Gamma-Ray Detectors

1 3 gamma -RayDetectorsHastingsASmith,Jr., , , , ,eitherdirectly(aswithaproportionalcount erorasolid-statesemiconductordetector)or indirectly(aswithascintillationdetector) , ,wewillpresentsomegeneralinformationonty pesof gamma -raydetectorsthatareusedinnondestru ctiveassay(NDA) , , ( )Inmostdesignstheouterelectrodeisthecyli ndricalwallofthegaspressurevessel,andthe inner(positive) (especiallyofionizationchambers) . 4 l%escintillationandsolid-statedetectorsa remuchmoredesirableforobtainingthe~spect roscopicdetailneededintheenergyrangetypi calofuraniumandplutoniumradiation(approx imately100-1000keV).Gascountersaredescri bedinmoredetailinChapter13,sincetheyarem orewidelyusedforneutrondetection. - - - * ~teCtO~Thesensitivevolumeofascintillatio ndetectorisaluminescentmaterial(asolid,l iquid,orgas)thatisviewedbyadevicethatdet ectsthegamma-ray-inducedlightemissions[u suallyaphotomultipliertube(PMT)].

2 Thescintillationmaterialmaybeorganicorin organic; ,plastics, (thestandardagainstwhichotherscintillato rsarecompared).Somecommoninor-ganicscint illationmaterialsaresodiumiodide(NaI),ce siumiodide(CSI),zincsulfide(ZnS),andlith iumiodide(LiI).Themostcommonscintillatio ndetectorsaresolid, ,bismuthgermanate(BiqGesOl2),commonlyref erredtoasBGO,hasbecomepopularinapplicati onswhereitshighgammacountingefficiencyan d/oritslowerneutronsensitivityoutweighco nsiderationsofenergyresolution(Refs. 3and4). ,2, ,ionized(excited)atomsinthescintillatorm aterial relax , ,the (calledactivators) [NaI(Tl)].Thescintillationlightisemitted isotropically;sothescintillatoristypical lysur-roundedwithreflectivematerial(such asMgO)tominimizethelossoflightandtheniso pticallycoupledtothephotocathodeofaPMT.

3 ( )Scintillationpho-tonsincidentonthephoto cathodeliberateelectronsthroughthephotoe lectriceffect, ,theycollidewithelectrodesinthetube(know nasdynodes) ,causingalargemultiplication(byafactorof 104ormore)oftheelectronfluxfromitsinitia lvalueatthe gamma -RayDetectors47 solid -STATECRYSTALSIGNAL,-1* :./.:,;.~.:~nDEPLETED(SENSITIVE)REGION~c ::: ~::: ::R$,..,,:,,.:.pHIGH:VOLTAGE+~ (seeChapter4).(a)(b)(c) :(a)open-endedcylindricalortruecoaxial,( b)closed-endedcylindrical,and(c) , (theso-calledtruecoaxial)orclosed-endedc rystals[ (a-b) ,Jr.,andMarciaLucastheelectricfieldforch argecollectionisprimarilyradial, ,theradialeiectricfieldgeometrymakesthec oaxial(especiallytheopen-endedcoaxial) [forexample, (c)]. ,withthesmallthicknessesoptimumforlow-en ergymeasurements(forexampleintheL-x-rayr egionforspecialnuclearmaterial).]

4 Planardetectorsusuallyachievethebestener gyresolution,becauseoftheirlowcapacitanc e;theyarepreferredfordetailedspectroscop y, , ,inenvironmentswhereneutronlevelsarehigh (suchasaccelerators,reactors,orin-strume ntswithintenseneutronsources), ,radiationdamageeffectscanbeofconcerninN DAapplicationswherelargeamountsofnuclear materialarecontinuouslymeasuredwithhigh- resolution, gamma -rayspectroscopyequipmen t forexample, ,reducestheamplitudesofsomefull-energypu lses, ,theresolutionisdegraded, ( ).Ithasbeengenerallyobservedthatsignific antperformancedegradationbeginswithaneut ronfluenceofapproximately109n/cm2,anddet ectorsbecomeunusableatafluenceofapproxim ately1010n/cm2( ).However, ~iirming(annealing)thedetectorcrystal( ).

5 ,2, ,thereverse-bias-diodeconfigurationofage rmaniumsolid-statedetectorresultsinveryl owcurrentsinthedetector(usuallyinthepico -tonanoampererange).Thisleakagecurrentca nbefurtherreducedfromitsroom-temperature valuebycryogeniccoolingofsolid-statemedi um,typicallytoliquidnitrogentemperature( 77K).Thiscoolingreducesthenatural,therma llygeneratedelectricalnoiseinthecrystalb utconstitutesthemaindisadvantageofsuchde tectors:thedetectorpackagemustincludecap acityforcooling,andthisusuallyinvolvesad ewarforcontainingtheliquidcoolant. Inrecentyears,attemptshavebeenmade- gamma -RayDetectors491I,~n= ,! <!,,!t *. ~.:; ;.#~ viI_ . ~\-*<#nqt&@$#$:, keV;!,d-t,~~..,. :t10,/i4006008001000I 0= ~ ,:.;>;!%.*.* .; 1I1[.. ~-d~;&.]

6 J~~~ III qj$s$i,- !~;&/$102004006008001000 -2 XIO*ln/cm21$ (GAMMAENERGY)Thedeteriorationofahigh-res olutionsolid-statedetectorgammaspec-trum withincreasingneutronjhence(4). ( ) ,Jr.,andMarciaLucastocoolthedetectormate rialelectronically( ),buttheseeffortsarestillintheexperiment alstages, [Si(Li)].Theloweratomicnumberofsiliconco mparedtogermaniumreducesthephotoelectric efficiencybyafactorofabout50(seeChapter2 ),butthistypeofdetectorhasbeenwidelyused inthemeasurementofx-rayspectrainthe1-to5 0-keVenergyrangeandfindssomeapplicationi nx-rayfluorescence(XRF)measurements(seeC hapter10).Thelowphotoelectricefficiencyo fsiliconabove50keVisanadvantagewhenmeasu ringlow-energyxraysandgammarays, ,aswellasmanyotherapplica-tionsofgamma-r ayspectroscopy,itwouldbeadvantageoustoha vehigh-resolutiondetectorsoperatingatroo mtemperature, ,Hg12,andGaAshasbeenextensivelyresearche d( ).

7 Theirhigheraverageatomicnumbers, ,thesedetectormaterialshaveenjoyedlimite dap-plicationtoNDAproblemstodate, ,thesedetectorsmaybecomemoreattractiveas convenient, ~-RayEnergyAtomice-hPair(6)aResolutionat 122keV~MaterialNumbers(eV)(keV)Ge(77K) (300K)48, (300K)80, (300K)31, (300K)c11, Thisquantitydeterminesthenumberofchargec arriersproducedinaninteraction.( )bRepresentativeresolutiondata, Whilenotasemiconductormaterial, , (MCA), (spectrum)ofthedetectedoutputpulses, , (a). :(1)thequantumuncertaintiesintheenergies ofthetransi-tions(theso-calledHeisenberg Uncertainty),and(2) , ,the ideal , , (b). (b) ,afterleavingthesample,willbescatteredby externalmaterialsbeforeenteringthedetect or,andthiseffectcanshowupinthefinalenerg yspectrum(seebelow).

8 Whenthegammarayentersthedetectionmedium, ittransferspartorallofitsenergytoanatomi celectron, ,inaseriesofcollisions, (seeTable3-1).Aphotoelectricinteractiont ransfersalloftheincidentphoton senergytoaphotoelectron; ,Jr., (~~~j~EOPULSESFROMSINGLEPHOTONINTERACTIO NSINTHEDETECTOREc1t1 -RayEnergyE;(a)(b)(c)Anidealizationofthe photonspectrum(a)producedbyPeenuclei,(b) emergingjiomamaterialsample,and(c) - . ]ypartoftheincidentphoton senergytoanionizedelectron; ;however, (c).Themaximumenergythatcanbedepositedin thedetectionmediumfromaComptonscattering eventcomesfromaneventwherethephotonissca tteredby180 .TheCompton-generateddetectorpulsesareth ereforedistributedbelowthismaximumenergy [ (c)]andconstituteasourceof background (c) ( ).)

9 A3n MONOENERGETICGAMMA-RAYFLUXGEcBEoDETECTOR PULSEAMPLITUDE(y-RAYENERGY) ,Jr., , ( ). ,distributedsmoothlyuptoamaximumenergyEC ( ), (EC) head-on collisionbetweenthephotonandtheelectron, wheretheelectronmovesforwardandthegamma- rayscattersbackwardthrough180 ( ). Comptonl%lley. Foramonoenergeticsource,pulsesinthisregi onarisefromeithermultipleComptonscatteri ngeventsorfromfull-energyinteractionsbyp hotonsthathaveundergonesmall-anglescatte ring(ineitherthesourcematerialsorinterve ningmaterials) , -120 ,amonoenergeticsourcewillgiverisetomanys catteredgammarayswhoseenergiesarenearthi sminimumvalue( ). ,eventsinthisregionarefromhigh-energygam maraysandcosmic-raymuonsinthenaturalback groundandfrompulse-pileupeventsifthecoun trateishighenough(seeChapter4).

10 Inamorecomplexspectrum, ,verynearthe zero-pulse-height-amplitude region,arisestypicallyfromlow-amplitudee lectronicnoiseinthedetectionsystemthatis processedlikelow-amplitudedetectorpulses (seeChapter4), ,containingmanydifferentphotonenergies,t heCompton-edgeandbackscatterpeakfeatures tendto ~ullwidthofthe(full-energy)photopeakatHa lfits~aximumheight(FWHM).IfastandardGaus sianshapeisassumedforthephotopeak,theFWH M isgivenbyFWHM=2um(3-1) (smallFWHM) , ,suchasnoise,pulsepileup,improperpole-ze rosettings, , (Tl) ~ , , ,Jr.,andMarciaLucas ~3in, (Tl)$.[ !.L~2iii10qp.: I Averageenergy(ineV)requiredtoproduceonee lectron-ionpairinthedetectingmedium.~The ratioE/dforE=300keV(Equation3-2).cThequa ntityo(n)/n,or(l/n)l/2,withouttheincorpo rationoftheFanofactor.]