COTSMOSDosimetryontheMeMOSatBoard,ResultsAfter2.5YearsinOrbit
JoséLipovetzky(1*,2,3,4),MarianoGarcia-Inza(1,3),MacarenaRodríguezCañete(1),GabrielRedin(1),SebastiánCarbonetto (1,3),MartínEcharri(5*),FedericoGolmar(3,5,6),FernandoGomezMarlasca(2),MarianoBarella(2,3,5),GabrielSanca(6),Pablo
Levy(2,3,6),AdriánFaigón(1,3) (1)FacultaddeIngeniería,UniversidaddeBuenosAires||PaseoColón850,CiudaddeBuenosAires,Argentina||
Phone +541152850819 (2)ComisiónNacionaldeEnergíaAtómica||Bustillo9500,CentroAtómicoBariloche,Bariloche,Argentina
Phone+542944445100x5349 (3)ConsejoNacionaldeInvestigacionesCientíficasyTécnicas||(4)InstitutoBalseiro,Argentina
Bustillo9500,CentroAtómicoBariloche,Bariloche,Argentina (5)CentrodeMicroyNanoelectrónicadelBicentenario-InstitutoNacionaldeTecnologíaIndustrial(INTI),Argentina
(6)EscueladeCienciayTecnología,UniversidaddeSanMartín(UNSAM),Argentina Mails:jose.lipovetzky@ieee.org,fgolmar@inti.gob.ar,magarcia@fi.uba.ar,afaigon@fi.uba.ar
*formerly ThisworkwassupprotedbyCONICETwithPIP2014-2016-GI-Faigon,UBACyTwith20020150200085BA,andANPCyTwithPICT
2014-1812
Abstract:Wepresenttheresultsafter2.5yearsinor- I.Introduction
bitofTotalIonizingDose(TID)measurementsdone MetalOxideSemiconductor(MOS)dosimetersare
usingMetalOxideSemiconductor(MOS)dosime- MOStransistorswhichallow thequantificationof
tersontheMeMOSatboard.TheMeMosatboard theTotalIonizationDose(TID)throughtheshiftof
waslaunched on July19th 2014attheBugSat-1 “Tita”microsatellitedevelopedbySatellogictostay
the Threshold Voltage (VT) caused by buildup of positivechargeintheoxideandthegenerationofin-
atLEO.Weusedasdosimetersp-channelCommer- terfacetraps[1,2].MOSdosimetershavebeenused
cialOffTheShelf(COTS)MOStransistorswithgate inspaceapplications[3-6]andmedicalapplications
oxidesof250nm.Beforelaunch,asubsetoftransis- [7,8].Toallowhighsensitivities,MOSdosimetersare
torswith similardrain currentto voltage(I-V) usuallymanufacturedinad-hocprocesseswithgate
curveswhereselectedfrom agroupof100devices. oxidesofhundredsofnanometersorevenfewmi-
Thetemperaturedependenceofthe(I-V)curveswas crometers,thickerthanregularsMOStransistors[1-
studiedtofindtheminimumtemperaturecoefficient 4,7,9].SincetheresponsivitytoTIDisapproximately
biasing point. Then, a calibration subgroup of sensorswasirradiatedusinga60Cogammasourceto
proportionaltotheoxidethicknesstox[10,11]. However,severalworkshavedealtwiththeuseof
studytheirresponsetoTID,showingresponsivities CommercialOffTheShelfMOStransistorsasMOS
of~75mV/krad when thesensorsareirradiated dosimetersinmedicalapplications[12]orindustrial
withoutgatebias.Also,thepostirradiationresponse applications[13,14].Thisworkproposestheuseofa
ofthesensorswasmonitored,inordertoincludea COTStransistorasalowcostsensorforon-board
correctionforlow doserateirradiations,yielding dosimetryinaCubesat.
30mV/krad.Abiasingandreadingcircuitwasdevel-
oped in order to allow the reading of up to 4 sensors. Thethresholdvoltagewasmonitoredduringdiffer-
II.Devicesandinitialcalibration ThedevicesusedinthisworkassensorsareCOTSp-
entperiodsofthemission.After2.5yearsinorbit, channelMOSenhancementtransistorswithagate
thethresholdvoltageofthesensormountedonthe oxidethicknessof~250nm.Theoxidethickness---
MeMOSatBoardhadaVTshiftofapproximately35 notinformedbythemanufacturer---wasestimated
mVcorrespondstoadoseof1.2krads.
usinggatetunnelcurrentvsgateoxidevoltagechar-
acteristics of the sensors during Fowler Nordheim
268
currentinjection[13].TheMOSdevicesarepack- B.TIDResponse
agedinaTO-72metalcase.
TheresponseofthesensorstoTID wasstudiedto
ItisknownthattheresponsivitytoTIDofdifferent obtaintheresponsivityofthesensors.Itisknown
MOSdosimeterscanhaveahighdispersion—upto thattheresponsivityofaMOSdosimeterdependson
30%---ifaremanufacturedindifferentSiwafer[13- thegatevoltagesustainedduringtheirradiation.
14].Thisdispersionwouldintroduceerrorsinthe Usually,apositivegatevoltagewillincreasethere-
quantificationofdoseifthedevicesarenotindividu- sponsivity.However,aconstraintimposedbythe
allycalibrated.Thus,initially,asubsetof15devices satelliteplatform totheboardwiththedosimeters
from abatchof100MOStransistorswhereselected wasthattheboardwouldnotbepoweredduring
usingascriteriontohavesimilardraincurrentto mostofthetimeofthemission.Thisimpliesthatthe
gatevoltage(I-V)characteristics,i.e.similarVT and sensorswouldbebiasedduringmostofthemission transconductancevalues.Thesedeviceswherecali- timewithzerovoltsbetweenalldeviceterminals.
brated.
Theresponsetoirradiationofthesensorswasdone
usinga60Cogammarayssource,atadoserateof
A.TemperatureDependence
~0.5Gy/min(50rads/min).Somesensorswherebi-
ThetemperaturedependenceoftheI-Vcurveofthe asedwithagatevoltageofzerovolts(aswouldbeon
devicesselectedforcalibrationwasstudied.TheI-V thesatellite),andotherswithagatevoltageof9V.
curveofMOSdosimetersisaffectedbytwomain Fig.2presentstheVTshiftasafunctionofdosedurtemperaturedependencies,thedecreaseofVT and ingirradiationupto1.6and1.8Gywithgatevoltthedecreaseofthetransconductancewithtempera- agesof9Vand0Vrespectively.Theresponsivities
ture.Usually,botheffectscompensateforagiven ofthedeviceswhere58mV/Gy and6.6mV/Gyfor
draincurrentvalue,knownasZeroTemperatureCo- 9Vand0Vbias.Figure3showstheshiftoftheI-V
efficient(ZTC)current(IZTC).Usually in MOS curveafterirradiationinadevice,showinghow a dosimetry,thispointoftheI-Vcurveisusedtomea- shiftinthegatevoltagerequiredtoapplyagiven
suretheVT shift[4].Fig.1presentstheI-V curves draincurrentcanbeusedasadosimetricmagnitude. measuredatfourdifferenttemperaturesfrom 0oCto
60oC,atemperaturewiderthanthetemperaturefluc-
tuationsexpectedduringthemission.Itcanbeob-
served that the ZTC point ofthe curve is at
IZTC=270µA.However,duetocircuitalrestrictions, thedosimeterswherefinallybiasedwitha200µA
draincurrenttoquantifytheVT shiftontheboard. Thisintroducedasmalltemperaturedependenceof
thereadingaswillbeseenlater.
Fig. 1. I-V characteristics of the selected devices.
Fig. 2. Response to TID during 60Co irradiation of one sample of MOS dosimeters with 0V and 9V gate bias.
TheexperimentofFig.2wasrepeatedwithanincrementaldoseof1.5Gywithzerovoltsbiastothe15 devices of the set,yielding responsivities from 6.6mV/Gyto7.5mV/Gy,provingthateachdevice needs an individualcalibration despite ofhaving similarinitialVT values.Thispre-useirradiationallowstheinitialcalibrationofeachdevice.
269
C.LongTermAnnealing.
III.MemosatBoardandReadingCircuit
AdifficultywiththeuseofMOSdosimetersisfad- TwoCOTSMOStransistorswereusedassensorwas
ing,i.e.thelongterm recoveryofVT withtime,a mountedontheMeMOSat-01board,whichwaspart problem associatedtotheneutralizationofoxide ofthepayloadoftheBugSat-1,amicrosatellitedevel-
trappedchargeviatunnelingorthermalexcitationof oped by Argentine company Satellogic [15],
electrons[1-4].Annealingcanintroduceerrorsin launchedfromtheDombarovskyairbase(Russia)in
thequantificationofTIDifthesensorsarecalibrated June19th,2014.TheBugSat-1isonaLEOorbitwith
atahighdoserateandthenthedosimetryiscarried 620kmofaltitudeandaninclinationof97.9degrees
outatsuchalowdoseratethatannealingcausesa andweights22kg.
significantrecoveryinVT,underestimatingtheab- MeMOSat-xxisareconfigurableplatform fortestsorbeddose.Thisfeatureisalsoknownasapparent ingReRAM nonvolatilememoriesinthesatelliteen-
doseratedependencyoftheresponseofthesensor vironment.Thefirstlaunchedboard,MeMOSat-01
[4].IthasbeenreportedthatthefinalVTshiftaftera wasdesignedtoperform targetspecifictestsontwo shortirradiationatahighdoseratefollowedbylong ReRAM HfO2,generatingreportsincludingexperiterm annealingisapproximatelysimilartotheVT mentalandsystem parameters,reportedperiodically shiftobtainedaftertheexposuretothesamedosein toEarthviainteractionwiththesatellite.MeMOSat-
thesameannealingtimeatamuchlowerdoserate. xxwasthefirststepofLabOSat,acomprehensive
Toinvestigateintothis,oneofthecalibrateddevices projecttoperform experimentsinspace.TheMeM-
wasirradiatedwiththe60Cosourceupto20.3Gyus- OSat-01boardismountedclosetothecenterofthe
ingagatebiasvoltageof0Vbeforethelaunchofthe Bugsat-1satellite,surroundedbyalargeamountof
satellite.ThedevicewaskeptonEarthduringthe massinmostdirections.
missiontimewithallterminalsgrounded,toevaluate
therecoveryofVT afterthemissiontime.Figure3 showstheI-VcurvesofthefreshMOSsensor,after
irradiationandafter2.5yearsofannealing.Initially,
VTshifts-152mV.Butthen,afteranneal,VTrecovers 92mV,a60%.Thus,thefinalafterannealVT shift happenstobe60mV,meaningthatthereallowdose
ratesensitivityofthedosimeterisonly3.0mV/Gy.
Thisisanimportantresult,sincetheirradiationin
spacewillbeperformedinascaletimeofyearsin-
steadofminutes.
Fig. 4. The MemoSat board before mounting the MOS dosimeter.
The sensors are mounted on the bottom corner of the board.
Fig. 3, change in I-V curves after a 20.3Gy (2.03 krads) irradiation with 0 V gate bias and the recovery after 2.5
years.
ThecircuitdesignedtoreadVT shiftsofthesensor wasdesignedtakinginaccountthattheplatform of thesatellitewouldnotprovidepowerduringmostof thetime.Moreover,sincethemainobjectiveofthe missionistotestthesatellite itself,theMeMOSat boardwouldbeturnedononlyduringshortintervals,duringspecificperiodsoftime.Thus,thecircuit hadtoensurethatthesensorwouldhaveallthetime thezerovoltsgatebias,exceptforveryshortreading timesin which thereadingdrain current(IREAD) wouldbeappliedtomeasuretheVTshift.Also,tobe
270
abletomeasureinfuturemissionsthedoseindiffer-
entpointsofthesatellite,thecircuitshouldallowthe IV.ResultsandDiscussion
readingoffourdevicesusingthesameanalogtodigi- ThesensorsweremountedontheMeMOSatboard,
talconverter(ADC).
included on the satellite and launched.Before
Theschematicdesignofthecircuitusedinthe launch,theinitialVTvaluewasread.Thesatellitebeboard ispresented in Fig.5.When theanalog gantotransmitdosimetricinformationafterallthe
switchesS10andS20areopen,thesensor(QRAD- initialchecksandtestswereperformedontherestof
FET)isinthe“Exposure”mode,withallterminals theplatform,164daysafterlaunch.Onlyoneofthe
aregroundedthrough1MΩ resistors.Whenthe twosensorsresultedtobefunctionalafterthispe-
switchesareclosed,theoperationalamplifierU2ap- riod.
pliesonthegateofthesensorthegatevoltagere- Figure6presentstheVTvaluesobtainedduringthe quiredtohaveafixeddrainbiascurrentofIREAD(9V- mission.Therearetwoperiodsoftimeinwhichthe Vref)/Rref.Thegatevoltagewasreducedusingare- informationwasnotrecorded,atthebeginningof
sistivedividerandreadusingana12-bitADC,yield- themissionandduringthesecondyear.
ingaresolutionof2.4mVperADCcount.Theana- ItcanbeobservedthatVT hasacleartendencyto logswitchesareinfacttheswitchesofananalog reduce,withahighdispersion.Thisdispersion---of
multiplexer,allowingthesequentialreadingoffour ~+2ADCcountsor~+5mV---isattributedtother-
deviceswiththesameelectronics.
malvariationsinthesatellite,causinganuncertainty
inthereadingofVTduetothefactthatthereadcurrentwasnotsetexactlytotheZTCpointoftheI-V
curveofthedevice.Thisdispersionisconsistentwith
atemperatureamplitudeof32oC,whichagreeswith
thermalinformationprovidedbythesatellitemanu-
facturerwhichreportedlessthan 40oCoftempera-
tureamplitudeontheboard.Unfortunately,wedid
not have available measurements of temperature
doneatthesametimesofVTreadings.Neweditions oftheboardincludeonboardtemperaturemeasure-
mentsto allow thermalcorrection ofdosimetric
magnitudes.
The threshold voltage shift recorded during the
missionisapproximately-35mV.Takinginaccount
that the calibrated sensitivity of the sensor at low
doserateswas3.0mV/Gy,thedosemeasuredbythe
detectoris~12Gy---i.e.1.2krads.
Thisdoseissurprisinglylowerthanwhatisusually
expectedinaLEOorbit,higherthantensofGy/year
[16]inmostsatellites.However,theexplanationfor
Fig 5. Bias and reading circuit.
thelow dosemeasuredontheexperimentisthat since the board is placed closed to the center of the
DevicerestrictionsforcedtheuseofIREAD=200µA, satellite,mostofthemassoftherestofthesystem whichwas,asexplainedbefore,notexactlytheIZTC servesasashieldingformostparticles.Accordingto value.Thus,thetemperaturedependenceofthevolt- [16]mosthighenergyelectronswillbeshieldedbya
agereadingwasestimatedfromI-Vcurvesatdiffer- layerthickerthan300mils(7.6mm)ofAl,being
enttemperatures,yielding~296µV/oC.
mostofthedosecaused,behindshielding,byhigh
energyprotons.Ontheotherhandthedoserate
causedbyhighenergyprotonsfrom theVanAllen
Beltsreducesapproximately10timesperinchofAl
equivalent shielding.The shielding done by the
271
satellitemassof22kgdistributedaroundthesensor
wouldexplainthelow doseratemeasuredonthe References:
board.Unfortunately,proprietaryinformationofthe [1]AndrewHolmes-Siedle,Thespace-chargedosimeter:
satellitemanufacturercouldnotbedisclosedtousto makeamorecompleteanalysisbased on Monte Carlosimulations.
Generalprinciplesofanewmethodofradiationdetection,NuclearInstrumentsandMethods,Volume121, Issue1,1October1974,Pages169-179,ISSN 0029554X,DOI:10.1016/0029-554X(74)90153-0.
[2]Holmes-Siedle,A.andAdams,L.RADFET:areviewof
the use ofmetal-oxide-silicon devices as integrating
dosimeters,InternationalJournalofRadiationApplica-
tions\\&Instrumentation.PartC,RadiationPhysics\\&
Chemistry},Vol.28No.2,pp.235—244,1986.
[3]Adams,Leonard;Holmes-Siedle,Andrew,"TheDevel-
opmentofanMOSDosimetryUnitforUseinSpace,"
NuclearScience,IEEETransactionson,vol.25,no.6,
pp.1607-1612,Dec.1978
[4]Holmes-Siedle,A.;Ravotti,F.;Glaser,M.,"TheDosi-
metric Performance ofRADFETs in Radiation Test
Beams,"RadiationEffectsDataWorkshop,2007IEEE,
vol.0,no.,pp.42-57,23-27July2007
Fig. 6. Threshold Voltage Shift as a function of time for the dosimeter on the MeMOSat board.
[5]Andrew Holmes-Siedle,Thespace-chargedosimeter: Generalprinciplesofanewmethodofradiationdetec-
tion,NuclearInstrumentsandMethods,Volume121,
V.Conclussionsandfuturework
Issue1,1October1974,Pages169-179,ISSN 0029-
TheworkpresentsthedesignandrealizationofMOS dosimetryontheMeMOSatboard,mountedonthe BugSat-1satellite,ona620km lowearthorbit.The sensorsusedinthisworkwhereCOTSMOStransistors,provingthefeasibilityofusingsuchdevicesin
554X,DOI:10.1016/0029-554X(74)90153-0. [6]MacKay,G.F.;Thomson,I.;Ng,A.;Sultan,N.,"Appli-
cationsofMOSFETdosimetersonMIRandBIONsatellites,"NuclearScience,IEEETransactionson,vol.44, no.6,pp.2048-2051,Dec1997 [7]Rosenfeld,A.B.;Lerch,M.L.F.;Kron,T.;Brauer-
low costexperiments.Thedosemeasuredonthe Krisch,E.;Bravin,A.;Holmes-Siedle,A.;Allen,B.J.,
boardafter2.5yearsis~1.2krad,avaluecompatible "Feasibilitystudyofonlinehigh-spatial-resolutionMOS-
withthefactthattheboardwasprotectedbythe FET dosimetryin staticand pulsed x-rayradiation
massoftherestofthesatellitefrommostofthehigh fields,"NuclearScience,IEEETransactionson,vol.48,
energyparticlesinthespaceenvironment
no.6,pp.2061-2068,Dec2001
Theworkisbeingcontinued,andtheMeMOSat boardhasbeenreplacedbyamorecomplexexperimentnamedLabOSat[20].Duringthepastyears,the FOXFET,anewthickgateoxideMOSdosimeterwas developedandtestedindifferentapplication[9].The
[8]Rosenfeld,A.B."Electronicdosimetryinradiationtherapy",Rad.Meas.41pp134-153,2006.
[9]J.Lipovetzky,M.A.GarciaInza,S.Carbonetto,M.J. Carra,E.Redin,L.Sambuco Salomone,A.Faigon , “FieldOxiden-channelMOSDosimetersFabricatedin CMOSProcesses.”,IEEETransactionsonNuclearSci-
FOXFETprovedtohavelowerfadingandamuch ence,may2013.
highersensitivity,andwillreplacetheCOTSsensors [10]Oldham,T.R."IonizingradiationeffectsinMOSox-
used in thiswork.However,low costacademic ides",WorldScientificPubCoInc,1999.
CubeSatprojectsmightuseCOTSMOStransistors [11]Holmes-Siedle,A.G.andAdams,L.,"Handbookof
asaneffectivewaytoestimatedoseandperform ex- radiationeffects,2002,OxfordUniversityPress,USA.
perimentsinorbit.Theuseofdifferentsensorsin differentpositionsofthesatellitetoevaluatetherelative dose rates observed after differentshielding thicknesses.Alsothelongterm fadingneedstobe takeninaccountforareliabledeterminationofthe
[12]L.J.Asensio,M.A.Carvajal,J.A.Lopez-Villanueva,M. Vilches,A.M.Lallena,A.J.Palma,Evaluationofalowcostcommercialmosfetasradiationdosimeter,Sensors andActuatorsA:Physical,Volume125,Issue2,10January2006,Pages288-295,2006
[13]Lipovetzky,J.;Redin,E.G.;Faigon,A.;"Electrically
realdoseintheexperiment.
Erasable Metal–Oxide–Semiconductor Dosimeters"
272
IEEETransactionsonNuclearScience,Volume54,Issue4,Part3,Aug.2007Page(s):1244–1250 [14]Faigon,A.;Lipovetzky,J.;Redin,E.;Krusczenski,G.; "ExtensionoftheMeasurementRangeofMOSDosimetersUsingRadiation Induced ChargeNeutralization" IEEETrans.onNuclearScience,Volume55,Issue4, Part1,Aug.2008Page(s):2141-2147 [15]www.satellogic.com [16]Falguere,D.;Boscher,D.;Nuns,T.;Duzellier,S.; Bourdarie,S.;Ecoffet,R.;Barde,S.;Cueto,J.;Alonzo,C.; Hoffman,C.,"In-flightobservationsoftheradiationenvironmentanditseffectsondevicesintheSAC-Cpolar orbit,"NuclearScience,IEEETransactionson,vol.49, no.6,pp.2782,2787,Dec2002 [17]Jaksic,A.;Ristic,G.;Pejovic,M.;Mohammadzadeh, A.;Lane,W.,"Characterisationofradiationresponseof 400nm implantedgateoxideRADFETs,"Microelectronics,2002.MIEL2002.23rdInternationalConference on,vol.2,no.,pp.727-730,2002 [18]Ristić,G.andGoluboviv,S.andPejović,M.“SensitivityandfadingofpMOSdosimeterswiththickgateoxide”,Sensors\\&Actuators:A.Physical,Vol.,NO.2-3, pp.153—158,1996. [19]Holmes-Siedle,A.G.,Goldsten,J.O.,Maurer,R.H., & Peplowski,P.N.(2014).RADFETdosimetersinthe belt:TheVanAllenProbesonday365.IEEETransactionsonNuclearScience,61(2),948-954. [20]“LabOSat:Lowcostmeasurementplatform designed forhazardousenvironments”,M.Barella,,G.Sanca,F. GomezMarlasca,G.Rodriguez,D.Martelliti,L.Abanto, P.Levy,F.Golmar,Proceedingsofthe2016SeventhArgentineSymposium andConferenceonEmbeddedSystems(CASE):ISBN:978-987-45523-9-6PrintedIEEE CATALOG: CFP1646V-PRT IEEE XPLORE: CFP1646V-ART(2016).
View publication stats
273
Ver+/-
