but the twomost important are a fasterdigital readout in the130‐nmprocess, andreducedlayoutspaceandpoweraffordedbycombiningtheASICdesignsintoasingledie.Wealsohavemethodstoprotectthelow‐noiseanalogamplifierfrominducednoisefromthedigitalcircuit.

We have also characterized the first two separate ASICs in realistic radiation environments (totalionizingdose,TID)consistentwithMILSTD833TM1019forlowEarthorbit(~10krad),interplanetarymissions (~30 krad), and the higher doses associated with Jupiter missions (100 krad behindshielding),havingplacedMCPdetectorsinallsuchenvironments.ThesesuccessfulradiationtestsoftheCSAv3andHalfGRAPH2informedthedesignoftheGRAPH,andwehaveanopportunitytoimproveitsrobustnesstoradiationdosebyusingstandardindustrydesigntechniquesforradiationhardness.

2.FPGAsystemtoreadoutHalfGRAPHASICs

Ourproposedparallel‐XSreadoutsystemwasnotsimplycomprisedofthenewASICs.Newboardassemblieshadtobedesigned, laidout,andconstructedtocouple theseASICs toourexistingXSanodes,minimizingstray‐loadcapacitances.TheseboardsalsoincludedcontrolFPGAsthatnotonlyhaveanewinputinterface,butalsoanewoutputinterfacetocoupletothehigh‐bandwidthcomputerinterfacerequiredforourultimateeventrates.

3.Designof50‐mmand100‐mmXSMCPdetectorsincorporatingnewelectronics

Keyissuesforlarge‐areaXSMCPdetectorimplementationsincludelow‐massandrobustconstructionschemesthatcanaccommodatethecapabilityforahigh‐vacuumsealed‐tubeconfiguration.Withoutincurringexcessivecosts,areasonableformattoaccomplishthisfirstwasthe50‐mmdetector.Thisdetectorachievedspatialresolutionsof~20 μmFWHM,backgroundrates<0.1eventscm‐2sec‐1,lowfixed‐patternnoise,longlifetime,multi‐MHzratecapabilitywithlowdeadtime,anddetectormassofafewhundred grams. The design and construction of brazed‐body assemblies provides for the bestpackaginganddiversityofapplications,sothiswasoneofthecoretasks.

In the 100 mm × 100 mmeffort,we scaled up the brazed‐body assembly, detector backplate, XSanode,andnumberofchannelsbyafactorof2ineachdimension(320electronicchannelstotal),withthegoalofmaintainingtheperformanceparametersachievedwiththe50‐mmassemblyandtotestinappropriateenvironments,thermalandvibrational.

81

ProgressandAccomplishmentsWorkprogressesonallthreeobjectivesofthe100‐mmXSdetectorobjectives:ASICdesign,fabrication,andtestattheUniversityofHawaii;FPGAcontrolelectronicsatUCBerkeley;and50‐mmand100‐mmXSdetectordesignandtest,alsoatUCBerkeley.TwoversionseachoftheASICshavebeenfabricatedandtested,ashasthe100‐mmdetector(usingthePXSreadoutelectronics).The160‐channelFPGA‐controlled amplifier and digitizer boards were fabricated along with their electronics boxes, andphotoneventsarenowbeingreadout,albeitata loweventrate,butenoughtodemonstratehigh‐resolution imaging. Radiation testing of both the CSAv3 and theHalfGRAPH2 has been completedsuccessfullyup to460‐kRadTID level.The100‐mmbrazedbody,anode,andbackplatedesignwascompleted; and the detectorwas fabricated and tested in vacuumwithUV photons using the PXSelectronics. It has been successfully vibrated to the levels specified in the General EnvironmentalVerificationStandard(GEVS,GSFC‐STD‐7000A),14grms(rootmeansquare)inthreeaxes.

ASICDesign

Wehavesuccessfully fabricatedand tested twoworkingASICdesigns: the16‐channelCSAcalled“CSAv3”(Fig.2)andthe16‐channelanalogsamplinganddigitizingHalfGRAPH2.Thesectionsbelowdiscussthesedesignsandtheirtestingtodate.

Fig.2.CSAv3dielayoutshowing16channelsalongwithregionsofcontrolcircuitryincludingDigital‐to‐AnalogConverters(DACs)andregisters.

CSAv3

AtanMCPgainof106(160fC), thelargestsignalexpectedonasinglestripis~50fC.TheCSAv3amplifierwasdesignedtodrivetheinputoftheHalfGRAPH2,includingtheshorttracebetweenthem.Wedesignedtheamplifiertohavelinearoutputswingsof±600mV.

Thecomponentsofasingleamplifierstage(Fig.3)consistofacharge‐sensitivepreampfollowedbyapole‐zerocancellation(PZC)network,ashaper,anoptionalpolarityinverter,andabufferamptodrivethedownstreamHalfGRAPH2.Thepreampisaninvertingfolded‐cascodeintegrator,withthefeedback‐circuittimeconstantcontrolledbyaPMOS(P‐typemetal‐oxide‐semiconductor)transistoroperatingasavoltage‐controlledresistor.ThePZCcircuitcancelsthepoleofthepreamp,removingthe long baseline recovery. Additional programmable shaping and buffering circuits allow us tooptimizethegainandnoiseoftheamplifier.

82

Fig.3.AgeneralizedschematicofthenewCSAv3.

Initialperformancetestswiththeamplifiershowthatitmeetsthespecificationsforgain,noise,rise‐time,andfall‐time[6].Insummary,theoverallgainis~9.5mV/fCoverarangeof0to60fCinputcharge.Within thisrange, thegain linearity iswithin7%,and theoutputpulsewidth isconfinedwithin100ns.Thermsnoisewasestimatedfromafitofnoisemeasurementsfordifferentinputloadcapacitances, and results in a noise equivalent of 586 e– + 96 e–/pF input load. An example of ameasuredwaveformforachargeinjectionof50fCisshowninFig.4.Detailsofthedesign,simulation,andfabricationofCSAv3canbefoundin[6].

Fig.4.Amplifierresponsetoa50‐fCinputpulsewitha5‐pFload.Notethe20‐nsrisetimeandareturntobaselinein75nswhileretaininganoisevalueof~1000e–rms.

HalfGRAPHASIC

HalfGRAPH2 [7] is a 16‐channel, 1 giga‐sample‐per‐secondwaveform‐digitizing chipwith 12‐bitresolution(Figs.5and6). Itwasdesigned in theTaiwanSemiconductorManufacturingCompany(TSMC) 0.25‐μm CMOS technology, using Tanner EDA design and simulation tools. The circuit’sheritageistheTeVArrayReadoutwithGSa/ssamplingandEventTrigger(TARGET)ASICusedforphotomultiplier‐waveformsamplingintheCherenkovTelescopeArray.Eachchannelofthisdigitizerchip has a two‐stage analog storage mechanism. In the first stage, a short sampling array issubdividedintotwosamplewindows,eachwith32switched‐capacitorstoragecells.Inthesecondstage,inaping‐pongfashion,asonesamplewindowisfilling,theotheristransferredintoalargerstoragearray.Thisstoragearrayhas8192cells,organizedastwobanksof64rowsof64sampleseach(alsocalledastoragewindow)foreverychannel.Thisresultsinacontinuoussampleof8.192‐μslengthbeforebeingoverwritteninacircularbuffer.

83

Fig.5.HalfGRAPH216‐channelASIClayout(left)anddiemountedincarrier(right).

Fig.6.FunctionalschematicoftheHalfGRAPH2ASICdesign(describedindetailinthetext).

TheinputanalogtriggercircuithasfourdigitaloutputlinesthatnotifyacontrolFPGAofanewevent.Atriggerpulseissetinplaceifthemeasuredsignalonachannelishigherthanapresetthreshold.Triggerlinesareorganizedtocoverchannels1‐4,5‐8,9‐12,and13‐16.Thisallowsfinerlocalizationofthestripwheretheeventoccurred.ThetriggerarrivalintotheFPGAmarkstheaddressinthestoragearrayofthesamplesforthatevent.

Inordertodigitizetheacquiredsignal(e.g.,Fig.4),eachchannelusestwobanksof12‐bitWilkinsonADCs.AnFPGAselectsthestoragewindowtobedigitized.Thirty‐twoanalogsamplesinparallelonall16channelsofthestoragewindowareconvertedconcurrentlyintodigitalvaluesusingcomparators,avoltageramp,andcounting(12‐bitcounter)witha500‐MHzdual‐phaseclockuntilthecomparatorfireswhentherampexceedstheanalogvalueonthecell.Ittakes4.1μstocompletethedigitization.

AstheFPGAhastheaddressoftheeventsinthestoragearray,itcanpreventtheoverwritingofthosecells.Therefore,thetheoreticalthroughputofthesystemislimitedbythe4.1‐µsconversiontimeoftheADCs.Themaximumthroughputofonechannelis240kHz,butthereisnodeadtimeatratesbelowthisfrequencyduetothemultiplebuffering.

FPGAControllerandIntegratedReadoutElectronicsfor50‐mmDetector

InparallelwiththeASICtestingdescribedabove,wedesignedtheelectronicslayoutforthefull160channelsofthe50 mm × 50 mmanodereadout(Fig.7).Thesensitivepreampsmustbemountedclose

84

totheanodetominimizecapacitanceloadingoftheinput,andtheHalfGRAPH2digitizingASICsmustbeclosetothepreampoutputstominimizetheiroutputload.Giventheseconstraints,aswellasthedesireto fit the electronics into a small enclosuremounted directly to the detector backplate (Fig. 8),wedecidedtoconstructtwoprintedcircuitboards:anamplifierboardwith10CSAv3chipsandadigitizerboardwith10HalfGRAPH2chipsandtwoFPGAs(ArtixAX200),oneforeachaxis,XandY.ThedigitizerboardwillinterfacewithadownstreamFPGAdevelopmentboard(e.g.,SP601)via11LVDSpairs.

Fig.7.High‐levelschematicof50‐mmdetectorelectronics, includinga160‐channelamplifierboard(left)and160‐channeldigitizerboardwithtwoFPGAstocontrolASICsandcalculateeventcentroids(right).

Fig.8.Amplifier(bottom)anddigitizer(top)boxesmountedto50‐mmdetectorbackplate.

The160totalanodechannels(80 × 80)areinputtoaCSAv3amplifierboardusingtwoSamtecTMS‐140‐01‐S‐D connectors, and the 160 amplified signals pass through an inter‐board connector(Samtec LSS‐150‐02‐F‐DV‐A) to the 10 HalfGRAPH2 chips on the digitizer board. Additionalconnectorsalsoexistforinter‐boardpower,control,andcommand/dataout.Themaindatainterface(using 11 LVDS lines running at 100 MHz double data rate, DDR, to the development board)communicateswithbothFPGAs,XandY,andiscontrolledbythedownstreamhost.Initiallyinrawoutput mode, we transmit all 160‐channel samples of 128‐ns duration just for triggered events,achieving8000events/susingbothXandYinterfaces.

85

The boards have been fabricated, installed in their boxes, and coupled to the 50‐mm detectorbackplate(Fig.8).Wecantestthemeitherwithtestpulsescapacitivelycoupledintotheirinputordirectlywithphotoneventsfromthe50‐mmdetector.LearningtocontroltheHalfGRAPH2chipwiththeFPGAhasbeenalongandlaboriousprocess,asthetriggertimingoftheanaloggiga‐samplesintheringbufferisverysophisticatedwhilemaintainingsynchronicitybetweentheXandYchannels.We are now able to read out in full raw‐datamode photon events into our computer at 360Hz(limitedbythe“ZED”boardtoEthernetinterface),whichallowsustocheckourevent‐centroidingalgorithmsandfinite‐impulse‐response(FIR)filteringtechniques.

WediscoveredfourissueswiththeHalfGRAPHchipwhichwewanttoimprove/fixintheGRAPHdesign.Thetwoleastsignificantbits(LSBs)oftheWilkinsonADCdonotseemtofunction,soourpulsesaresampledat10bitsinsteadof12.Thisisduetoaraceconditioninthelogicforthepseudo‐synchronouscounterusedinthegray‐codecountgenerator,whichhasbeenreproducedinsimulation.Thisprobablywon’taffectthespatialresolution,sincewewillbecombiningmanyofthenanosecondsampleswithaFIRfilter,effectivelyincreasingthesampleresolution.Second,thecombinedtotalpowerdissipationofthetwoboardscomesto11.4W(includingthetwoFPGAsandthevoltageregulatorinefficiencies).Wehavenotyetdesignedamethodtoremovethispowerinvacuum,sowewanttodesigntheGRAPHchiptoachievealowerpowerlevel,andthe100‐mmenclosureboxandboardstoextractthispowertoabetterheatsink.WealsowanttocorrectthelayoutissuethatcausesthepedestalfeaturesseeninFig.9.ThefourthissuewiththeHalfGRAPHarchitectureisitsactualthroughput.TheWilkinsonclockwasonlyreliable to250MHz.Themultiplexingcircuit required threereads (rather than two) to readoutallchannelsandthereadoutclockcouldonlyoperateupto100MHz.Atbest,thethroughputcouldreach50to100kHz.ThisisalsoaddressedbytheGRAPHchip(below).

Fig. 9. Examples of raw mode output from all 160 channels (vertical/spatial axis) of 128 ns of data(horizontal/temporalaxis).Left:RawdatafromasingleamplifiedphotoneventshowingdetectorXaxisinthetopsetof80channelsanddetectorYaxisinthebottom80channels.LeftCenter:Pedestalofthosememorylocations.RightCenter:Pedestal‐subtractedimageshowingjustthephotondata.Right:3‐Drepresentationoftherightcenterimage,withchargeeventsoccupying~8channelsinboththeXandYaxes.

50‐mmXSDetector

Therearetwokeyaspectstoourflight‐like50‐mmand100‐mmXSdetectordesigns.Thefirstisaphotolithographic and laser‐cut XS anode design made with polyimide. Polyimide’s dielectricconstantislowerthanthatofaluminaceramicbyafactorofthree,resultinginlowerindividualstripcapacitanceandthusloweramplifiernoise.Thetopstrippatternisfirstetchedinthecopper,afterwhichalaserablatesthematerialbetweenthestrips.Thistoplayeristhenbondedtothebottomstrippatternetchedonamuchthickerpolyimidesubstrate.TheinputsideoftheanodeisshowninFig.10,installedinthe50‐mmXSdetector,withmeasuredstripcapacitancesmatchingourdesignmodel. Outputs from the 80 × 80 strips go through a hermetic seal consisting of 2 × 80‐pinconnectorssealedwithvacuumepoxy(Fig.11).TheotherkeyaspectofourdetectorisusingaKovarandceramicbrazedbodytomounttheMCPsovertheXSanode.Thistechniqueisusedinvacuum‐image‐tubeconstructiontomakeastrong,robust,andcleandetectorthatcansurvivelaunchstress.

86

Figure10showsthebrazedbodyofthe50‐mmdetectormountedoverourXSanodeontoavacuumbackplatewiththreehighvoltage(HV)feedthroughs(MCPsremovedtoshowtheanodebelow).

Fig. 10. View of windowless 50‐mm XS detectormounted on a vacuum flange with three HV feed‐throughsshowingtheXSanode(MCPsremoved).

Fig.11.Externalsideofdetectorshowing80‐contactfeed‐throughs(×2)sealedwithepoxyanda64‐channelpreampboardpluggedintooneaxis.

ImagingResultswith50‐mmDetector

Sincethereadoutfirmwareisstillbeingdeveloped,weusedourexistingPXS2electronicsand64‐channelamplifierboardstoreadoutthecentral64 × 64stripsofthis80 × 80XSanode.Thiscorrespondstoacentralactiveareaof40 mm × 40 mm.TheresultsbelowuseastackoftwoMCPsfromPhotonisUSAwith10‐μm pores on 12.5‐μm centers, 53.7 mm × 53.7 mm, and 60:1 L/d ratio (600‐μm‐thick each). Wemeasured the spatial resolution and linearity and acquired flat fields to measure the UV responseuniformityto183‐nmlightfromaHgpen‐raylamp.

Tomeasurethespatialresolution,weusedapinholemaskgridmounteddirectlyontheinputMCP.Thepinholesare10μmindiameterandspaced1mmapartonasquaregrid(Fig.12),anexcellentmethodofsamplingthePointSpreadFunction(PSF)acrossthefieldofview.Tomeasurethespatialresolution,wehadtobintheX,Yeventdatato8192 × 8192(5‐μmpixels)toresolvethePSF.Theinset of Fig. 13 shows the X dimension PSF (top strips) of a single pinhole. The average spatialresolutionsintheXandYdimensionswere17.5μmFWHMand22μmFWHM,respectively.

Fig.12.UVimageofpinholemaskof10‐μmholeson1‐mmsquaregrid.InsetshowsthePSFofasingleholewithresolutionof~4pixelsFWHM(20μm).

Fig. 13. Residuals to linear fit of derived pinholeposition across central 40 mm of the 50‐mm XSdetector.

87

Thedetectorlinearitycanalsobemeasuredwiththepinholemaskdata,asthepinholesareuniformlyspacedat1mm.Figure13showsresiduals(inμm)toalinearfittopixelpositionofthepinholevs.pinhole number. The ±15‐μmdeviation from zero is smaller than the detector spatial resolutionmeasuredabove,andcomparabletothehexagonal12.5‐μmporespacing.Thismeasurementatteststotheaccuracyofthephotolithographicanodestripregularity.

EnvironmentalTestingof50‐mmand100mmDetector

One of themain goals of this SAT programwas to advance the TRL of the 50‐mm and 100‐mmdetectorsystems.Sincethenew50‐mmdetectoronitsbackplateisanewdesign,wedecidedtogobeyondtestingitsperformanceonthebench,andconfirmitsperformanceattemperatureextremes,aswellasitsabilitytosurvivetheg‐forcesofstandardrocketlaunches.

The detector and front‐end electronics were mounted onto a vacuum housing inside a thermalchamberandcycledfrom–30°Cto+45°C(Fig.14).Wemeasureddetectorperformancefrom–15°Cto+45°C,allowingthetemperaturetoequilibrateforaboutonehourateachtestlevelbeforetakingadeep‐UVflat‐fieldimage.Inallmeasures – resolution,backgroundrate,anddynamicrange – thedetectorworkedflawlessly.

Fig.14.DetectortemperatureandMCPresistanceduringthermaltestofXSdetector.

WealsousedtheSSLvibrationtabletovibratebothofthe50‐mmand100‐mmdetectors(withoutelectronics) to the 14.1 g rms level specified by GEVS. Figure 15 shows the 100‐mm detectorbackplate(airside)onthetestfixture,andTable1showsthevibrationspectrumlevelsappliedtothedetector.UV imagingperformancewas identicalbeforeandaftervibration,showingthisnewdetectordesignisreadyforflightvibrations.

Fig.15.100‐mmXSdetectormountedonvibrationtableatSSL.

Frequency (Hz) g2/Hz dB/Octave

88

20 0.026 5.97

50 0.16 0

800 0.16 -5.97

2000 0.026 –

Table1.Vibrationfrequencyspectrum(14.1grms).

RadiationTestingoftheCSAv3andtheHalfGRAPH2ASIC

TotesttheresilienceofthenewASICstoTID,wespenttwodaysattheUniversityofMassachusettsLowell Radiation facility. The facility allows exposure of microelectronics to 60Co gamma rays atcontrollable rates and durations, achieving a calibrated dose traceable to NIST standards using acalibratedBrukerBiospindosimeter.FortheCSAv3,wetestedtheASICsbothunpoweredandbattery‐poweredduringtheexposure,firstinstepsof25‐kRaddose,withperformancetestinginbetween.Wemeasuredbiascurrent,pulsegain,andnoiseofeverychanneloftheeightchipsanddidnotdetectachange(Fig.16)orfailureuptoaTIDof486kRad,moredosethanwewouldexpectforamissiontoJupiter.FortheseparateexposureoftheHalfGRAPH2,ourperformancetestconsistedofdrivingtheinputwithalinearvoltagerampandmeasuringthelinearincreaseinthedigitaloutputcodesforallchannels,andagain,didnotseeachangeinperformanceto236kRad.WehavenotyettestedeitherASICwithenergeticionstoseeiftheysufferfromsingle‐eventupsets(SEUs)orlatch‐ups.

Fig.16.Gain(left)andrmsnoise(right)oftheCSAv3CSAASICsvs.TIDfrom60Cogammarays.Sixchipswith16channelseachweretestedandthestandarddeviationwasderivedfromthevariationofthechannels.Notallchipsreceivedthefulldose.

The100 mm × 100 mmXSDetector

The latest SATprogram scales theXS technology to larger format size plus other improvements toincreaseitsTRL.Wedoublethesize(quadruplethearea)ofaXSdetectorto100mm×100mmactivearea,whiledemonstratinghighTRLwithvibrationandthermaltests.Thislargersizeisnottargetingaspecificmission,butdemonstratesalarge‐formatdesignthatcanbescaledeasilytoasimilarformatwhile reducing risk. Alongwith a new detectormechanical design, we combine our working ASICdesignsintoasingleASIC(the“GRAPH”chip)using130‐nmCMOStechnologytoreducereadoutvolume,mass,andcomplexity.Assuch,weareconvertingalloursub‐circuitlibrariesusedinthe250‐nmprocessHalfGRAPH2totheTSMC130‐nmprocess,andthencombiningwiththeCSAv3intotheGRAPHASIC(TheCSAv3wasalreadydesignedinthe130‐nmprocess).WealsodemonstratetheradiationhardnessoftheexistingASICdesignswithenergeticparticles,andmaintainthishardnessinthenewGRAPHASIC.

The100 mm × 100 mmMechanicalDesign

89

Thebrazed‐bodydesign(Fig.17)isscaledupbyafactorof~2,withanoutputaperturebelowtheMCPsof103mmwhiletheactivefieldofviewissetbya100 × 100openmaskof25‐µmthicknessbetweentopandbottomMCPs.ItisdesignedtoholdsquarePhotonisUSAMCPsupto110mminsize.TheXSanodedesign(Fig.18)isalsoscaledto160stripsperaxis,butwithaslightincreaseinstrippitchfromthe635µmofthe50‐mmanodeto645µmforthe100‐mmanode.Thisresultsina103 mm 103 mmanodewithenoughstripsattheedgestosamplethe100‐mmfieldofviewfully.Theanodeoutputsarenowdistributedonall fourcornersof theanode ina“pinwheel” fashionsothemechanicalcutoutsthroughthevacuumbackplatecanbemoresymmetric for increasedstrength. Italso facilitates theelectronicsboardlayoutwheretheGRAPHchipscanbespreadoutontheprintedcircuitboard.

Fig. 17. View of windowless 100‐mm XS detectormounted on a vacuum flange with two HV feed‐throughsshowingtheXSanode(MCPsremoved).

Fig.18.Fabricated100‐mmXSanode.Notethefourdistributedoutputconnectorsalongtheedgesoftheanode.

103‐mmXSAnode

Wefabricatedthenew103‐mmsquareanodewiththenewpinwheeloutputarrangementusingthesametechniquesoflasercuttingandbondingofapolyimidetoplayerusedwiththe50‐mmanode.Resistance and capacitivemeasurementswere as expected,with about a 2× increase due to thedoubledstriplength.ThisshouldincreaseCSAreadoutnoisefromthatofthe50‐mmanodebyonly50%giventhebaseline,no‐loadnoiseoftheamps.Figure18showsthenew,gold‐coatedanode.

100‐mmDetectorBrazedBodyandBackplate

Thebrazedbodydesign ismostlyascaled‐upversionofthe50‐mmdesign,butthebodyismoresymmetric as the signal outputs are now coming off all four sides of the anode, so thismust beaccommodatedbyaddingthecantileveredshelftomounttheMCPsdirectlyovertheanodecenter.

Thebackplatenowhasfourcutoutstoletthefoursignalconnectors(80channelseach)through,againwith an epoxied hermetic seal. The pinwheel design puts much more steel between the cutouts,increasing the strength of the backplate.Wewelded two commercial‐off‐the‐shelf (COTS) safeHV(SHV)connectorstoprovidetheMCPbiashighvoltage.Figure17showsthebrazedbodyattachedtothebackplatewiththeXSanodeshowing(MCPsnotinstalled).

GRAPHASIC

Givenoursuccesswith theCSAv3 in130‐nmtechnologyanddesign libraries,weareplanning tocombine the amplifier and digital converter on the same chip using this technology. Designtechniqueswill be employed to shield the low‐noise inputs from the high‐speed digital outputs.However,tohedgeourbets,weareusingmultiplexorstoallowbypassingoftheGRAPHCSAcircuitincasetheinducednoiseprovesproblematic.TheresultingASICwouldthenactlikeanimprovedHalfGRAPH,whichcouldbeusedwiththeexistingCSAv3.

90

Figure19isasimplifiedschematicwithanoverall16‐channelGRAPHASICdesign,includingmoredetailforonechannelandthenmoredetailforonesamplingcell(2048analogsamplesperchannel).EachGRAPHchannelhasaninputCSA(whichcanbebypassed),atriggercircuittodetectarisingedgepulse,andanADCwhoseoutputsaremultiplexedto fastLVDSlinestoadownstreamFPGAcontroller.Eachchannelhasatrigger,andtwoconsecutivetriggersareOR‐edintoasingleoutputpin,signalingtheFPGAofaneventandonwhatchannelsthechargehaslanded.Foreachchannel,theGRAPHcorecircuitrycontinuouslysamplestheanalogoutputoftheCSA,storingthevoltageonthesamplecellcapacitors.Thetimebasecontrolsthesamplingswitches,andthereare64samplesperwindow,16windowsperbankand2banks.After2048samples,thetimebaseoverwritesthesamplesstartingatthebeginning.Whileonebankisbeingfilledwithnewsamples,theotherbankgoesthroughaWilkinsonconversion,wherebyananaloglinearrampisinputtoacomparatorwhilea digital register is being incrementedwith a grey‐code binary counter.When the ramp voltageexceedsthestoredsamplevoltage,thecomparatoroutputlatchesthecounter,andthedigitalvalueisproportionaltotheinputvoltage,henceanADC.

ADC LV

DS

(12 pairs)

ADC

MU X

Triggers

ADC

×16

Readout control Trigger

16-Ch. GRAPH ASIC

Address

Timebase Ext

Dig. outRamp

Wilk grey code

Signal 12b counterLatch

Sample

Single Sample of Channel (64 per window):

Reset

Charge amplifier

Amplifier bypass

1Bank = 16 windows=1024 samples

64 samples

64 samples

Bank A

64 samples

64 samples

64 samples

Bank B

64 samples

Sampling Timebase

GRAPHCORE

Wilk. Conversion Controller A

Wilk. Conversion Controller B

Ramp Wilk grey Analog “SPY”

Ramp

Convert or Sample

Sync (A/B)

CSA_out

Input Read address

12b parallel

Readout Module

Latch Rd_Cl Start Address

OutputMUX

(16 ch.) ×16

Single Channel of GRAPH:

91

Fig.19.FunctionalSchematicofGRAPHASIC.Topleftistheoverall16‐channelASICwhereeachchannelhasanamplifierandaWilkinsonADC(topright),whoseoutputismultiplexedtofastLVDSlinestoadownstreamFPGAcontroller.ThebottomschematicisamoredetaileddescriptionofasinglechannelshowingtheGRAPHcorewhichsamplestheinputanalogsignal,andconvertstodigitalvalueswhoseregistersareaddressable.Amoredetaileddescriptionisfoundinthetext.

Theotherbankthatisundergoingsamplinghasinitsregistersthedigitalvaluesfromthepreviousconversion,andcanbereadout.TheFPGAinstigatesthisprocessbysendingthechannel,window,andsampleaddressseriallytothereadoutcontrolwhichtransfersthedatatotheoutputMUXforfasttransmissiontotheFPGAofacontiguoussetoftemporalsamplevaluesstoredintheregisters.Whileloadinganewaddressviatheserialport,theinitiallysetvalueautomaticallyincrements.Thetimebasecontrolstheping‐pongsynchronizationofthetwobanks,onesampling/readoutwhiletheotherconverting.Theanalog“SPY”outputenablesprobingoftheanalogCSAoutputsorramps,usefulforoptimizationofsignal‐to‐noiseratio.NotshownaretheslowserialinterfacetotheFPGAusedtosetallstaticparametersofthechip.

PathForwardAsofJuly1,theGRAPHASICchipdesignisfinishedandundergoingfinalchecksandsimulationstotheTMSCfoundryattheendofJuly.Fabricationshouldtakeapproximatelythreemonths.Duringthattime,testboardswillbebuiltattheUniversityofHawaiitomeasureitsperformance,andreadoutboardsthatincorporatethenewdesignwillbestartedinBerkeley,takingadvantageofthelowerpowerandsimplifiedlayout,providedbythe2‐for‐1deceaseinchipcount.

Thiseffortisdirectlytransferableinthedevelopmentofa100‐mmdetectorreadoutusingthesamelogicasthe50‐mmsystemscaledupbyafactoroftwoineachdimension.Infact,someaspectsareeasier in that thereare twiceasmanycomponents,but four timesthearea.Thismightmakethethermaldesigneasieraswell.Boardswillhavetobelaidoutandelectronicsboxesfabricatedforthenewsize,thistimeincorporatingmethodsforbetterheatextraction.

References[1] J.Vallerga,J.McPhate,A.Tremsin,O.Siegmund,R.Raffanti,H.Cumming,A.Seljak,V.Virta,and

G.Varner,“Developmentofaflightqualified100×100mmMCPUVdetectorusingadvancedcrossstripanodesandassociatedASICelectronics,”Proc.SPIE,9905,99053F(2016)

[2] K.Hoadley,K.France,N.Nell,R.Kane,T.B.Schultz,M.Beasley,J.C.Green,J.R.Kulow,E.Kersgaard,and B.T. Fleming, “The assembly, calibration, and preliminary results from the Colorado highresolutionEchellestellarspectrograph(CHESS),”Proc.SPIE,9144(2014)

[3] J.Vallerga,R.Raffanti,A.Tremsin,O.Siegmund,J.McPhate,andG.Varner,“Large‐formathigh‐spatial‐resolutioncross‐stripreadoutMCPdetectorsforUVastronomy,”SPIE,7732(2010)

[4] X.Michalet,R.A.Colyer,J.Antelman,O.H.W.Siegmund,A.Tremsin,J.V.Vallerga,andS.Weiss,“Single‐quantum‐dot imagingwithaphoton‐countingcamera,”CurrentPharmaceuticalBiotechnology,10(5),543‐558(2009)

[5]F.B.Berendse,R.G.Cruddace,M.P.Kowalski,D.J.Yentis,W.R.Hunter,G.G.Fritz,O.Siegmund,K. Heidemann, R. Lenke, A. Seifert, and T.W. Barbee Jr., “The joint astrophysical plasmadynamicexperimentextremeultravioletspectrometer:resolvingpower,”SPIEConferenceSeries,6266,31(2006)

[6] A.Seljak,H.S.Cummings,G.Varner,J.Vallerga,R.Raffanti,andV.Virta,“Afast,lowpowerandlownoisechargesensitiveamplifierASICforaUVimagingsinglephotondetector,”JINST,Volume12,Issue04,pp.T04007(2017)

[7] A.Seljak,H.S.Cummings,G.Varner,J.Vallerga,R.Raffanti,andV.Virta,“PrototypereadoutsystemforamultiMpixelsUV single‐photon imagingdetectorcapableof space flightoperation,” JINST,Volume13,Issue02,pp.T02003(2018)

92

Foradditionalinformation,contactJohnVallerga:jvv@ssl.berkeley.edu

93

UltrasensitiveBolometersforFar‐IRSpectroscopyattheBackgroundLimitPreparedby:CharlesM.(Matt)Bradford(JPL/Caltech)

SummaryThis isa three‐yearStrategicAstrophysicsTechnology (SAT)awardbegun inMarch2018(whenfundswereavailabletothePI).Theaimistodemonstratereadinessforaflightprogramwiththeworld’s most sensitive bolometers: transition‐edge‐sensed (TES) devices which meet therequirementforbackground‐limitedspectroscopyfromcoldspacetelescopes.

Wehavedevelopedasetofrequirementsandgoalsforourdetectordevelopmentprogram.Weareworkingtowarddevicesthat:

demonstrateaper‐pixelnoiseequivalentpower(NEP) lessthan10−19WHz−1/2withagoalof3×10−20WHz−1/2;

providesufficientspeedofresponse, f3dB=5Hz(requirement),15Hz(goal), tobeuseful inarangeofinstrumentconfigurationsforSPICA,theOriginsSpaceTelescope(OST),andapotentialNASAfar‐IRprobe‐classmission;

arereadoutwithafrequency‐domainmultiplexingschemescalabletothousandsofdetectorsat50mKandcompatiblewithspaceflight;and

canmaintainhighdutycycleinthefaceofcosmic‐rayinteractionsinspace.

The first challenge is the sensitivity, which requires very long and thin isolation legs for thebolometers.Thishasbeenourprimarythrustthusfar.

BackgroundFar‐IRwavelengthsareuniquelywell‐suitedtofundamentalquestionsthatguideNASAastrophysicssuchas:“HowdoestheUniversework?”and“Howdidwegethere?”Far‐IRmeasurementsarevitalfor probing the inner workings of galaxies, which are typically obscured by dust at opticalwavelengths. The Spitzer and Herschel missions have shown us that the dust obscuration is soprevalent that most of the energy ever produced in galaxies by stars and accreting black holesthrough itshistoryhasbeenabsorbedandre‐radiated.Similarly,planetsassemble fromopticallythick diskswhich cool primarily in the far‐IR.Most of our information about the dust‐obscureduniversecomesfrombroadband‐imagingdatasets.Butthese2‐Ddatasetsfallshortwhenitcomestoproviding astrophysical insight. Spectroscopy is a farmore powerful tool, providing quantitativediagnosticstorevealthemachineryattheheartofgalaxiesandformingplanetarysystems.

Astronomers are aware of this toolset, but sensitive far‐IR spectroscopy has remained elusivebecauseitrequiresadvanced,exquisitedetectorsensitivitiesinthefar‐IR,aregimewhichdoesnotbenefitfromthecommercial/industrialinfrastructure.Sensitivefar‐IRdetectorsandreadoutsmustbedevelopedandpurpose‐builtbyNASAscientistsandtechnologistsforfar‐IRastrophysics.

WithalatestartinFY2018,wehavedevelopeddesignsforprototypearraysofatransition‐edge‐sensed(TES)bolometer.WedemonstratedanewreleaseprocesswiththesedesignsintheMicrodevices Laboratory.We designed a cryogenic electronics system to support thetesting and specified and procured a new dilution refrigerator to house the experiment

94

This work on low‐noise‐equivalent‐power (NEP) detector development is directly applicable tofuture far‐IR spacemissions such as SPICA [1] and theOrigins Space Telescope (OST)[2]. Thesemissionsfeaturecryogenictelescopes,whichwhencombinedwithdispersivespectrographsatthebackgroundlimit(hereafterreferredtoasBLISS‐type)createpowerfulspectroscopicfacilitieswith1–4orders‐of‐magnitudesensitivityimprovementoverthecurrentstateoftheart.Figure1showsthesepossiblegains,alongwithaconceptforSPICA,thenearest‐termopportunityforsensitivefar‐IR spectroscopy, and the quantitative sensitivity gains possible. This advance will bring far‐IRsensitivitiesintoparitywiththoseofpowerfulflagshipprogramsatshorterandlongerwavelengths,JWSTandALMA,butprovidinguniqueaccesstothemostdust‐enshroudedstarformationandblackholegrowthintheuniverse’sfirstbillionyears.

Fig.1.Left:SPICAconcept,a2.5‐metertelescopeactivelycooledtobelow8K,withaplannedfive‐yearlifetimeattheEarth‐sunL2haloorbit.Right:Far‐IRspectralsensitivities.Aspectrometeronacoldtelescopewillbe100–1000timesmoresensitivethanpresent‐dayfar‐IRfacilities,andobservingspeedscalesastheinversesquareoftheplottedsensitivity.TheSAFARIcurveassumesdetectorNEPof1×10−19WHz−1/2,therequirementadoptedforthisprogram.OST(previouslytheFar‐IRorFIRSurveyor)sensitivityplottedhereassumesa4×6‐mtelescopewithdetectorsoperatingat2×10−20WHz−1/2,justabitbetterthanthisprogram’sgoalsensitivity.Modelgalaxyspectraareover‐plottedatarangeofredshifts,fromtheepochofreionizationtothepresentday.Inadditiontothefine‐structurelines,thesespectrometersalsohaveexcellentsensitivitytobroadmid‐IRfeaturessuchasthepolycyclicaromatichydrocarbon(PAH)bands—mid‐IRfluxesarescaledtoshowthesensitivitywhenbinnedtoR=100.

Theper‐pixelsensitivityrequirementdescribedaboveexceedsthestateoftheartforflightfar‐IRbolometers(flownontheSPIREinstrumentonHerschel)byafactorof~300.Wewillachievethissensitivity increase through a combination of lower temperature (we target 60‐80‐mK detectoroperationcooledwitha50‐mKbath)andmoreextremeisolationleggeometry(wearebaselining1000‐micron‐longby0.4‐micronwideby0.25‐micronwidebolometerlegs).Thelowertemperaturerequiresdevelopmentofnewsuperconductingmaterial,andtheextremeisolationlegsrequirenewfabrication procedures. These have been the focus of our work in FY 2018. An additional keychallengewe are beginning to address is integrating the new detectors with a superconductingmultiplexing(MUX)readoutsystem,whichenablesarrayswiththousandsofpixels.ForthiswearecollaboratingwithadetectorgroupinSRON(theNetherlandsInstituteforSpaceResearch)thathasdevelopedanRFfrequency‐domainmultiplexer(f‐MUX),showninFig.2.

95

Fig.2.Schematicofthefrequency‐domainmultiplexerwewillusetoreadouttheTESbolometers.TheschematicisprovidedbyourpartnersatSRON[3],butthisreadoutissimilartothatbeingusedforthebolometerinstrumentsfortheSouthPoleTelescope(SPT)[4]andPolarbear.Acombofbiasfrequenciesissentonasingleline,witheachfrequencyaccessingasingleTESthroughanLCresonator.CurrentsaresummedataSuperconductingQuantuminterferencedevice (SQUID) input,andabaseband feedback systemactivelynulls theSQUID flux tomaintainlinearity.Ourultimategoalistodemonstratethissystemat176× multiplexingformat,butourfirststepunderwaynowisan18‐channelsystemreadoutpixelbypixelwithasingle‐toneRFbias.

ObjectivesandMilestonesOurhigh‐levelobjectivesarepresentedaboveinthefirstparagraphofthereport.Wehavedevelopedastrategyforachievingthesegoals.Ourinitialtestingwilluseadarktestpackage(wecallv0)withan18‐channelpassive‐LC‐filterchipfromourSRONcollaborators,andanewprototypebolometerarraywearebuilding.Thispackagewillbecooledinanewrefrigeratorto50mK,andbiasedwithabenchtopRFsourcemeasuredwithacommercialSQUIDsystem.ThiswillallowustodebugstraylightandstrayRFproblems,aswellasdevelopexperiencewiththef‐MUXsystem.Wewillthenpushtoalarger‐format(176×,asenvisionedforSPICASAFARI)MUXdemonstration,alsodarkv1‐D.Theworkwillculminatewithopticalsensitivitymeasurementswiththisarraypackage,but includingfeedhornscouplingtothebolometers.

Wehaveidentifiedthefollowingmilestoneswithtargetcompletiondates:

Specifyandordertestfacilitycryostat(CompletedMarch2018,uponreceiptoffunding); Developnewwet‐releaseprocessforbolometerstructures(CompletedJune2018,seebelow); Develop design for RF cables, custom filter boards, order components, and test equipment

(CompletedJuly2018); Designv0mechanicaltestenclosureandfabricate(underway,expectSeptember2018,seebelow); CommissionfridgeandbeginRvsTtestingofbilayerfilms(deliveryexpectedinAugust,testing

underwayinOctober2018); BeginTEStestingwithv0setupincryostat(November2018); DemonstratedarkNEPatorbelow2×10‐19WHz‐0.5and≥50%yieldwithv0setup(April2019); Designv1 focalplanepackage, includingMUXchips,backshorts, feedhorncoupling, flight‐like

design(February2019); Procure/buildmulti‐channelf‐MUXwarmelectronicsincollaborationwithSRON.(May2019); Dark‐testv1packageunderwaywithmulti‐channelelectronics(June2019);

96

Demonstratewithv1packageNEP≤2×10‐19WHz‐0.5(goal1×10‐19)and≥50%yield;understandyield(December2019);

Verifyexpectedcosmic‐rayinteractionusingv1package(March2020);and Testopticallyv1packageunderwaywithcryogenicblackbody;demonstrateopticalefficiency≥70%

withnofeedhornimpactonarrayyield(June2020).

ProgressandAccomplishmentsSinceourfundingarrivedinMarch2018,wecompletedthefirstthreemilestonesdescribedabove.Figures3,4,and5showournewprototypebolometerstructuresunderdevelopmentinthemicro‐deviceslab.Developingawet‐releaseprocessisessentialtopracticallow‐NEPdevices,becauseourpreviousdry‐plasma‐etchedbolometers,whiletheydemonstratedsufficientthermalisolationandNEP,hadexcessheatcapacitycausinglarge(~1sec)timeconstants.Thechallengeofthewet‐releaseisthatliquidsolventstendtodestroythefragilestructuresastheyemerge.Weareexploringbotharangeofwax‐mounting/demountingschemesandrelease‐supportbuttresseswhichcanberemovedwithalasertrimmerafterrelease.AsFig.4shows,aviableschemehasemerged.

Fig.3.MasklayoutnotesforfirstprototypeTESarray.Weareexploringbothmeanderedandstraightbeams,anddeviceswithandwithoutpost‐releaselasertrimmingapproachtoaidinthesurvivalofthefragiledevices.

C1 C2 C3 C4 C5

R1

R2

R3

R4

C1 C2 C3 C4 C5

R1 TES; beam‐straight,0.18 x 0.4 x 1000 um3

TES; beam‐meander,0.18 x 0.4 x 2000 um3

Small TES; beam‐straight,0.18 x 0.4 x 1000 um3

JTD; beam‐straight,0.18 x 0.4 x 1000 um3

JTD; beam‐meander,0.18 x 0.4 x 2000 um3

R2 TES; beam‐straight,0.18 x 0.4 x 1000 um3

TES; beam‐meander,0.18 x 0.4 x 2000 um3

Small TES; beam‐straight,0.18 x 0.4 x 1000 um3

JTD; beam‐straight,0.18 x 0.4 x 1000 um3

JTD; beam‐meander,0.18 x 0.4 x 2000 um3

R3 TES; beam‐straight,0.18 x 2 x 1000 um3

TES; beam‐meander,0.18 x 2 x 2000 um3

Small TES; beam‐straight,0.18 x 2 x 1000 um3

JTD; beam‐straight,0.18 x 2 x 1000 um3

JTD; beam‐meander,0.18 x 2 x 2000 um3

R4 TES; beam‐straight,0.18 x 2 x 1000 um3

TES; beam‐meander,0.18 x 2 x 2000 um3

Small TES; beam‐straight,0.18 x 2 x 1000 um3

JTD; beam‐straight,0.18 x 2 x 1000 um3

JTD; beam‐meander,0.18 x 2 x 2000 um3

97

Fig.4.Examplebolometerteststructuresafterrelease.Devicesshownwithbuttressesbothbefore(above)andafter(below)lasertrimmingofsupportbuttresses.

100um

100um

0.18um thick x 2um wide x 1000mm long (Straight)

0.18um thick x 2um wide x 1000mm long (Straight)

0.18um thick x 2um wide x 1000mm long (Straight)100um

100um0.18um thick x 0.4um wide x 1000mm long (Straight)

98

Fig.5.Layoutofourv0bolometertestexperiment.ThedetectorarrayandSQUIDcarrierarefabricatedinthemicrodevices labat JPL.TheLC chipwasbuiltbyour collaboratorsatSRON.Theprintedwireboard (PWB)includesanMDMconnectortoa loomoftwisted‐pairwires.TheSQUIDwillbe fromMagnicon(acommercialSQUID vendor),and includes itsown readout system outside the cryostat.RF interference ismitigatedusingdouble‐poleRCfiltersoneachconductorasitcomesthroughthePWB,andthealuminumenclosureincludesapartitionnearthisfilterarraytoseparatethe‘noisy’regionwherethecableenterstheboxfromthe‘clean’volumewherethebolometersare.

AnotherkeyaccomplishmentsinceonsetoffundinginMarchhasbeentodesigntheinitialbolometertestapparatus(layoutinFig.5).Thisassemblywillbehousedinanaluminumboxandcooledto50mKin our dilution refrigerator. The electrical system consists of four chips interconnectedwithwire‐bonds:thebolometerarrayonasilicondie,theexisting18‐channelLCchipfromSRON,acustomSQUIDcarrierchippatternedonsilicon,andacustomPWBwhichinterfaceswiththetwisted‐paircableloomfromthewarmsideofthecryostat.Alltheseelementshavebeendesignedandarebeingfabricated(orprocured,inthecaseofthePWB).

Inadditiontothev0experimentboxshowninFig.5,thecryostatalsohasafilteredbreakoutPWBatthewarmvacuumshellofthecryostat.Thisfilteredboardalsousesapartitiongalvanicallyconnectedto the cryostat jacket, so that the jacket forms anRF‐tight Faraday cagewith all signals passingthroughdouble‐polefilters.

PathForwardStillunderwayinFY2018aredeliveryandcommissioningoftherefrigerator(expectedAugust2018),andthefirstmeasurementsofresistanceversestemperature(RvsT)forthermistorfilms(unreleased).BytheendofFY2018,wewillhaveafullboxmechanicaldesigntoaccompanythev0functionalchiplayout(Fig.5),soweexpecttomeasurebolometersearlyinFY2019.

References[1] P.Roelfsema et al., “SPICA–a large cryogenic infrared space telescopeUnveiling theobscured

Universe,”PublicationsoftheAstronomicalSocietyofAustralia,SPICASpecialIssue(2018)[2] C.Battersbyetal.,“TheOriginsSpaceTelescope.,”NatureAstronomy.v.2(August2018)

Printed Wire Board(PWB)

16.4 mm

15.54 m

mChip Layout

Detector Array

LC ChipSQUIDCarrier

SQUID

10.7 mm

16.5 mm

0.2 mm gap

4.6 m

m

20.0 m

m

0.2 mm gap

Holes in substrateCan add more ‘meat’ 4 clips

99

[3] R.Hartog,M.D.Audley,J.Beyer,D.Boersma,M.Bruijn,L.Gottardi,H.Hoevers,R.Hou,G.Keizer,P.Khosropanah,M.Kiviranta,P.Korte,J.Kuur,B.‐J.Leeuwen,A.C.T.Nieuwen‐huizen,andP.Winden,“Low‐noise readout of TES detectorswith baseband feedback frequency domainmultiplexing,”JournalofLowTemperaturePhysics,167(5),652–657(2012)

[4] M.A.Dobbs,M.Lueker,K.A.Aird,A.N.Bender,B.A.Benson,L.E.Bleem,J.E.Carlstrom,C.L.Chang,H.‐M.Cho,J.Clarke,T.M.Crawford,A.T.Crites,D.I.Flanigan,T.deHaan,E.M.George,N.W.Halverson,W.L.Holzapfel,J.D.Hrubes,B.R.Johnson,J.Joseph,R.Keisler,J.Kennedy,Z.Kermish,T.M.Lanting,A.T.Lee,E.M.Leitch,D.Luong‐Van,J.J.McMahon,J.Mehl,S.S.Meyer,T.E.Montroy,S.Padin,T.Plagge,C.Pryke,P.L.Richards,J.E.Ruhl,K.K.Schaffer,D.Schwan,E.Shirokoff,H.G.Spieler,Z.Staniszewski,A.A. Stark, K. Vanderlinde, J.D. Vieira, C. Vu, B. Westbrook, and R. Williamson, “Frequencymultiplexedsuperconductingquantum interferencedevicereadoutof largebolometerarrays forcosmicmicrowavebackgroundmeasurements,”ReviewofScientificInstruments,83(7)(2012)

[5] A.D.Beyer,M.E.Kenyon,P.M.Echternach,T.Chui,B.‐H.Eom,P.K.Day,J.J.Bock,W.A.Holmes,andC.M.Bradford,“Ultra‐sensitiveTransition‐EdgeSensorsfortheBackgroundLimitedInfrared/Sub‐mmSpectrograph(BLISS),”JournalofLowTemperaturePhysics,p.143(December2011)

Foradditionalinformation,contactMattBradford:matt.bradford@jpl.nasa.gov

100

RaisingtheTechnologyReadinessLevelof4.7‐THzLocalOscillatorsPreparedby:QingHu(MIT)

SummaryThe 63‐μm (4.744 THz) [OI] fine‐structure line is an important spectral line for astrophysicsobservations.Despite thegreatpotential, however, astrophysical observationof the [OI] linehasrarelybeenperformedbecausethe4.744‐THzfrequencyisbeyondthereachofmostimplementedlocaloscillators(LOs)insensitiveheterodynereceiversinvolvingcryogenicmixers.Inthisthree‐yearNASAStrategicAstrophysicsTechnology(SAT)project,weplantoraisetheTechnologyReadinessLevel(TRL)ofTHzquantum‐cascadelasers(QCLs)forLOapplicationsto5orbeyond,bridgingthe“mid‐TRL gap” between a promising enabling technology and a mission‐ready component. TheprojectstartedinMarch2016andisplannedtoconcludeinFebruary2019.

Theobjectivewillbeachievedbydevelopingantenna‐coupled3rd‐orderdistributedfeedback(DFB)lasingstructures,and inparallel,designingandgrowinghigh‐performancequantum‐cascadegainmediawithpeakfrequencyaround4.7THz.Bytheendoftheproject,wewilldevelopsingle‐modeDFBlaserswithfrequencywithin10GHzofthetarget4.744‐THzline,continuouswave(cw)outputpowerof~5mW,wall‐plugpowerefficiency(WPE)of~0.5%atanoperatingtemperatureof~40K,andbeampatternsnarrowerthan1010degrees2.

The project will be carried out in the Principal Investigator’s (PI) laboratory at MIT, with themolecularbeamepitaxy(MBE)wafersprovidedbyDr.JohnRenoatSandiaNationalLaboratoriesthroughauseragreement.

Duringthepreviousyear,wecompletedthedesignofantenna‐coupled3rd‐orderDFBstructures,andgeneratedfabricationmasksbasedonthedesign.Wecompletedthefabricationprocessandtestedthedevices.Thelasingfrequency(4.74xTHz)andpowerlevels(>1mW)alreadymeetourfinalgoal.However,thepeakfrequencyofthegainmediumistoofarfrom4.744THz,sodeviceperformancesuffered.Themaximumoperatingtemperatureisonly~30K,whichdoesnotmeettherequirementofanactualobservatory,butitissufficientforlabtestingat~10K.SeveraldeviceshavebeenshippedtoourcollaboratoratSRONtotesttheircomponents(beamsplittersandhot‐electronbolometer,HEB, mixers) for the recently initiated the Gal/Xgal U/LDB Spectroscopic/Stratospheric THzObservatory (GUSTO) project. We have also developed a novel scheme to electrically tune thefrequencyby~10GHz(thisisatadifferentcenterfrequencyof~3.9THz,butthemethodcanalsobeappliedto4.7THzoncethegainmediumisavailable).

BackgroundThe63‐μm(4.744THz)[OI]fine‐structurelineisthedominantcoolinglineofwarm,dense,neutralatomicgas.Becauseofitsgreatintensityinhigh‐UVphotodissociationregions(PDRs)andshocks,the [OI] 63‐μm line is superior for probing regions ofmassive star formation and the centers ofgalaxies.ItisauniqueprobeofPDRs,shockwavesfromstellarwinds/jets,supernovaexplosions,and cloud‐cloud collisions. These radiative and mechanical interactions shape the interstellarmediumofgalaxiesanddrivesgalacticevolution.Thesizescaleoftheinteractionscanexcite[OI]emission over many parsecs. Moreover, the emission regions are often complex, with multipleenergeticsourcesprocessingtheenvironment.Spectrallyresolvedobservationsofthe[OI]linewithaheterodynereceiverarraywillallowuserstodisentanglethisconvolutedinteractionandpermitthestudyoftheenergybalance,physicalconditions,morphology,anddynamicsoftheseextended

101

regions. In this way, such a receiver array will provide new and unique insights into theinterrelationshipofstarsandgasinawiderangeofgalacticandextragalacticenvironments.

ThisprojectmainlyaddressesNASA’sStrategicSubgoal3D,“Discovertheorigin,structure,evolution,anddestiny of theuniverse,and search forEarth‐likeplanets.” It also addresses NASA’s StrategicSubgoal 3A, “StudyplanetEarth from space toadvance scientificunderstandingandmeet societalneeds”;andNASA’sStrategicSubgoal3C,“Advancescientificknowledgeoftheoriginandhistoryofthesolarsystem,thepotentialforlifeelsewhere,andthehazardsandresourcespresentashumansexplorespace.”ThedevelopmentwillsignificantlyreducetheriskofseveralcurrentandfuturesuborbitalprojectssuchasGUSTO,along‐durationballoon(LDB)payload.Theproposedsystemsthisprojectwilldevelopincludeaneight‐elementheterodynereceiverarrayforthe4.744‐THz[OI]line.

DFBstructuresarerequiredtogeneraterobustsingle‐mode lasingoutput.Currently,1st‐order,2nd‐order,and3rd‐orderDFBlasershavebeendemonstrated(thegratingperiodofannth‐orderDFBgratingisntimesthehalf‐wavelengthinthegainmedium).Amongthese,the3rd‐orderDFBstructuresshowthe greatest promise for LO applications, because of their compact size and good beam patterns.Despitethepromiseof3rd‐orderDFBlasers,thephasemismatchlimitstotallengthtolessthan1mmsothebeamisstillquitedivergent.Inaddition,thelowextractionefficiencyyieldsalowWPEof0.1%.InthisSATproject,weplantofurtherdevelopthispromisingtechnologybydevelopingthefollowing.

• Perfectlyphase‐matched3rd‐orderDFB lasers, to generate evennarrowerbeampatternsandhigheroutputpowerlevels;

• Novelantenna‐coupled3rd‐orderDFBstructuretoincreaseWPE;and• Better‐performancegainmedium,peakedat4.74THz.

ObjectivesandMilestonesTheprojectobjectivesaretodevelopsingle‐modeDFBlaserswithfrequencywithin10GHzofthetarget4.744‐THzline,cwoutputpowerlevelgreaterthan5mW,WPE≥0.5%atanoperatingtemperatureof~40K,andbeampatternsnarrowerthan10 10degrees2.TheannualmilestonesarelistedinTable1.

Year Milestones

Year 1

(3/2016 – 2/2017)

• Complete the design of perfectly phase-matched 3rd-order DFB lasers aimed for ~4.7 THz• Develop a high-yield dry-etching process using inductive-coupled plasma (ICP) to

achieve clean and smooth sidewalls with high aspect ratios • Grow ~3 MBE wafers based on improved QCL designs • Fabricate devices using a combination of dry and wet etching

Year 2

(3/2017 – 2/2018)

• Continue to improve the fabrication process for higher quality and higher yield • Grow ~3 MBE wafers based on improved designs • Develop perfectly phase-matched (neff = 3.00 ± 0.02) 3rd-order DFB with a modest

value of m ≈ 2 cm-1 to ensure a robust single-mode operation and WPE ≈ 0.1% • Design phase-matched 3rd-order DFB structures integrated with half-wave antennas

Year 3

(3/2018 – 2/2019)

• Grow ~3 MBE wafers based on improved designs • Develop phase-matched 3rd-order DFB coupled with integrated antennae with a more

aggressive value of m ≈ 10 cm-1, achieving WPE ≈ 0.5% with ≥ 5 mW power at ~40 K. The phase matching should be good enough for a long device for the high output power and with beam divergence ≤ 10 10 degrees2

Table1.MilestonesofthisSATproject.

ProgressandAccomplishmentsPriortothebeginningdateofthecurrentSATproject,wesuccessfullydevelopedasingle‐mode3rd‐orderDFBlaserthatlaseswithin3GHzofthetarget4.744‐THz[OI]line[1].Althoughthepowerlevelislower

102

thandesiredduetoimperfectphasematchingandthelackofintegratedantennas(whichasdiscussedaboveispursuedintheSATproject),itisadequatetopumpasingle‐elementHEBmixer.OnedevicewasintegratedwiththeStratosphericTerahertzObservatory(STO‐2)LDBobservatory.Unfortunately,thecontrolelectronicsofthisdeviceweredamagedbythesunduringtheflight,sonodatawasobtained.

Thedesignofnovel antenna‐coupled,perfectlyphase‐matched3rd‐orderDFB structures requirescleveranalysisandaccuratenumericalvalidation,whichiswhatweaccomplishedduringthefirstyearoftheproject.Fabricationofthedesignedstructuresrequiresasophisticatedcombinationofdry and wet etching, which started in the past three months. The design and growth of high‐performanceTHzgainmediaisalsoahighlychallengingprocess.AidedwithsophisticatednumericalpackagesincludingSchrödingerandPoissonsolvers,wecompletedthedesignofquantum‐cascadestructuresandthreeMBEwafersweregrown.ThegrowthwascarriedoutbyDr.JohnRenoatSandiaNationalLaboratories.Sandiahasalltherequiredequipment,andDr.Renohasalongtrackrecordinthegrowthofrecord‐settingTHzgainmedia.

The design of high‐power and single‐modeDFB lasers requires innovative thinking and detailednumerical simulations, and both were accomplished following a long history of innovation andtechnicalknow‐howinthePI’sgroup.Wecompletedthedesignofantenna‐coupled3rd‐orderDFBstructuresandgeneratedfabricationmasksbasedonthedesign.Wealsodesignedquantum‐cascadegainstructureswiththecenterfrequencyaround4.7THz.ThreeMBEwafersweregrownbasedonthedesign,butthegrowthtookplaceduringtheendoftheMBEmachinecycle.Asaresult,thedopinglevelissignificantlyoffsotheperformancewaspoorinoperatingtemperatureandoutputpowerlevels.WewillregrowthesethreestructuresinAugust2018whentheMBEmachineisbackonlineandwillcarryoutthefabricationanddevicecharacterizationshortlyafterward.Wedonotanticipateanychangeofdirectionintheproject.Weareontracktoachieveourmilestonesasshownaboveandanticipateachievingourobjectives.

PathForwardAtthisstage,weexpectourremainingworkwillcloselyfollowtheannualmilestoneslistedabove.

Reference[1] J.L.Kloosterman,D.J.Hayton,Y.Ren,W.Kao, J.N.Hovenier, J.R.Gao,T.M.Klapwijk,Q.Hu,C.K.

Walker, and J.L. Reno, “Hot electron bolometer heterodyne receiverwith a 4.7‐THz quantumcascadelaserasalocaloscillator,”Appl.Phys.Lett.,102,011123(2013)

Foradditionalinformation,contactQingHu:qhu@mit.edu

103

AdvancingFocalPlaneTRLforLiteBIRDandotherNextGenerationCMBSpaceMissionsPrepared by: Adrian Lee (PI; UC Berkeley and Lawrence Berkeley National Laboratory); ShawnBeckman,JulianBorrill,andBenjaminWestbrook(UCBerkeley);KamArnold(UCSanDiego);NilsHalverson(UniversityofColorado,Boulder);HannesHubmayr(NIST);Chao‐LinKuoandKeithL.Thompson(StanfordUniversity);andAritokiSuzuki(LawrenceBerkeleyNationalLab)

SummaryThis three‐year technical development project started in October 2017. The project’s goal is toincreasetheTechnologyReadinessLevel(TRL)ofsuperconductingmm‐wavedetectorsandreadoutelectronicsforacosmicmicrowavebackground(CMB)spacemission,withafocusonapplicationtotheLiteBIRDmission.Thismissionenvisionsafocalplanewiththousandsofdetectorsspanning15spectral bands – two orders of magnitude more than the most recent CMB space mission. Thisproject’stechnologydevelopmentgoalsarenecessarytorealizethenext‐generationCMBfocalplane.

TheUSteamisworkingaspartoftheLiteBIRDinternationalcollaboration,ledbyMasashiHazumi.IntheUS,therearefivepartnerinstitutions:TheUniversityofCaliforniaBerkeley,TheUniversityofCalifornia San Diego (UCSD), The University of Colorado Boulder, Stanford University, and theNationalInstituteofStandardsandTechnology(NIST)inBoulder,CO.

BackgroundTheCosmicMicrowaveBackground(CMB)hasbeenagoldmineforcosmology.Thetemperaturefluctuations measured by the COBE, WMAP, and Planck space missions revolutionized ourunderstandingofcosmology,usheredintheeraofprecisioncosmology,andformthebasisformuchoftoday’sstandardcosmologicalmodel.TheCMBgivesussuchfundamentalinformationabouttheuniversesinceitisemittedatatimeveryclosetotheBigBang,0.003%ofthecurrentageofuniverse.

Anext‐generationCMBimagermission,suchasLiteBIRDorPICO,willdigintothenextarcheologicallayerbymeasuringthepolarizationanisotropyoftheCMB.Allpreviousmissionswereoptimizedfortemperaturemeasurements,but thepolarizationcontainsnewmessengers that informuson thebeginning and evolutionof theuniverse.At the very firstmomentof theuniverse, space itself isthoughttohaveundergoneanexponentialexpansion.ThisexoticbutsimpleideacalledInflationiscompellingsinceitexplainssimultaneouslybasicconundrumsabouttheuniversesuchas:whyistheentireskythesametemperatureeventhoughtheregionsweseewouldnotnaivelybethoughttobecausallyconnected?Whyisthegeometryofspaceflat(Euclidean)?

TheorysuggeststhatthereisonemeasurementthatpositivelydetectsInflationasthecorrectmodelforthebeginningoftheuniverse–toseetheswirlingfingerprintsof“B‐mode”polarizationinthe

Akeytechnologicalrequirementforthedetectorsiscontroloftheirsensitivitytocosmicrays.Wehavedemonstratedthatourdetectorarraydesignisfairlyinsensitivetoenergydepositionfrom ionizing particles in the bulk silicon, which is a positive step toward achieving therequirement on dead‐time due to cosmic‐ray interactions in flight.

104

CMB.Ifweareabletoseethissignature,wewouldnotonlyhaveaclearidentificationofInflation,butcouldalsomeasuretheenergyscaleofInflationandbeabletonarrowinontheprecisemodelfor Inflation.Evenifasignal isnotseen,wewilleliminateanentireclassof Inflationmodelsandtherebychangethedirectioninthesearchforthephysicsatworkatthebeginningoftheuniverse.

TheUSisparticipatingintwoimagermissionconcepts,LiteBIRDledbyJAXA,andPICO,aProbe‐classNASAmissionstudy.OurPCOSinvestigationfocusesonTRLadvancementwiththespecificLiteBIRDmissionconceptasafocus,butallthetechnologiesareleadingcandidatesforPICOandotherfutureCMBmissionconcepts.TheLiteBIRDmissionisoneoftwocandidatesinthe“StrategicLargeMission”categoryinJAXA(alongwithSolarSail),andadownselecttoonemissionisscheduledlatein2018.

The focal‐plane array technology required for LiteBIRD has been matured for ground‐basedobservationsbutrequiresadaptationstoachievetheTRLrequiredforaspacemission.Thelistofareasforadvancementinclude:

i. ThedetectorarrayresponsetocosmicraysatL2(thelikelyorbitforaCMBspacemission)needstobeunderstood.Ifthepredicteddeadtimeduetocosmicrayswouldsignificantlyreducescienceyield,thearraysneedtoberedesignedwithfeaturestomitigatetheeffectofcosmicrays.

ii. Thedetectorarraysmustbetestedtoseeiftheywouldsurvivelaunchvibration,andmodifiedifrequired.

iii. The transition‐edge sensors (TESs)must be adapted to the lower optical power theywouldreceiveinaspace‐bornetelescopewhichisstraightforwardengineering.

iv. Thefrequency‐domainreadoutelectronicsthatourteamhasdevelopedmustbeadaptedtothelongercablelengthsexpectedinaspacecraft.

v. Techniquestodevelopsimulationsoftherelativelylargedatasetgeneratedinanall‐skymissionwiththousandsofdetectorshavetobedeveloped.

Aswillbedescribedindetailbelow,ourinvestigationwilladdresseachoftheseareastoadvancetheTRLofthedetectorarrays,readoutelectronics,anddatasimulation.Detectorarrayperformancewillbeevaluatedusingfocusedtestingsuchaswithradioactivesourcestomimiccosmic‐rayeventsandshake‐tableteststosimulatelaunchvibration.Wewilliterateonthedesignofthedetectors,readoutelectronics,anddatasimulationpipelinetoprogressfromTRLinthe3‐4rangetoTRLinthe5‐6rangeoverthethree‐yearprogram.

ObjectivesandMilestonesTheobjectivesofthisresearcharetoincrease,inparallel,theTRLoffocal‐planesuperconductingdetectorsandreadoutelectronicsfortheJAXALiteBIRDCMBmissionandotherCMBspacemissions.ThemilestonesidentifiedintheproposaltodothisworkareshowninTable1below.Goodprogressisbeingmadetowardtheinitialmilestones.Detailsaregivenbelow.

Date Deliverable2018-09 All necessary testing facilities validated

2018-09 Initial designs for TRL-4 demonstration of technology all complete

2019-03 Produce supporting documents for JAXA System Requirements Review (SRR), including both technology requirements and instrument simulations based on those requirements

2019-09 Demonstrate TRL 4 of all focal-plane and cryogenic readout technology

2019-09 Demonstrate acceptable level of magnetic-field and vibration sensitivity

2020-03 Demonstrate cosmic-ray mitigation appropriate for detectors

2020-03 Demonstrate end-to-end simulations of instrumental systematic effects, at their demonstrated levels in observations

105

2020-09 Demonstrate TRL 5 or 6 of all focal-plane and cryogenic readout technology

Table1:Milestonesidentifiedinoriginalproposal.

2018‐09:Allnecessarytestingfacilitiesvalidated:o ThetestcryostatatUCSanDiegothatwillbeusedfordevelopmentoftheLiteBIRDcryogenic

readout components is cryogenically fully operational. Wiring to support the DfMuxdevelopmentforLiteBIRDhasbeendesignedandordered,andwillbeinstalledintimetomeetthismilestone;

o ThetestcryostatatUCBerkeleythatwillbeusedfordevelopmentofLiteBIRDdetectorshasbeenordered.ItwillbedeliveredinJanuary2019.Untilthattime,othercryostatsalreadyvalidatedatUCBerkeleyarebeingusedfortesting;and

o The test cryostatatStanfordUniversity is in thedesignphase.A100‐mK fridgehasbeenidentifiedandwill soonbeordered.Wehavedecided todelayacquisitionof thiscryostatsinceitwillnotbeneededuntillaterintheprogram.

2018‐09 Initial designs for TRL‐4 demonstration technology all complete. Key issues fortechnologydevelopmentdescribedintheproposalareaddressedinmoredetailbelow:o Detectorcosmic‐raysensitivity:wehavedemonstratedtheabilitytomeasureeventsfroma

radiationsource,andhavedesignedseveralmitigationstotest.Thesub‐taskoncosmic‐raysensitivityiscomplete.Demonstrationisunderway.

o Detectorvibrationsensitivity:vibrationtestsofexistinghardwaresimilartothatproposedfor LiteBIRD has been successfully completed. However, no specific TRL developmenthardwarehasbeenproduced.Weareawaitingagreementwithourinternationalpartnersonthefocal‐planemechanicallayouttobuildthesemodules.Workonthislayoutisunderway.

o Detectormagneticfieldsensitivity:constructingHelmholtzcoilsandrelatedtestequipmenttocharacterizedetectorresponsetoknownmagneticfields.Weplantoachievethemilestonein2018‐09withLiteBIRD‐likedetectors.

o Cryogenicdetectorreadoutsystem:thecryogenicreadoutsystemhasbeenre‐designedtoincludeasuperconductingquantuminterferencedevice(SQUID)arrayamplifieratthe0.1‐Kfocal‐planestage.Thisdesigniscomplete,andsometestinghasoccurred.Furthertestingisunderway. On the topic of cryogenic readout, this milestone is complete, and nowdemonstrationisunderway.

2019‐03: Produce supporting documents for the JAXA SRR, including both technologyrequirementsandinstrumentsimulationsbasedonthoserequirements.Whilethismilestoneis scheduled for 2019‐03, theUS team actually has to support a JAXA Phase A1 Exit Reviewaround2018‐09.Wearemakingprogresstowardthisgoalfirst.o Technology requirements:workingwith our international partners,we have detailed the

instrument requirements to level 3, and are currently detailing the US deliverablerequirementstolevel5.Thatrequirementsbreakdownwillbedeliveredby2018‐09.

o Instrumentsimulations:withfocal‐planedesignexpertise,wehavesupportedtradestudiesfortheinstrumentopticaldesignandupdatesoftheprojectedinstrumentsensitivity.Wewilldefineabaselinedesignbefore2018‐09,andsupportsimulationsofthatdesignincludingthesensitivityoftheinstrumentandtheeffectofinstrumentaluncertaintiesonthedata.

2020‐03:Demonstrateend‐to‐endsimulationsof instrumentalsystematiceffects,at theirdemonstratedlevelinobservations.o Simulation framework performance: we have ported the Time‐Ordered Astrophysics

Scalable Tools (TOAST) simulation and reduction framework to the new flagshipsupercomputerattheNationalEnergyResearchScientificComputingCenter(NERSC),andbeguntooptimizethisnovelarchitecture.

106

o Simulationofsystematiceffects:wehaveusedsimulationsofrealistic,correlateddetectornoise tooptimizeour scanstrategyparameters, including investigatingour robustness tocatastrophichalf‐waveplatefailure.

Othermilestonesintable1arediscussedinthe“PathForward”sectionattheendofthisdocument.

ProgressandAccomplishmentsIn this section,wereportonprogressandaccomplishmentsonCosmic‐RayDeadtimeMitigation,Focal‐PlaneDesign,BolometerDesignOptimization,andDataSimulation.

Cosmic‐RayDeadtimeMitigation

TheLiteBIRDmissionrequirementistolimitdatalossfromcosmic‐rayinteractionsinthefocalplaneto 5%. To achieve this requirement, our team is first developing an understanding of energypropagationmechanismsinourdetectorwafers.Ourcurrentunderstandingincludestwoprimaryenergy transportmechanisms.First, energydeposition in the siliconsubstrate generatesballisticphonons which penetrate the thin detector layers, producing both ballistic and diffusive lateralphononpropagationtotheTES.Second,Cooperpairsintheniobiumgroundplanearebrokeneitherdirectlybytheincidentparticle,orbydiffusephonons.Theresultantquasi‐particlesthenrecombinetocreatediffusephononsneartheTESthataresubsequentlyabsorbed.

Toverifythesemechanisms,atestfacilityandanalysistoolkitforcosmic‐raymitigationtestinghavebeendevelopedatUCBerkeleyoverthelast12months.ThistestsetupusesanAmericium‐241alpha‐particlesourcetobeaccuratelypositionedinthreedimensionsoveratestchipcooledto250mK.Thesourceprovidesa5.5‐MeValphaparticleaccompaniedbya60‐keVgammaraytosimulateenergydeposition into our focal plane via cosmic rays (see Fig. 1 for example alpha events). The alphaparticleiscapableofpenetratingthesiliconnitrideandniobiumsurfacelayerswhenincidentonthedetector‐sideofourdevicewafers,andwillstopinthesubstratewhenincidentonthesubstrate‐side.

Fig.1.TimestreamsofAmericium‐241alphaeventsonthedetectorsideofthesmalldetectorarraywithpulsefitsdisplayed.Top:Atimestreamofdataof22‐seconddurationshowingtherateofinteractionandpulseheight

107

variation.Bottom:Azoom‐inof1.5secondsofdatashowingtherapidriseandslowerfalltimesofthepulses.Thefalltimesareconsistentwiththethermaltime‐constantoftheTESdetectors.

Wecancompareeventsfromilluminationofthetwosidesofthedetectorchiptodistinguishbetweentransportmechanismsoriginatinginthesurfacelayersvs.inthesubstrate.Wefindthatthehitrategiven by alpha particles impacting the substrate‐side of our test chips is nearly identical to thebackgroundhitrate,settinganupperlimitondetectionofthebacksidealphainteractions.Thedataareconsistentwiththehypothesisthatballisticphononscreatedinthesubstrateby5.5‐MeValphaparticles do not transmit energy to the TES appreciably when compared with mechanismsoriginating in the detector‐side surface layers.When testingwith alphaparticles incident on thedetectorsideofourtestchips,wehaveverifiedthattheprimaryenergytransportmechanismshaveameanfreepathof1‐2mm.

In future testing,weplan tocontinue to investigateenergy transportmechanisms.Also,wehavedesignedandbuilttestdetectorsthathavefeaturestomitigateenergypropagationtotheTESwhichweplantotestinthecomingyear.Currenttestshavebeendoneat300mK,butwhentheBerkeley100‐mKtestfacilityisoperational,wecantestthebulkthermalresponseoftheentiredetectorarraytoenergydepositionbyionizingradiation.

Focal‐PlaneDesign

In collaboration with LiteBIRD teams from Japan and Europe, we have been optimizing telescopeconfigurationsandfocal‐planeconfigurationstomaximizesciencereturnforagivensetofinstrumentresources [2 ,3, 5].We optimize the focal‐plane configuration to achieve required sensitivity at allfrequency bands, meet optical coupling requirements, minimize the number of detectors to relaxreadoutrequirements,andminimizethenumberofdetectormodulestoreducemassandcost.Anexampleofafocal‐planedesignisshowninFig.2.Thisnewdesignhasfewerwaferscomparedtothe2016NASAconceptstudyreport.Thenewfocalplanehashighersensitivitybecauseithaslessfocal‐planespacetakenupbymountinghardware.Weareintheearlystageofthisre‐optimizationprocess,andtheinternationalcollaborationisstudyingvarioustrade‐offstoratifyanewbaselinedesign.

Fig.2.Left:Exampleofre‐optimizedfocalplanes.Right:Sensitivitycomparisonbetweenthecurrentbaselinedesignandthenewdesign.

Space‐OptimizedBolometers

A technical requirement of our Technical Development program is to realize TES bolometersoptimizedforspace.Wehavetakentwosignificantstepstowardrealizingsuchdetectors.First,wedevelopedAlMnfilmsoptimizedfor100‐mKoperatingtemperaturewithsuitableresistancetobecoupledtothebaselinefrequency‐divisionSQUIDmultiplexer.Second,wedesignedandaresoontofabricatebolometerprototypesthatusetheseAlMnfilmsandarespace‐optimized.

108

TosatisfytheL1sciencerequirementtoconstraininflationandtheresultingL2sensitivityrequirement,LiteBIRD will use arrays of TES bolometers operated at 100 mK, with a target superconductingtransitiontemperatureTc=170mK[4].Anadditionalconstrainttothesensordesigniscompatibilitywith the readoutmultiplexer. For the frequency‐divisionmultiplexer (fMUX) used in LiteBIRD, therequiredsensorimpedanceis~1Ohm.LiteBIRDisthefirstCMBexperimenttousethecombinationof100‐mKbolometersandR=1Ohmsensors,andthereforeTESfilmdevelopmentisrequired.

Byuseof90‐nm‐thick,220‐ppmAlMnfilmswehaverealizedfilmsthatcovertheneededparameterspace in Tc and R. We demonstrated high spatial uniformity in the superconducting transitiontemperature,Tc,acrossthewafer.ThisuniformityisessentialtoachieveconsistentsensitivityofthedetectorsforLiteBIRDandothernext‐generationCMBpolarizationsatellitemissions.Figure3showsameanTc~170mKandapeak‐to‐peakvariationthatis6%withina100‐mmdiameterwafer,theplannedwafersizeforLiteBIRD.ThislevelofuniformityenablesatightdistributionindetectorNEPacrossthearray.

Fig.3.AlMnfilmsdevelopedforLiteBIRDachievethetargetTcandexhibit6%peak‐to‐peakuniformityacrosstheplanned100‐mmdiameterwafer.

Second,UCBerkeleyandNISTco‐fabricatedadetectortestchiptomeetLiteBIRDspecifications.AphotographofthebolometersisshowninFig.4.Thebolometerdesignsspanaleg‐lengthrangeof250to2000µminordertocovertherangeofsaturationpowersrequiredforthemedium‐frequency(MF) pixels to be fabricated at Marvell Nanofabrication Facility. A second design iteration isunderway and will likely provide us with baseline designs to achieve the specifications of thebolometersforalltheMFwafers.

109

Fig.4.PhotographofaLiteBIRDtestchipfabricatedatUCBerkeleywithfabricationsupportfromNISTfortheAlMnTES.Resultsfrominitialtestshowthatthesebolometersarewitha10%oftheTctargetandafactorof2ofthenormalresistancetargets.

Finally,wedesignedspace‐optimizedbolometers.WeplantofabricatedetectorchipswithseveralvariationsonthebolometershowninFig5.Theaimofthefabricationsplitsistospanthesaturation‐power requirements of LiteBIRD and probe the thermal circuit of these devices for stability.Maximizing the instantaneous sensitivity of the detectors given the low photon‐loading of spacedictates the need for low‐thermal‐conductance bolometers. For LiteBIRD, the smallest saturationpower (the totalpowerwhich thebolometercansupport, aquantitydirectlyproportional to thethermalconductance)isapproximately100fW.Inpreviouswork,TESbolometersoptimizedforaspace‐based spectrometer, with parameters similar towhat is required for LiteBIRD, have beenrealizedonthesinglepixellevel[1].Morerecentlyandwithinourowncollaboration,TESbolometersoptimizedforballoon‐basedobservationusing300‐mKrefrigerationhasbeendemonstrated[2].ThebolometerdesignshowninFig.5buildsontheseresults;themaindifferencebeingoptimizationfor100mKasopposedto300‐mKrefrigeration.

Fig.5. Space‐optimizedTES bolometer design.The four long thermally isolating legs achieve low saturationpower,yetaremechanicallyrobust.Thesizeofthecentralbolometerislandis150µmx150µm.

Cryogenicreadoutdesign

Toachievetherequirementsonthedetectorreadout,theSQUIDarrayamplifierusedtoamplifythedetectorsignalsneedstobeoperatedatlowertemperatureandclosertothedetectors.Wedesigned

110

areadoutcircuitwiththeSQUIDarrayamplifiersatthedetectorfocalplane(Fig.6),andhavebeguninitialtestsat250mK.Testsat100mKwillcommencesoon.

Fig.6.ImageofcryogenicreadouttestcircuitwithSQUIDarrayamplifier,inductor/capacitorboard,and

bolometertestchips.

ThermalCalculations

In support of new collaborators in Europewhowill focus on the sub‐K cooler and high‐frequencytelescope(HFT)telescopedesign,Co‐IThompsonre‐calculatedthethermalcoolingpowerrequirementsof the sub‐4‐K stages for various focal‐plane options for the low‐ and high‐frequency LiteBIRDtelescopes.Manyof thesechoicesare interdependent,sothate.g.choiceof telescopewith its f‐ratioaffectsthephysicalsizeofthefocalplane,evenintheabsenceofapixel‐countchange,thuschangingthestructuralsupportthermalloadingaswellastheradiativeloadingonthecoldstages.Inthecomingyear,weplantoconvergeonmaturedesignswithspacethermalexpertsattheSSLatBerkeley.

DetectorTestSystems

AtUCSanDiego,ABlueForsdilutionrefrigeratorcryostatwasprocuredandiscryogenicallyfunctional.ThiswillbetheprimaryLiteBIRDcryogenicreadouttestfacility.Itisbeingoutfittedwithfullwiringnow.Inparallel,westartedsometestinginthiscryostatthatdoesnotrequirethefullwiring.

AtStanfordUniversity,Co‐IThompsonhasbeendesigningtheoptical‐testcryostatinpreparationfordetectortesting.Itwillhaveinternalsourcesformeasurementsofopticalbeams,opticalefficiency,integratedsidelobepower,andpolarization.Thetest facilitywillalsohavea light‐pipeoptionforspectral‐responsemeasurementsforbothanFTSsourceandafiltered(e.g.,usingthickgrillfilters)thermalsource.Helmholtzcoilsaroundthecryostatwillsupportmagnetic‐field‐sensitivitytestingupto~100Hz.A100‐mKfridgewillbeorderedsoon.Thetime‐scaleforcomingon‐lineshouldbebythebeginningof2019.

AtBerkeley,ABlu‐Forsdilutionrefrigeratorwasordered,withdeliveryexpectedinJanuary2019.AnewcryostatsectionwasdesignedwithsufficientsizetoholdtheentireLow‐FrequencyTelescope(LFT) or High‐Frequency Telescope (HFT) focal planes. This refrigerator will be the main testplatformforfullfocal‐planeintegrationtests.

Time‐DomainDataSimulations

111

LiteBIRD'sabilitytoachieveitsscientificgoalswilldependonprecisecontrolofbothstatisticalandsystematicuncertainties.Statisticaluncertaintiesrequireustogatherasufficientlylargedataset,butthisthenposesahostofcomputationalchallenges.Systematicuncertaintiesrequireusfirsttodesignthemissionsoastominimizethemattheoutset,andthentodesign,implement,validate,andverifystrategiestomitigatewhateverisleftinthedataanalysis.Generatingandreducingsyntheticrealisticdata–bothinvolumeandcomplexity–isthereforeamission‐criticalcapability.ForLiteBIRDweareextending the TOAST CMB data simulation and reduction software, originally developed anddeployedbythePlanckmission,tothescaleandrealismrequiredbyLiteBIRD.

AtBerkeley,Co‐IBorrill'steamhasportedTOASTtothenewestNERSCsupercomputer,Cori.BuiltontheXeonPhiKnightsLandingcores,eachCorinodesupportsupto272threads,requiringahighlevelofvectorizationtoachievegoodperformance.Wehavefullyinstrumentedthecode,identifiedanumberofperformancebottlenecks,andaddressedthesesufficientlyforthecodetoscaletothefullsystem's600,000+cores.

AtStanfordUniversity,newpostdoctoralhireGiuseppePuglisihasbeendoingsimulationson1/fnoise contributions to LiteBIRD‐like scanning CMB satellites, with and without polarizationmodulators, and for ranges of satellite spin rates and 1/f knees. In coordination with othercollaboratorsintheLiteBIRDproject,hehasidentifiedsomekeyissuesthatneedtobeaddressed,such as the distribution of pixel orientations in the focal plane. He is also beginning to addresssystematicsquestionswithregardtogainvariation.

PathForwardAsdescribedintheObjectivesandMilestonessection,sofar,weareontracktomeetourearlyprojectmilestonesanddeliverablesshowninTable1.Ourpathforwardfortheremainingworkcontinuestoadheretotheplandescribedintheproposal.Themajorremainingwork,byprojectmilestone,isoutlinedbelow,withnotesoncurrentprogresswhereapplicable.

Allnecessarytestingfacilitiesvalidated(due2018‐09)o FabricateandcommissionUCSDtestcryostat

Deliveredandcryogenicallyoperational.Orderedreadoutwiring.o FabricateandcommissionUCBerkeleytestcryostat

Ordered;deliveryexpectedJanuary2019o FabricateandcommissionStanfordtestcryostat

Designcompletedandconstructionbeginning InitialdesignsforTRL‐4demonstrationtechnologyallcomplete(due2018‐09)

o Designdetectorcosmic‐ray‐sensitivitytestprogram Complete;demonstrationisunderway.

o Designdetectorvibration‐sensitivitytestprogram Awaitingagreementwithourinternationalpartnersonthefocal‐planemechanicallayout

o Designdetectormagnetic‐field‐sensitivitytestprogram Constructing Helmholtz coils and related test equipment to characterize detector

responsetoknownmagneticfields;testsplannedforlaterin2018o Redesigncryogenicdetectorreadoutsystemanddevisetestprogram

Complete,demonstrationisunderway ProducesupportingdocumentsforJAXASRR(due2019‐03)

o Finalizetechnologyrequirements Completeddraftinstrumentrequirementstolevel3;currentlydetailingtheUSdeliverable

requirementstolevel5;workingwithinternationalpartnerstofinalizerequirementso Developinstrumentsimulationstovalidaterequirements

112

Supportingtradestudiesfortheinstrumentopticaldesignandupdatesoftheprojectedinstrumentsensitivity;afterbaselinedesignisfinalized,willsupportsimulationsofthatdesign including the sensitivity of the instrument and the effect of instrumentaluncertaintiesonthedata

o Developinterfacecontroldocuments(ICDs) DraftingICDmatrix;whenbaselinedesignisfinalized,willbegindraftICDdocumentsfor

theUSdeliverableso Developvalidationplan

Workingwithinternationalpartners,developvalidationplanforUSdeliverables DemonstrateTRL4ofallfocal‐planeandcryogenicreadouttechnology(due2019‐09)

o Design/fabricate/testsingle‐pixelLF/MF/HFdetectorsandopticalcouplingo Design/fabricate/testlow‐fidelityLF/MF/HFdetectormoduleso Design/fabricate/testlow‐fidelityfocal‐plane‐towermechanicalassemblyo CompleteTRL‐4demonstrationonredesignedcoldreadout

Demonstrateacceptablelevelofmagneticfieldandvibrationsensitivity(due2019‐09)o WorkwithJapaneseandEuropeanpartnerstodefinelaunchvehicle/spacecraft/instrument

magnetic‐fieldandvibrationalenvironmento Iterateondetector/module/focal‐planedesignandshieldingtoreducemagnetic‐fieldand

vibrationsensitivitytomeetrequirementso Iterateoncold‐readoutdesignandshieldingtoreducemagnetic‐fieldandvibrationsensitivity

tomeetrequirements Demonstratecosmic‐raymitigationappropriatefordetectors(due2020‐03)

o WorkwithJapaneseandEuropeanpartnerstodefineanticipatedcosmicrayfluxatL2andfinalizerequirements

o Iterateondetectordesign/fabrication/testusingin‐labradioactivesourceo IrradiatematuredesignwithprotonbeamatHIMACfacilityinJapan

Demonstrateend‐to‐endsimulationsofinstrumentalsystematiceffects(due2020‐03)o WorkwithJapaneseandEuropeanpartnerstoconductsimulationstofinalizescanstrategy,

useofhalf‐waveplate(HWP)andinformhardwarerequirementsforUSdeliverableso Finalizehardwarerequirementso Characterizedetectorandreadoutopticalandelectricalcontributionstosystematicso Incorporatemeasurementsintohigh‐fidelitysimulationsincludingdetector,telescope,and

HWPimperfections,representativeskymodel,andmissionoptimization DemonstrateTRL5or6forallfocal‐planeandcryogenicreadouttechnology(due2020‐09)

o Design/fabricate/testLF/MF/HFfulldetectorarrays,opticalcoupling,anddetectormoduleso Conductmagnetic‐field/vibration/cosmic‐raytestsonfullmoduleso Design/fabricate/testhighfidelityfocal‐plane‐towermechanicalassemblyo Integrate,test,anddemonstratedetectormodule,coldreadout,andfocal‐plane‐towerassembly

References[1] A.D.Beyeretal.,“Ultra‐sensitivetransition‐edgesensorsfortheBackgroundLimitedInfrared/Sub‐

mmSpectrograph(BLISS),”JournalofLowTemperaturePhysics,167.3‐4,182‐187(2012)[2] A. Suzuki et al., “The LiteBIRD Satellite Mission: Sub‐Kelvin Instrument,” Journal of Low

TemperaturePhysics,inpress(2018)[3] T. Matsumura et al., “LiteBIRD: Mission Overview and Focal Plane Layout,” Journal of Low

TemperaturePhysics,Vol.184,Issue3‐4,pp824‐831(2016)[4] C. Hill et al., “BoloCalc: a sensitivity calculator for the design of Simons Observatory,”

arXiv:1806.04316

113

[5] B. Westbrook et al., “Development of the Next Generation of Multichroic Antenna‐CoupledTransitionEdgeSensorDetectorsforCMBPolarimetry,” JournalofLowTemperaturePhysics,Vol.184,Issue1‐2,pp74‐81(2016)

Foradditionalinformation,contactAdrianT.Lee:Adrian.Lee@berkeley.edu

114

DevelopmentofaRobust,EfficientProcesstoProduceScalable,SuperconductingKilopixelFar‐IRDetectorArraysPreparedby:JohannesStaguhn(JHU)

SummaryThefundingforthistwo‐yearprojectstartedJanuary1,2018.Theprimaryobjectiveoftheprojectistodemonstrateaversatileschemetoconnecttwo‐dimensionalsuperconductingdetectorsandcold‐readoutelectronicsthroughaseparatesiliconstructurewithsuperconductingthrough‐wafervias.The team consists of a number of experienced scientists and technicians: S.HarveyMoseley, AriBrown, EdWollack, Karwan Rostem, Elmer Sharp and Steve Maher (NASA/GSFC); Gene Holton(NIST‐Boulder);andKentIrwin(StanfordUniversity).

Background

Thefar‐IRbandisuniquelysuitedtostudythephysicalconditionsintheinterstellarmediumfromnearby sources out to the highest redshifts. Incoherent superconducting bolometers are a high‐sensitivity, Far‐IR imaging and spectroscopy technology for many future suborbital and spacemissions,includingtheOriginsSpaceTelescope(OST).Suchmissionsrequirerobust,high‐sensitivitydetectorarrayswithseveral104pixels,high focal‐plane filling factors, and lowcosmic‐raycross‐sections,operatingovertheentirefar‐IRregime.

Thesearrayscouldconsistofsmallersub‐arrays, ifthosearetileable.Thetransition‐edge‐sensor‐(TES)basedBackshortUnderGrid(BUG)arrayarchitecture,whichourgrouphasfieldedinanumberoffar‐IRcameras,isagoodcandidatetomeettheserequirements:BUGsaretileable;andwiththeintegrationoftheSQUIDmultiplexer,scalablebeyondwafersizes;theyprovidehighfillingfactors(>90%at1‐mmpixelpitch),andaredesignedtohavelowcosmic‐raycrosssections.IndividualBUGarrays have been demonstrated in the HAWC+ camera on SOFIA to provide background‐limitedperformance for suborbital far‐IRapplications.However, theproductionofBUGswith integratedreadoutmultiplexershasmanytime‐andresource‐consumingprocesssteps.Inordertomeettherequirementofrobustnessandproductionefficiencyforfuturearrays,wedevelopedanewmethodtoprovidethesuperconductingconnectionofBUGdetectorstothereadoutmultiplexersorgeneralreadoutboardsbehindthedetectors.Thisapproachshouldallowustoproducereliable,verylargedetectorarraysforfuturefar‐IRmissions.

ObjectivesandMilestonesTheaimofthisprojectistodevelopandstreamlinethefabricationprocessesrequiredtoproducebackground‐limitedlargefar‐IRarrayswithlargepixelnumbers(~105).Wewillachievethisgoalby

Wehavedesignedallcomponentsneededtotestthemostcrucialtechnicalparametersrequiredfortheproposeddeliveryofacompletedetectorarray.Productionoftestchipsformetalinterfacesandtest detector arrays has begun.

115

combiningmaturedetectorandreadouttechnologiesfromourpreviouswork,tofabricatearobust,close‐packed,high‐sensitivitybolometerarraywithreliablehigh‐quantum‐efficiencyabsorbersthatoperatesovertheentire far‐IRrange,andcanbeefficientlyandreliablyproduced.Thesimplifiedprocesswillintegratedetectorarraysthroughsuperconductingbondstoacoldreadoutmultiplexer.Itisveryversatileinitsapplications,sinceitwillallowthematingofTESdetectorstotime‐domain,frequency‐domain,microwave,andcode‐divisionmultiplexers.

Themainobjectiveswillbeachievedbymeetingthefollowinggoals:

a. DevelopanovelBUGarchitectureinwhichthesuperconductingthrough‐viaprocessisseparatedfromdetectorproduction,improvingproductionspeedandreducingrisk;and

b. Producebackground‐limited5‐kilopixelarrayssuitableforthefar‐IRspectrometerSuper‐HIRMES.

Forthelatterwewilladditionally:

c. RefineourAlMnprocess for quickly and reliably fabricatingTESwithhighlypredictable anduniformtransitiontemperatures(lessthan5%variation)acrosstheentirewafer;and

d. Refineastandardprocessforreliablyfabricatingimpedance‐matchedandrobustabsorbersfortheentirefar‐IRwavelengthrange,whicharenotsusceptibletoroom‐temperatureaging.

ProgressandAccomplishments1.Through‐wafervias:

The existing through‐wafer via process involves placing the vias directly on the detector devicewafers,whichisextremelychallengingtechnically.Thisisprimarilybecausetheviasneedtobefilledwithapolymerthat isplanarwith thesiliconmembranes towithinamicron,sothatsubsequentmicron‐scalelithographyisnotcompromised.Inordertocircumventthischallenge,wehavemovedtheviastoaseparatesubstrate(Fig.1),whichrelaxesthetolerancesonthelithographyandgreatlysimplifiesthedetectorfabricationprocess.Furthermore,inordertoincreasethethrough‐waferviafabricationthroughput,wehavenotfilledthevias.

Fig.1.Schematicofthethrough‐waferviasubstrate.Thedashedrectanglesareregionswherethesiliconwillberemovedsothatlightcanpassthrough,whichwillbelocateddirectlybelowtheopticallyactiveregionofthebolometricdetectors.

The primary technical challenge associated with this new approach involves patterningsuperconductingtracesandindiumbumpsonthethrough‐waferviasubstrate.Wehaveaddressedthischallengebydevelopingtwonovelpatterningprocesses.

Thefirstprocess,usedtoetchmetallictracesonavia‐containingsubstrate, involvesusingatwo‐layeretchmask.Thefirstlayerconsistsofananti‐reflectioncoatingandthesecondconsistsofadry‐filmresist,whichislaminatedonthewafer.Thedry‐filmresistisphotolithographicallypatternedandtheanti‐reflectioncoatingisreactive‐ionetched.Incontrastwithaconventionalspin‐onresist,adry‐filmresistcaneffectivelycoverthevias.

116

Thesecondprocess,usedtoliftofftheindiumbumpsonavia‐containingsubstrate,involvesusingatwo‐layerlift‐offmask.Thefirstlayerconsistsofpositivephotoresistandthesecondconsistsofadry‐filmresist;andthelift‐offmaskispatternedinasimilarmannerastheetchmaskdescribedabove.

Figure2showsthesmallestfeatures,20µm,whichcanbereliablypatternedusingthefirstprocess.Ifwe conservatively allow for a 20‐µmbuffer region around the patterned features, in this casesuperconductingTiNleads,ournovelprocesswillpermitfabricationofbolometricdetectorarrayswithaBUGarchitectureandafillingfractionof~94%.

Fig.2.Micrographofadry‐film‐resistetchmask.

Figure3showsthesmallest indiumbumps,20µm,whichcanbereliablybe liftedoffasubstratecontainingthrough‐wafervias,oursecondprocess.Asinthecaseofallpatternedfeaturesfabricatedonthethrough‐waferviasubstrate,thesizeofthebumpswilllimittheoptical‐fillingfractionofthefocalplanebolometricdetectorarrays.Bumpplacementµµcanbeascloseas5µmawayfromthevias,illustratingtheversatilityofthisprocess.

Fig.3.Micrographoftwogroupsoffour4‐µmtallindiumbumpsnexttotwothrough‐wafervias.Thesiliconwaferinwhichtheviasweremicromachinedwas~0.4mmthick.

Wehavebegunproductionoftheseteststructures.Wehavealsodevelopedanovelmetal‐patterningprocess(Fig.4).Thisprocessinvolvesusingacombinationoflaminateddry‐filmresistandsprayed‐onpositivephotoresisttoliftofftracesfromawaferwhichcontainsthrough‐wafervias.Thisavoidstheneedof filling the vias,which is a high‐risk process. It also avoids theneed to use corrosivechemicalsforpatterningthemetal.

117

Fig.4.Lift‐offmaskusedtopatternmetalonthrough‐waferviacontainingsubstrates.

2.Metal‐metaltestchips:

Wehavedesignedtestboardstotestallrequiredmetal‐metalinterfaces(e.g.,In,Mo2N,andTiN)andmetal interfaces (e.g., Mo2N/In, TiN/In) for superconducting transition temperature and criticalcurrentdensity.TheboardsalsoenabletestsofInbumpswithfourdifferentdimensions.Thereare12unit‐cellsonasinglewafer.

Fig.5.Masklayoutforasingle“unitcell”testdevice.Thedieontheleftcanbeflip‐chipbondedontothedieontheright.

3.DetectorArray

Wehave designed our first TES detector pixelswhichwill provide uswith the capability to testthermalconductancethoughthedetectorlegsthatprovidetherequiredsensitivity.Thetestchipalsocontainsfeaturesthatletusmeasurethethermalizationwithinthemembrane.Figure6showsanimageofthetestarray,whichwillgointoproductionsoon.WedecidedtouseeitherMolly‐SiliciteorNiobium‐Silicyte.Patterningwillbeusedtotunefortherightimpedance.

118

Fig.6.Layoutofthetestarray

4.Detectorlandingwafer

WhileweonlyproposedtomatethenewdetectortoaSQUIDmultiplexer,wearealsoinvestigatinga more flexible solution to mate the detector to a wafer that contains a fanout to any type ofmultiplexer.Figure6showsasketchofthearchitectureusedtoconnectadetectorarraywiththe(multiplexer‐)readoutchips.Asshown,thewafercanbefoldedsothatthechipswillbebehindthedetector,allowingfortilability.Figure7showshowthisschemewilllookphysically.

Fig.6.Withtheabilitytoetchawayasectionofthewaferbelowthepolyamide,partofthewafercanbeflippedundertheotherpart,positioningitunderthearray.

Fig.7.CADimageofthewaferwithdetectorarray(right)andmultiplexerchips(left).Allconnectionsaremadewithsuperconductingmaterials.

PathForwardIntheremainderofYear1,weplanto:

Develop a through‐wafer via substrate fabricationprocess, refine double‐bondhybridization;preparefordevicetests;and

Carryoutafirstiterationofdetectorfabricationandtests/characterizationanddeliveruWaveSQUIDmultiplexers–completiontimeframe:12/2018

InYear2,wewillstartproductionofthefinalflightarray.Laboratorytestmeasurementswillbeusedtoverifyitsperformance.

119

Foradditionalinformation,contactJohannesStaguhn:johannes.staguhn@nasa.gov

120

High‐EfficiencyContinuousCoolingforCryogenicInstrumentsandSub‐KelvinDetectorsPreparedby:JamesTuttle(NASA/GSFC)

SummaryThis is a three‐year development effort,whichbegan in January 2017. Its goal is to advance theTechnologyReadinessLevel(TRL)ofamulti‐stagemagneticcoolingsystem,whichwillcontinuouslycooldetectors to0.05Kand rejectheat to10K.Thedevice, basedonadiabaticdemagnetizationrefrigerators(ADRs),willhavehighthermodynamicefficiencyandwillsimultaneouslycoolopticscontinuouslyat2–4K.Itwillexceedthecoolingrequirementsofallcurrentlyconceivedfuturespacemissions with cryogenic detector arrays. With a heat‐rejection temperature of 10 K, it will becompatible with recently demonstrated extremely‐low‐vibrationmechanical coolers. Since ADRshavenomovingparts,itwillenableanessentiallyvibration‐free300‐to‐0.05‐Kcoolingsystem.

Thedevelopment team is a groupof technologists atNASA/GSFCwith awealthof experience inspaceflight and laboratoryADRs. In fact, this groupdeveloped anddemonstrated the technologybeingadvancedhere.Duringthefirst18monthsofthiseffort,theyproducedadesigned‐for‐flightsingle‐stage10‐to‐4‐KADRandcompletedextensiveperformancetestingonit.Inaddition,theymadegoodprogressonthedesign,fabrication,andprocurementofparts,equipment,andsuppliestomakeasecond such stage. This was in preparation for assembly and testing of a two‐stage 10‐to‐4‐KcontinuousADRinlate2018.

BackgroundSeveralpast,andmanyfuture,astronomicalinstrumentsrequirecoolingtosub‐Kelvintemperaturesto obtain high sensitivity. Newer generations of detector arrays need several µW of cooling attemperaturesof0.05Kandlower.Previouslong‐lifespacemissionsusedsingle‐cyclecoolingdeviceswithmodestcoolingpower,butthesesub‐Kelvinrefrigeratorsarenolongeradequate.Asdetectortechnologymaturesandarraysizegrowslarger,demandhasincreasedforhighercoolingpowerandlower operating temperatures. Presently, several astrophysics flagshipmission concepts requirecoolingoflargesuperconductor‐basedfocalplanestosub‐Kelvintemperatures,includingtheOriginsSpaceTelescope(OST),Lynx,ProbeofInflationinCosmicOrigins(PICO),andGalaxyEvolutionProbe(GEP).Instrumentdesignstudiesforallthesemissions,includingthreedifferentOSTinstruments,arecurrentlyincorporatingourADRtechnologyintheirdesigns.Inaddition,thedevelopmentteamfor the Japanesemission,LiteBIRD,has expressed interest in this technology. Sub‐Kelvinenergy‐resolvingdetectorswouldalsoenhancetheHabitableExoplanetObservatory(HabEx)andLargeUV/Optical/Infrared (LUVOIR) mission concepts. High‐cooling‐power, high‐efficiency, high‐duty‐cyclesub‐Kelvincoolersarerequiredforthenextgenerationofsensitiveinstruments.BothCosmicOrigins(COR)andPhysicsoftheCosmos(PCOS)ProgramAnnualTechnologyReports(PATRs)listedsub‐Kelvincoolingastechnologygaps(“HighPerformanceSub‐KelvinCoolers”and“High‐efficiencycoolingsystemscoveringtherangefrom20Ktounder1K,”respectively).

Sub‐Kelvintemperaturesinspaceareproducedbyacombinationofmechanicalcryocoolersattheupper‐temperatureendandspecializedsub‐KelvincoolersoperatingfromafewKtolessthan0.05K.There are several ways to produce this low temperature: ADRs, dilution refrigerators, and acombinationofsorptioncoolingandadiabaticdemagnetization.AnADR,havingathermodynamic

121

efficiencyclosetoCarnot,isthemostefficientwaytoproducesub‐Kelvintemperatures.Itprovidescycliccoolingbyraisingthemagneticfieldinaparamagneticmaterial,removingtheresultingheatfromthematerial,andthencoolingitbyreducingthefield.Ourteaminventedamethodtoproducehigh‐heat‐liftcontinuouscoolingwithamulti‐stagecontinuousADR(CADR).WedemonstratedaTRL‐4CADRwith6.5µWofcontinuouscoolingat50mK,asignificantimprovementoverthe0.7µWcoolingprovidedforabout40hourspercycleonAstro‐H/SoftX‐raySpectrometer(SXS).ThatCADRcanalternativelyprovide31µWat100mKcomparedtothe0.2µWofcoolingachievedonPlanck.TheCADRrejectsitsheattoacryocoolerattemperaturesashighas4.5K.

Inadditiontodetectorcooling,severalproposedfar‐IRspaceobservatoriesrequiretelescopestobeat2–6Ktolimitself‐emission.Thisismuchcolderthanthe~30Kthatcanbeachievedviapassivecooling,asintheJamesWebbSpaceTelescope(JWST).ThistechnologyneedislistedintheCORPATR(“AdvancedCryocoolers”).Thecombinationofahighlyreliablemechanicalcryocoolerreaching10KandaCADRcoolingfrom10to4Kwillprovidealow‐input‐powersolutiontofillthistechnologygap.

Manyspaceobservatoriesrequireextremelystablepointing.TheCADR,whichhasnomovingparts,is vibration‐free,but it is justpartof a cooling chain.Currently, theCADRrejects itsheat at4K,limitingthechoiceofupperstagecoolerstolinearpistoncryocoolers.Jittercausedbysuchcoolershas been a problem for recent astrophysics missions. Miniature turbo‐Brayton coolers offer asolutiontothisproblem,aswasdemonstratedbythe70‐KcoolerfortheHubbleSpaceTelescope(HST) Near Infrared Camera and Multi‐Object Spectrometer (NICMOS) instrument. There aredifficulties extending this technology to 4 K, but an engineering unit has recently demonstratedsignificantcoolingpowerat10K.Thus,extendingtheCADRheatrejectiontemperaturefrom4to10Kwillenablesub‐Kelvindetectorsinobservatorieswithtightpointingrequirements.

Figure1showsourfullCADRschematically.Theleftsideoftheimageisthefour‐stage4‐to‐0.05‐KCADR.Thestagesarethermallyconnectedinseriesviaheatswitches.Thefirststage,ontheveryleft,remainscontinuouslyat0.05K.Theotherstagetemperaturesriseandfallastheheatswitchesopenandcloseinasequencethatpumpsheatfromthecoldsourcetothecontinuous4‐Kstage.The10‐to‐4‐KCADRisontherightside.Itconsistsoftwoparallelstages,whichalternativelycoolthe4‐Kstageandwarmuptojustabove10Ktodumpheattothecryocooler.Eachmagnetinthesystemhasitsownferromagneticshieldtominimizeitsstraymagneticfieldandtoenhancethefieldinsidethemagnetbore.Theentiresystemissurroundedbyasecondshield,whichfurtherreducesanyfieldfluctuationscausedbythemagnets.Figure2(left)showsapreliminaryconceptualdesignwiththeoverallshieldcutawayforclarity,and(right)across‐sectionofourdesigned‐for‐flight10‐to‐4‐KADRstage.The10‐Kmagnetisshowninsideitsferromagneticshield,withthesaltpillsuspendedinsidethemagnetbore.

Fig.1.Aschematicrepresentationofour10‐to‐0.05‐KCADR.

122

Fig.2.Left:ApreliminarylayoutofthefullCADRwiththeoverallmagneticshieldandthermalstrapscutawayforclarity(OD,outerdiameter).Right:Across‐sectionviewofourdesigned‐for‐flight10‐to‐4‐KADRstage.The10‐Kmagnetisshowninsideitsferromagneticshield,withthesaltpillsuspendedinsidethemagnetbore.

Ourdevelopmentplanistodemonstrateasingle‐stage10‐to‐4‐KADRinYear1,atwo‐stage10‐to‐4‐KCADRinYear2,anda10‐to‐0.05‐KCADRinYear3.Inaddition,wewilldevelopanoverallmagnetic‐shielding scheme which nearly eliminates any nearby field fluctuations caused by our CADRoperation.ThecompletesystemwillbetestedbeforeandaftervibrationtodemonstrateTRL6.

ObjectivesandMilestonesOurfirstmajormilestonewasmetearlyinYear2,whenwedemonstratedasingle‐stage10‐to‐4‐KADR.Thisdeviceincludesaunique10‐Kmagnet,developedseveralyearsagounderNASA’sSmallBusiness Innovative Research (SBIR) program. It also includes a passive gas‐gap heat switchconfiguredtoworkfrom10to4K,aferromagneticshieldforthemagnet,aparamagneticsaltpillinsideametalcan,andathermallyisolatingsuspensionsystemforthesaltpill.

BytheendofYear2,wewillmeetoursecondmajormilestonebydemonstratingcontinuouscoolingat4Kusingatwo‐stage10‐to‐4‐KCADR.Thiswillincludetwoshielded10‐Kmagnetswithsaltpillssuspendedinthem,twopassiveandtwoactivegas‐gapheatswitches,andthermalstrapsconnectingthepillsandswitches.Itwillrequireacustomizedcontrolalgorithmtokeepthe4‐Ktemperatureconstanttowithin0.001K.Thetargetcoolingat4Kwillbegreaterthan20mW.

BytheendofYear3,wewillassembleaflight‐worthyversionofourlaboratory4‐to‐0.05‐KCADRandintegrateitwiththe10‐to‐4‐KCADR.Theresulting10‐to‐0.05‐KCADRwillbesurroundedbyanoverallmagneticshieldtokeepexternalfieldfluctuationsbelow5µT.OurthirdmajormilestonewillbemetinthethirdquarterofYear3,whenthisassemblywillbeperformance‐tested,withatargetofatleast5‐µWofcoolingat0.05Kandbetterthan1‐µKroot‐mean‐square(rms)temperaturestabilityatthattemperature.Wewill thensubject it toavibrationtestchosentoenvelopetheacceptancelevelsexpectedforfuturecryogenicspacemissions.Thefinalmajormilestonewillbeasuccessfulpost‐vibrationperformancetestbytheendofYear3.

ProgressandAccomplishmentsDuring theperiodbetweenmid‐2017andmid‐2018, the teamaccomplished theYear‐1goal,andmadesignificantprogressonitemsrequiredtostayonscheduleinYears2and3.

4‐stage 4K ‐ 0.05K CADR O.D. ~ 35 cm

2 or 3‐stage 10K ‐ 4K CADR

123

Wefabricatedpartsandassembledadesigned‐for‐flightsingle‐stage10‐to‐4‐KADR.Figure3(left)showscomponentsofthesaltpillassembly,and(right)theKevlarsuspensionsystemusedtoholdthispillassemblyinsidethe10‐Kmagnet.WeinstalledtheADRinacryostatandperformedextensivetestingonitinearly2018.Figure4showstheADRbeforeandafterinstallationinthecryostat.Ineachtest,weoperatedtheADRthroughacompletecoolingcycle.First,wefullychargedthemagnetandwaited until the salt pill cooled back to 10 K. Then,we rapidly lowered themagnetic field,bringing thesaltpill to itsoperating temperatureof4K.At thispointweestablisheda feedbackcontrol loopwhich slowly rampeddown themagnetic field to keep thepill at 4K until the fieldreachedzero.Atthispointthe“holdtime”at4Kwascompleted,andwerampedthemagnetbackuptoitsfullfield,beginninganewcycle.

Fig.3.Left:Thecopperhousingwhichcontainsthesaltpillinthe10‐to‐4‐KADR,andthestainless‐steelshellwhichsurroundsit.Right:AmodularKevlar‐suspensionelementusedtosupportthesaltpillinsidethe10‐Kmagnet.

Fig.4.Left:Thesingle‐stage10‐to‐4‐KADR.Right:TheADRinstalledinacryostatfortesting.

Thiswasperformedseveraltimeswithdifferentconstant‐powervaluesapplied(whileat4K)toaresistiveheatermountedonthesaltpill.Theparasiticheatloadleakingintothesaltpillfromthe10‐Kstageexactlymatchedthecalculatedcontributionfromtheopenheatswitchandthepill’sisolatingsuspensionsystem.Also,thepillwasabletoabsorb2.4Joulesofheatat4K,independentofappliedpower valueswithin the range relevant for systemperformance. This indicates that our salt pillassemblyremainsisothermalevenwhenabsorbingthemaximumpowerthatitwillsee.

124

Wethenstudiedthespeedwithwhichwecouldre‐cycletheADRafter itsmagnetic fieldreachedzero.Weusedthemaximumpracticalmagnetramprate,andwewereableshortenthecycletimebynotwaitinguntilthesaltpillhadcooledto10K.The4‐Kcoolingenergydependsonthetemperatureat whichwe began themagnet’s rapid de‐magnetization. Thus, our tests provided a correlationbetween4‐Kcoolingenergyandtotalrecycletime.Inoureventualtwo‐stagecontinuous10‐to‐4‐KADR, two identical stages will operate in parallel. While one stage holds the constant 4‐Ktemperature,theotherstagewillberecycled.Thus,the4‐Kholdtimewillbeequaltotherecycletime.Thetotal4‐Kcoolingenergydividedbytheholdtimeisthemaximumheaterpowerthatourtwo‐stageADRcanabsorbcontinuously,andcalculationsusingourcycledatapredictedapowerofabout13mW.Finally,wedemonstratedthatwecouldcyclein132secondsandthenholdthepillat4Kfor132secondswhileabsorbing13mW.

Thetargetcoolingpowerforthe10‐to‐4‐KCADRis20mW,ofwhichlessthan1mWwillbeabsorbedfromthe4‐to‐0.05‐KCADRsubsystem.Ourtestingindicatesthatourexistingdesignwillfallshortofthatgoal,andweperformedextensivestudiestodeterminethecauses.Onecontributoristheheatcapacityofthemetallicpartswhichcyclebetween4and10Kalongwiththesaltpill.Ofthese,thestainless‐steelfastenersandthethickstainless‐steelflangeontheheatswitch’scoldendaremostsignificant.Weexpecttobeabletoincreaseourcoolingcapabilityby1–2mWbythinningthisflangeandswitchingtotitaniumfasteners.

We also discovered a subtle error in our performance prediction based on the publishedmagnetizationdataforthegadoliniumgalliumgarnet(GGG)salt.Itturnsoutthatforacylindricalpillinasolenoidmagnet,themagnetizationisnotuniform,andthereisaneffectivede‐ratingofthesalt’scoolingperformance.Weupdatedourmodeltoaccountforthis,anditnowpredictsourmeasuredperformancequitewell.Welearnedthatotherresearchersinthisfieldhavemadethesameerror,andweplantopublishourresultswithanexplanationofthiseffect.Wealsostartedaparallelstudyof the feasibilityofmakingsaltpills fromadifferent salt, gadolinium floride,withhighercoolingcapabilityperunitvolume.Thismaterialisonlyavailabletousasapowder,however,sowewouldneedtocompressitintosmallpillstobestackedinsideourcoppersaltpillbus.Dependingonthethermalconductivityofthecompressedpowder,wemightneedtoincludesomeadditionalcopperpartsbetweenthepillstoenhancetheheatconduction.Summerstudentsareassistinguswiththisstudy,anditisindependentofourbaselineddevelopmenteffort.

Our team members performance‐tested a 4‐to‐0.05‐K CADR built for an external customer forlaboratoryuse.ThissystemusedmanycomponentsidenticaltothoseplannedforuseinouroverallCADR,sothetestingwasveryrelevanttoourplannedactivitiesin2019.Thetestcryostatincludedacryocoolerwhichproducessignificantvibrations,and thesegotworseas thecooler’soperationalhoursrose.Whenthevibrationswereattheirworst,significantheatingwasseeninthecoldesttwostagesoftheCADR.Temporarilyshuttingoffthecoolereliminatedthisheating,confirmingthatthevibrations caused the heating.While thiswould not be expected for a space‐flight cryocooler, itcertainlyimpactedthelabtestingofthisparticularCADR.WesuspectthattheheatingoccursintheKevlarstringsusedtosuspendthesestages,andweinitiatedastudytoverifythis.Summerstudentsare assisting with this test, in which we have suspended a dummy mass and will vibrate it atcryogenictemperaturesusingavoicecoil.Figure5showsapreliminarytestconfigurationinstalledinacryostat.TinythermometersinstalledatthemidpointsoftheKevlarstringswillmeasureanytemperaturerisesduetovibration,andwewillcorrelatetheseriseswithanymeasuredheatreachingthesuspendedmass.Preliminarytestingatroomtemperaturelookspromisingandindicatesthatweshouldbeabletoresolvetheheatingeffectsoncethesystemiscooledto4K.

125

Fig.5.AnexperimentwhichwillstudytheheatingofourCADR’slow‐temperaturestagesduetocryocoolervibrations.

Inparallelwiththiswork,ourteaminitiatedprocurementofthenew10‐Kmagnets,andweexpectthefirsttoarrivethissummer.Wealsoprocurednearlyall thesuppliesandequipmentneededfortheremainderoftheresearcheffort.MostoftheADRpartsneededforthe10‐to‐4‐KCADRarealreadybeingfabricatedbyoutsidemachineshops,andtheremainingdrawingswillsoonbesubmittedaswell.

InJuly,2017,wepresentedapaperdescribingthisworkattheCryogenicEngineeringConferenceinMadison,Wisconsin.Ourpaperwaswell‐receivedbythecryogenicscommunity,andmanyconferenceattendeesexpressedsupportforthistechnologybeingavailableforfuturespacemissions.InMay,twomembersofourteamparticipatedinapreliminarydesignstudyfortheOriginsSurveySpectrometerinstrumentbeingproposedforOST.Thebaselinedcoolingsystemforthisinstrumentincludesour4‐to‐0.05‐Ksubsystem,providinganotherpotentialcustomerforthistechnology.

PathForwardInthesecondhalfofYear2,wewillproduceasecondstageforthe10‐to‐4‐KCADRandintegrateitwiththeexistingstage.Wewillperformance‐testtheCADRtomeetourYear‐2milestone.Inparallel,wewillcontinuetofabricatepartsneededforthe4‐to‐0.05‐KCADRinYear3,aswellastheoverallmagneticshield.InYear3,wewillproceedwithouroriginaldevelopmentplan,asdescribedabove.

Foradditionalinformation,contactJamesTuttle:james.g.tuttle@nasa.gov