ebook img

Fermi's golden rule applied to the gamma decay in the quasicontinuum of 46Ti PDF

0.53 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Fermi's golden rule applied to the gamma decay in the quasicontinuum of 46Ti

Fermi’sgoldenruleappliedtotheγ decayinthequasicontinuumof46Ti M.Guttormsen,1,∗ A.C.Larsen,1 A.Bu¨rger,1 A.Go¨rgen,1 S.Harissopulos,2 M.Kmiecik,3 T.Konstantinopoulos,2 M.Krtic˘ka,4 A.Lagoyannis,2 T.Lo¨nnroth,5 K.Mazurek,3 M.Norrby,5 H.T.Nyhus,1 G.Perdikakis†,2 A. Schiller,6 S. Siem,1 A. Spyrou†,2 N.U.H. Syed,1 H.K. Toft,1 G.M. Tveten,1 and A. Voinov6 1Department of Physics, University of Oslo, N-0316 Oslo, Norway 2InstituteofNuclearPhysics, NCSR”Demokritos”, 153.10AghiaParaskevi, Athens, Greece 3Institute of Nuclear Physics PAN, Krako´w, Poland 4Institute of Particle and Nuclear Physics, Charles University, Prague, Czech Republic 1 5Department of Physics, A˚bo Akademi University, FIN-20500 A˚bo, Finland 1 6Department of Physics and Astronomy, Ohio University, Athens, Ohio 45701, USA 0 (Dated:January21,2011) 2 Particle-γ coincidencesfromthe46Ti(p,p(cid:48)γ)46Tiinelasticscatteringreactionwith15-MeVprotonsareuti- n lizedtoobtainγ-rayspectraasafunctionofexcitationenergy. Therichdatasetallowsanalysisofthecoinci- a dencedatawithvariousgatesonexcitationenergy.Formanyindependentdatasets,thisenablesasimultaneous J extractionofleveldensityandradiativestrengthfunction(RSF).Theresultsareconsistentwithonecommon 0 leveldensity. ThedataseemtoexhibitauniversalRSFasthededucedRSFsfromdifferentexcitationenergies 2 show only small fluctuations provided that only excitation energies above 3 MeV are taken into account. If transitionstowell-separatedlow-energylevelsareincluded,thededucedRSFmaychangebyafactorof2−3, ] x whichmightbeexpectedbecausetheinvolvedPorter-Thomasfluctuations. e - PACSnumbers:21.10.Ma,25.20.Lj,27.40.+z,25.40.Hs l c u n I. INTRODUCTION tainedfromthestudyofphotonuclearcross-sections[4],and [ thuslimitedonlytoenergiesabovetheparticleseparationen- ergy.Itwasestablishedthatthegiantelectricdipoleresonance 2 Fermi’s golden rule predicts the transition rate from one v state to a set of final states in a quantum system. The the- (GEDR)dominatestheRSFinallnuclei. Theinformationon 9 RSF below particle separation energy is significantly less. It oretical foundation, which has been successfully applied in 0 hasbeenobtainedmainlyfromtheOslomethodand(n,γ)and manydisciplinesofphysics,wasfirstestablishedbyDiracin 4 (γ,γ(cid:48))reactions. 1927 [1] and emphasized by Fermi in his book in 1950 [2]. 3 The Oslo method, which is applicable for excitation ener- . The rule is based on first-order perturbation theory, where 9 gies below the particle separation, is described in detail in the transition matrix element is assumed to be small. In nu- 0 Ref. [5]. In this work, we report on results obtained for clearphysics,thisassumptioniswellfulfilledforβ andγ de- 0 the 46Ti nucleus, which has two protons and four neutrons 1 cay. Thus, wetakethevalidityoftheFermi’sgoldenruleas outside the doubly-magic 40Ca core. The advantages of the : granted, rather than testing this rule. In this work, we study v theγ decaybetweenstatesinthequasi-continuumof46Tiand (p,p(cid:48)) reaction compared to the commonly used (3He,3He(cid:48)) i and (3He,4He) reactions are much higher cross-sections and X apply Fermi’s golden rule to disentangle the γ strength and better particle resolution. This allows us to make a detailed r leveldensity. a studyofγ-decayasafunctionofexcitationenergy. TheOslonuclearphysicsgrouphasdevelopedamethodto It can be discussed if the concept of one unique RSF is determine simultaneously the level density and the radiative validforlightnucleiandinparticularatlowexcitationener- strength function (RSF) from particle-γ coincidences. These gies. ThisquestiontogetherwiththeapplicabilityoftheOslo quantitiesprovideinformationontheaveragepropertiesofex- methodforlightsystemsastitaniumisthemainsubjectofthis citednucleiandareindispensibleinnuclearreactiontheories work. astheyaretheonlyquantitiesneededforcompletedescription In Sec. II the experimental results are described. The nu- oftheγ decayathigherexcitationenergies. clear level density and RSF are extracted in Sec. III, and in Thenuclearleveldensitycanbedeterminedreliablyuptoa Sec.IV,theapplicationofFermi’sgoldenruleandtheBrink fewMeVofexcitationenergyfromthecountingoflow-lying hypothesisisdiscussed. Summaryandconclusionsaregiven discreteknownlevels[3]. Thereisalsoreliablelevel-density inSec.V. informationfromneutronresonancesathigherexcitationen- ergies;however,thisinformationisrestrictedinenergyaswell asspinrange. II. EXPERIMENTALRESULTS ThemostrichexperimentalinformationontheRSFwasob- TheexperimentwasconductedattheOsloCyclotronLabo- ratory(OCL)witha15-MeVprotonbeambombardingaself- supporting target of 46Ti with thickness of 1.8 mg/cm2. The †Currentaddress:NationalSuperconductingCyclotronLaboratory,Michigan targetwasenrichedto86%46Tiwith48Ti(11%)asthemain StateUniversity,EastLansing,MI48824-1321,USA. ∗Electronicaddress:[email protected] impurity. Thissmalladmixtureof48Tiisnotexpectedtoplay 2 · 103 ) V V 2500 (a) Singles 46Ti 2+ 0+ (ke10000 9.3 ke 12500000 4+ E i 110044 3 s / 8000 nt 1000 0+ u o C 500 6000 110033 0 · 103 1400 (b) Coincidences 2+ 4000 V 1200 e k 3 1000 110022 9. s / 3 680000 2000 ount 400 4+ C 0+ 200 0 2000 4000 6000 8000 10000 0 4 6 8 10 12 14 E (keV) g Proton energy E (MeV) FIG. 2: (Color online) The particle-γ coincidence matrix for 46Ti. FIG.1:(Coloronline)Singles(a)andcoincidence(b)protonspectra recordedwith15-MeVprotonson46Ti. The γ-ray spectra have been unfolded with the NaI response func- tions. a major role. This is supported by the fact that at low exci- informationontheleveldensityandRSFcanbeextractedsi- tationenergieswheretheleveldensityissmall,wecouldnot multaneously. identifyanyimportantcontributionsof48Tiintotheγ-spectra. An iterative subtraction technique has been developed to Inaddition,bothtitaniumsareexpectedtobehavesimilarly. separate out the first-generation (primary) γ-transitions from Particle-γ coincidencesweremeasuredwitheightSi∆E− the total cascade [8]. The subtraction technique is based on E particle telescopes and the CACTUS multidetector sys- theassumptionthatthedecaypatternisthesamewhetherthe tem [6]. The Si detectors were placed in forward direction, levels were initiated directly by the nuclear reaction or by γ 45◦ relativetothebeamaxis. Thefront(∆E)andend(E)de- decay from higher-lying states. This assumption is automat- tectorshadathicknessof140µmand1500µm,respectively. icallyfulfilledwhenstateshavethesamerelativeprobability TheCACTUSarrayconsistsof28collimated5”×5”NaI(Tl) to be populated by the two processes, since γ-branching ra- γ detectorswithatotalefficiencyof15.2%atE =1.33MeV. tios are properties of the levels themselves. If the excitation γ Thesingles-protonspectrumandprotonsincoincidencewith binscontainmanylevels,itislikelytofindthesameγ-energy γ-raysareshowninFig.1. distributionfromthissetoflevelsindependentofthetypeof In total, 110 million coincidence events were collected in population. However, the assumption is more problematic if oneweekwithabeamcurrentof1.5nA.Usingreactionkine- the decayinvolves only a few(but not one) levelswithin the matics,themeasuredprotonenergywastransformedintoex- energybin. citation energy of the residual nucleus. In this way, a set of Fermi’sgoldenrulepredictsthatthedecayprobabilitymay γ-rayspectraisassignedtoaspecificinitialexcitationenergy be factorized into the transition matrix element between the E in 46Ti. Furthermore, the γ-ray spectra are corrected for initialandfinalstates,andthedensityatthefinalstate[2]: i theknownresponsefunctionsoftheCACTUSarrayfollowing 2π theproceduredescribedinRef.[7].Theunfoldedcoincidence λi→f = |(cid:104)f|H(cid:48)|i(cid:105)|2ρf. (1) h¯ matrix(E,E )of46TiisshowninFig.2. i γ Thecoincidencematrixdisplaysverticallinesthatrepresent Realizing that the first generation γ-ray spectra P(Ei,Eγ) yrast transitions from the last steps in the γ-cascades. How- are proportional to the decay probability from Ei → Ef = ever, there are also clear diagonal lines where the γ energy Ei−Eγ,i.e.λi→f,wemaywritetheequivalentexpressionof matchesthedirectdecaytothegroundstate(E =E )andto Eq.(1)as: i γ the first and second excited states at 889 keV (2+) and 2010 P(E,E )∝T ρ , (2) keV (4+), respectively. These γ-rays are primary transitions i γ i→f f in the γ-cascades. By studying the energy distribution of all where T is the γ-ray transmission coefficient, and ρ = i→f f primary γ-rays originating from various excitation energies, ρ(E −E )istheleveldensityattheexcitationenergyE af- i γ f 3 tertheprimaryγ-rayemission.Thisexpressiondoesnotallow Todemonstratehowwelltheprocedureworks,wecompare ustoextractsimultaneouslyT andρ fromtheexperimen- inFig.3forsomeinitialexcitationenergiesE theexperimen- i→f i tal P(E,E ) matrix. To do that, either one of the factorial talfirst-generationspectraPwiththeonesobtainedbymulti- i γ functionsmustbeknown, orsomerestrictionshavetobein- plying the extracted T and ρ functions. The agreement be- troduced. According to the Brink hypothesis [9], the γ-ray tween calculated and experimental first-generation spectra is transmission coefficient is independent of excitation energy; excellentfordecayfromhigherexcitations;howeverthereare onlythetransitionalenergyE playsarole. Thus,wereplace locallystrongdeviationswherethecalculatedspectrafallout- γ T withT(E ),giving sideoftheexperimentalerrorbarsforpopulationsoflowerex- i→f γ citationsenergies,asseenintheE =5.6and6.4MeVgates. P(E,E )∝T(E )ρ , (3) i i γ γ f ThesedeviationscouldbetheconsequenceofPorter-Thomas which permits a simultaneous extraction of the two multi- fluctuations [11], which will be further discussed in Sec. IV. plicativefunctions. WethenfitaboutN2/2datapointsofthe The general good agreement holds also for all the other 40 Pmatrixwith2N freeparameters. Aleast χ2 fitisthenpos- spectra (not shown) included in the global fit with the same sible,sinceitismanymoredatapointsthanfitparameters;in T(E )andρ(E)functions. γ thepresentcasewehave150freeparameterstofit2240data The experimental statistical errors are very small as seen points. in Fig. 3. Thus, the deviations are due to other sources of At low excitation energy, the γ decay is highly dependent errors. For example for E =6.4 MeV, the T ·ρ prediction i ontheindividualinitialandfinalstate;therefore,wehaveex- overestimatesaroundE =3MeVandunderestimatesaround γ cluded γ-ray spectra originating from excitation energy bins E =5.5 MeV with several standard deviations. Thus, there γ belowEi=5.5MeV. areindicationsthatacommonT andρ functionthatsimulta- ItiswellknownthattheBrinkhypothesisisviolatedwhen neouslyfitstheP(E,E )matrix,couldnotbefound.Thesys- i γ hightemperaturesand/orspinsareinvolved,seeRef.[10]and tematicerrorsbehindthesedeviationscouldbeduetoseveral referencestherein. However,inthepresentOsloexperiment, factorsasexperimentalshortcomings,assumptionsbehindthe thetemperaturereachedislow(T ∼1.5MeV)andisassumed Oslo method, the Brink hypothesis [9] and, most probable, toberatherconstant. Thedependencyonspinisofminorim- Porter-Thomasfluctuations. Keepingallthesepossibilitiesin portance.Themeasuredratiosoftheγfeedingintotheground mind,theresultsofFig.3areverygratifying. band indicate a low spin window of I ∼0−6h¯. At excita- There exist infinitely many T and ρ functions making tionenergyE∼8MeVtheratiosare57:100:9forthe2+,4+ identicalfitstothedata[5]astheexamplesshowninFig.3. and6+ states,andatE ∼10MeVtheyare55:100:10,which Thesefunctionscanbegeneratedbythetransformations areequalwithintheerrorbarsforextractingtheseratios. Of course,atlowexcitationenergythespindistributionfluctuates ρ˜(Ei−Eγ) = Aexp[α(Ei−Eγ)]ρ(Ei−Eγ), (4) due to a few levels (and spins) present within each 118 keV T˜(E ) = Bexp(αE )T(E ). (5) γ γ γ excitationbin. Inprinciple,theBrinkhypothesisassumedinexpression(3) Inthefollowing,wewilltrytodeterminetheparametersA,α, could bias the analysis, so that the Oslo method in the next andB. Thisinformationisnotavailablefromourexperiment turn validates the Brink hypothesis. This issue will be ad- andhastobedeterminedfromotherexperimentalresults. dressed in Sec. IV, where we find one common level density First, the level density at high excitation energy is nor- (according to Fermi’s golden rule), which in turn results in malized to the level density at the neutron separation energy oneuniversalRSFinthequasi-continuumof46Ti. ρ(S ). This data point is calculated from neutron resonance n spacingsD (see,e.g.,Ref.[12])withaspindistributiongiven 0 by[13] III. LEVELDENSITYANDRADIATIVESTRENGTH FUNCTION g(E,I)(cid:39) 2I+1exp(cid:2)−(I+1/2)2/2σ2(cid:3), (6) 2σ2 In our first investigations, we rely on the Brink hypothe- whereE isexcitationenergyandIisspin. sis from the expression (3) and factorize the first-generation Thereexistsnoneutronresonancedatafor46Ti. Wethere- γ-matrix P(E,E ) into one transmission coefficient T(E ) i γ γ foreestimateρ(S )fromtheparameterizationsofvonEgidy n andoneleveldensityρ(E). Sincethedecaybetweenstatesat andBucurescu[14]usingtheback-shiftedFermigas(BSFG) lowexcitationenergycannotbetreatedwithinastatisticalap- model,whichreads proach,acutofthematrixwithE >1.8MeVand5.5MeV γ √ <Ei <10.0 MeV was used to exclude clear non-statistical exp(2 aU) decay routes1. The least χ2 fitting of the two multiplicative ρBSFG(E)=η √ , (7) 12 2a1/4U5/4σ functionsfollowstheiterativeprocedureofRef.[5]. whereaistheleveldensityparameter,U =E−E1isthein- trinsic excitation energy, and E1 is the back-shift energy pa- rameter. Thespincut-offparameterσ isgivenby[14] 1ThelowerexcitationcutconcernsonlytheinitialexcitationenergyEi;one √ stillneedsγ-spectraoriginatingfromexcitationregionsdowntotheground σ2=0.0146A5/31+ 1+4aU, (8) stateinordertosubtracthigher-ordergenerationsofγ-rays. 2a 4 0.1 V e k E = 5.6 MeV V E = 6.4 MeV V E = 7.2 MeV 8 0.08 i ke i ke i y / 11 0.06 y / 118 y / 118 bilit 0.04 abilit abilit Proba 0.02 Prob Prob 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 V 0.1 Eg (MeV) Eg (MeV) Eg (MeV) e V V k E = 8.0 MeVe E = 8.9 MeV e E = 9.7 MeV y / 118 00..0068 i y / 118 k i y / 118 k i abilit 0.04 babilit babilit b o o Pro 0.02 Pr Pr 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 Eg (MeV) Eg (MeV) Eg (MeV) FIG.3:(Coloronline)Comparisonofexperimentalfirst-generationspectra(squares)andtheonesobtainedfrommultiplyingtheextractedT andρfunctions(redline).TheinitialexcitationenergybinsEiare118keVbroad.Theerrorbarsrepresenttheexperimentalstatisticalerrors. Abeingthenuclearmassnumber. [18] is shown in Fig. 5 for comparison (see triangle point at Figure 4 shows the extracted total level densities for tita- 15.5MeV). niumisotopeswithknownresonancespacingsD0atSn(filled Oneshouldstressthattheleveldensityfoundfromtheex- triangles). The resonance spacings for (cid:96)=0 neutrons only perimentisbasedonthespinandparitylevelspopulatedinthe givetheleveldensitiesforoneortwospinsandonlyonepar- (p,p(cid:48))reaction.Thus,thenormalizationtothetotallevelden- ity.Inordertoextracttheleveldensityforallspinandparities, sity described above, rests on the assumption that the struc- weuseEq.(6)andassumeequallymanypositiveandnegative tureoftheleveldensityremainsapproximatelythesameifall paritystatesatSn. Thisissupportedbycombinatorialquasi- spinsandparitiesareincluded. Thisassumptionisreasonable particle models [15–17], where the numbers of positive and fulfilled according to Ref. [16], where the level density for negativeparitystatesatSnarepredictedtobethesame. spinwindowsof2−6h¯ and0−30h¯ havebeencalculatedfor The points connected with lines are based on the semi- 46Ti within a combinatorial quasiparticle model [17]. From empirical estimate of von Egidy and Bucurescu [14] with a these estimates, our measurement includes 70−80% of the common scaling of η =0.5 in order to match qualitatively totalleveldensityandtheleveldensityfinestructuresforthe theexperimentalρ(S )points. Theestimatedvaluefor46Tiis twospinwindowsareverysimilar. n ρ =(4650±1000)MeV−1 atSn=13.189MeV(seeFig.4). The measured level density describes nicely the known The error bar chosen reflects roughly the general deviation leveldensityuptoaround4MeVofexcitationenergy.Theex- betweentheglobalestimatesandthepointsderivedfromneu- perimentalresolutionofρ isabout0.3MeVatlowexcitation tronresonancedata. energy,asseenforthe2+ stateat889keV.Ataround3MeV, Now the scaling (A) and slope (α) parameters of the level weseeanabruptincreaseinleveldensityduetothebreaking density can be determined as shown in Fig. 5. The normal- of proton and/or neutron pairs coupled in time-reversed or- izationoftheleveldensityisfittedtotheknownleveldensity bitals(Cooperpairs[19]).Thisresultsin∼10timesmorelev- around 3.5 MeV of excitation energy, and to the extrapola- els,makingitdifficulttodeterminetheleveldensityathigher tionfromρ(Sn)usingtheBSFGmodelwithparameterssum- excitationenergieswithconventionalspectroscopicmethods. marized in Table I. By choosing other ρ(Sn) values (within ItremainstodeterminethescalingparameterBofthetrans- the assumed uncertainty of ±1000 MeV−1), the logarithmi- mission coefficient T(E ). Here we use the radiative width γ cal slope of the ρ and T will change accordingly, as the α- (cid:104)Γ (cid:105) at S assuming that the γ-decay is dominated by dipole γ n parameter of Eqs. (4, 5) has to be adjusted. The ρ curve is transitions. ForinitialspinI andparityπ, thewidthisgiven wellfixedat∼3.5MeVandwillrotatearoundthispointwith by[20] different choices of α. Level densities in the 14−19 MeV excitation region of 46Ti have been measured from Ericson 1 (cid:90) Sn flthuectvuaalutieosnsre,psoereteRdebf.y[t1h8e]vaanrdioruesfeerxepnecreims ethnetarelignr.ouHposwdeivffeerr, (cid:104)Γγ(cid:105)= 2πρ(Sn,I,π)∑If 0 dEγBT(Eγ) within a factor of ten. The measurement of the Ohio group ×ρ(S −E ,I ), (9) n γ f 5 Calculated level density Present experiment Level density from neutron res. data 104 Known levels Back-shifted Fermi gas Estimated level density From neutron res. data From Ericson fluctuation data 1) 104 1) -V -V Me Me103 (Sn E) ( y at 47Ti 45Ti 48Ti 46Ti r (y nsit 103 49Ti nsit102 de 50Ti de el el ev 51Ti ev L L 10 52Ti 53Ti 102 1 6 7 8 9 10 11 12 13 14 0 2 4 6 8 10 12 14 Neutron separation energy S (MeV) Excitation energy E (MeV) n FIG.4:Deducedtotalρ(Sn)fromneutronresonancespacings(trian- FIG.5:Normalizationofthenuclearleveldensity(filledsquares)of gles).Thedatapointfor46Ti(square)isestimatedbyextrapolations 46Ti. Atlowexcitationenergies, thedataarenormalized(between fromtheBSFGmodelwithglobalparameterizationofvonEgidyand thearrows)toknowndiscretelevels(solidline).Athigherexcitation Bucurescu[14](cirleswithlinestoguidetheeye). energies,thedataarenormalizedtotheBSFGleveldensity(dashed line)goingthroughthepointρ(Sn)(opensquare),whichisestimated from the systematics of Fig. 4. For comparison a data point [18] obtainedfromEricsonfluctuationsareshown(blacktriangle). wherethesummationandintegrationrunoverallfinallevels withspinI thatareaccessiblebyγ radiationwithenergyE f γ andmultipolarityE1orM1. Furtherdetailsonthenormaliza- tionprocedurearefoundinRefs.[5,21]. ties in the value of ρ at Sn will change the slope of the RSF, Sincethereexistsnoneutronresonancedata,weagainhave and thus also the degree of low-energy enhancement. How- to rely on systematics from other isotopes. From Fig. 6, we ever,thelinesofFig.7showthattheenhancementisnotvery estimate an average γ-width of (cid:104)Γγ(cid:105)=(1200±500) meV at sensitivetoreasonablechoicesofthevalueρ(Sn). S . The large uncertainty in (cid:104)Γ (cid:105) gives an absolute normal- Such or similar enhancement has been seen in several n γ izationofT(E )andtheRSFwithinabout±50%. However, lightnucleiwithmassA<100, seeRef.[24]andreferences γ thisuncertaintydoesnotinfluencetheenergydependence. therein. Thereisstillnotheoreticalexplanationforthisvery The deduced RSF for dipole radiation can be calculated interestingphenomenon. fromthenormalizedtransmissioncoefficientT(E )by[22] γ 1 T(Eγ) TABLE I: Parameters used for the extraction of level density and f(Eγ)= 2π E3 . (10) radiativestrengthfunctionin46Ti. γ ThenormalizedRSFisshowninFig.7. Forcomparison,the Sn a E1 σ ρ(Sn) (cid:104)Γγ(Sn)(cid:105) (MeV) (MeV−1) (MeV) (MeV−1) (meV) GEDR data from Ref. [23] are also shown, which have been translatedfromphotoneutroncrosssectionσ toRSFby[22] 13.189 4.7 -1.0 4.0 4650(1000) 1200(500) 1 σ(Eγ) f(E )= . (11) γ 3π2h¯2c2 Eγ Unfortunately, there is a large energy gap between our data IV. FERMI’SGOLDENRULEANDBRINKHYPOTHESIS ending at a E =10 MeV and the GEDR data that start at γ 14MeV. According to Fermi’s golden rule, it is possible to factor- Our data on the RSF display a minimum near 4−6 MeV ize out the level density from the γ-decay probability. Since and some structures at lower γ-ray energies. The uncertain- theleveldensityisauniquepropertyofthenucleus,thesame 6 Neutron resonance data Present experiment 3500 High r (Sn) Estimated Low r (Sn) (g ,abs) resonance data 3000 3) -V e M meV) 2500 49Ti 47Ti on (10-7 ( 2000 48Ti 46Ti cti æ g 51Ti 50Ti un GÆ 1500 h f t g n 1000 e r t s e 10-8 500 v ti a 0 di 6 7 8 9 10 11 12 13 14 Ra Neutron separation energy S (MeV) n FIG.6: Experimentalaverageγ-width(cid:104)Γγ(cid:105)fromneutronresonance 0 2 4 6 8 10 12 14 16 18 20 data(triangles)[22].Onlyaroughestimatefor46Ticanbeachieved g -ray energy E (MeV) g (square). FIG.7: (Coloronline)Experimentalradiativestrengthfunctionfor 46Ti (squares). The RSFs assuming high (5650 MeV−1) and low extractedleveldensityisexpectedfromthedecayprobability (3650MeV−1)leveldensityatSn arealsoshown. Forcomparison deduced at different excitation energy regions within the ex- theGEDRdatafromthe(γ,abs)reaction[23]areshown(triangles). perimental restrictions, in particular the energy resolution of CACTUS. In order to see if the Oslo method gives a unique level density, we have divided the data set into three statis- finedby tically independent initial excitation-energy regions, namely Ei=5.5−7.0MeV,7.0−8.5MeVand8.5−10.0MeV,and (cid:82)EidE T(E )ρ(E −E ) applied the same methodology as described in Sec. III. The N (E)= 0 γ γ i γ . (13) results shown in Fig. 8 are very satisfactory. The different i (cid:82)0EidEγP(Ei,Eγ) datasetsgiveapproximatelythesameleveldensity. Thus,the The degree of correctness of the approximation (12) is typi- disentanglementofleveldensityandtransmissioncoefficient callyillustratedbythefitsinFig.3. fromγ-particlecoincidencesseemstoworkverywellaccord- Inthefollowing,wewillinvestigatethedependencyofthe ingtoEq.(2). deduced RSF on initial and final excitation energy. If the Thefactthatwemeasureapproximatelythesamelevelden- Brinkhypothesisisvalid,thereexiststhesameRSFforallex- sity function for primary γ-ray spectra taken at different ex- citationenergiesinthisnucleus. Wewillcallittheuniversal citation regions indicates that the RSF depends only on γ- RSF.InrealitythePorter-Thomasfluctuationsinvolvedinflu- ray energy and not on excitation energy. This seems to in- encetheRSFobtainedfromdifferentexcitationregions. We dicate the validity of Brink hypothesis, which will be tested willusethetermdeducedRSFinthefollowingforthequan- morethoroughlyinthefollowing. Inprinciple,thetestcould tityobtainedfromexperimentaldata. ThededucedRSFsare be performed by dividing the data set into even more initial expected to fluctuate around the universal RSF and the fluc- excitation-energy regions. However, the statistics of the ex- tuations are expected to be stronger if the number of transi- periment does not permit such an approach, and a different tionsusedinthedeterminationofthededucedRSFissmaller. approachhasbeenused. Thus, the deduced RSFs extracted from data sets involving By accepting the level density obtained with the global fit initialandfinalregionsofhighleveldensityshouldbemuch ofallrelevantdata(seelowestcurveinFig.8),wemayfurther closer to the universal RSF than the RSFs deduced from re- investigatethetransmissioncoefficientindetail,andthusthe gionsoflowleveldensity. validityoftheBrinkhypothesis. Wefirstadoptthesolutions T andρ fromSec.IIIandwrite Since there exists only one unique level density, we con- structthecounterparttoEq.(12)inthecasethatthetransmis- N (E)P(E,E )≈T(E )ρ(E −E ). (12) sioncoefficientdependsontheinitialexcitationenergy: i i γ γ i γ Thenormalizationfactorforeachinitialexcitationbinisde- N (cid:48)(E)P(E,E )≈T(E ,E)ρ(E −E ), (14) i i γ γ i i γ 7 -3)eV10-8 FPi/tr r · T Ei = 5.5-7.0 MeV M 104 SF ( R 5.5-7.0 MeV 10-9 7.0-8.5 MeV 1) x27 8.5-10.0 MeV 0 2 4 6E = 7.0-88.5 MeV10 -MeV103 x9 5.5-10.0 MeV -3)eV10-8 i ( M ry F ( sit x3 RS n102 e 10-9 d x1 el 0 2 4 6E = 8.5-810.0 Me1V0 ev 3) i L 10 -MeV10-8 F ( S R 10-9 1 0 2 4 6 8 10 E = 5.5-10.0 MeV 3) i 0 2 4 6 8 10 -eV10-8 Excitation energy E (MeV) M F ( S R FIG. 8: Level density extracted from statistically independent data 10-9 sets, taken from various initial excitation energy bins Ei (the three 0 2 4 6 8 10 upper curves). The lower curve is the result for the whole energy Eg (MeV) regionandisidenticaltotheoneofFig.5. FIG.9:(Coloronline)Gamma-raystrengthfunctionsextractedfrom various initial excitation bins Ei. The data points of the three upper deduced RSFs are from statistically independent data sets. whereN (cid:48) isdeterminedanalogouslytoEq.(13). Weexpect The data points in the lower panel are identical with the RSF of thatT(E,E )fluctuatesontheaveragearoundT(E ).Thus, Fig.7. TheRSFsdisplayedasredlines,areevaluatedfromthera- i γ γ itisreasonabletoexpectthatN (cid:48)≈N ,whichgivesatrans- tio P(Ei,Eγ)/ρ(Ef) (see text). The resemblance between the two methods confirms that the level density is common for the various missioncoefficientof: excitationregions,inaccordancewithFermi’sgoldenrule. P(E,E ) T(E,E )≈N (E) i γ . (15) i γ i ρ(E −E ) i γ Similarly,thetransmissioncoefficientasafunctionofthefinal (15). Sincetheapproximation(16)isasimpletransformation excitationenergyE =E −E isgivenby f i γ usingE =E −E ,wemayalsoinvestigatethedependencies f i γ oftheRSFonfinalexcitationenergyE . P(E +E ,E ) f T(Ef,Eγ)≈N (Ef+Eγ) fρ(E γ) γ . (16) In Figs. 10 and 11, the RSFs f(Ei,Eγ) and f(Ef,Eγ) are f shown for eight excitation regions. For all these deduced Thevalidityoftheapproximations(14)and(15)isdemon- RSFs, we use one common level density in the evaluation strated in Fig. 9 by comparing the deduced RSF from this of the approximations (15) and (16). The data points of the approximation (lines) with the independent fits of T and ρ experimental P matrix cover only a certain region in Ei and (datapoints). TheRSFsdeterminedforthewholeenergyre- Eγ,makingrestrictionsonthededucedRSFs. Thus, f(Ei,Eγ) gion 5.5−10 MeV are very similar, especially for low Eγ, is limited to 5.5 MeV <Ei <Sn and 1.8 MeV <Eγ <Ei. as shown in the lower panels. Also the similarity in the de- The limits for f(Ef,Eγ) are 0 < Ef < Sn−1.8 MeV and tailed structures of the two methods is recognized, although 1.8MeV<Eγ <Sn−Ef. Foraconsistencycheck, wehave some differences are present. The deviations are largest for testedthattheaverageRSFforallEf energiesequalstheone thehighγ-energypartoftheRSFspopulatingthelowest0+, forallEienergies(notshownhere). 2+, and 4+ states, where large Porter-Thomas fluctuations Figure 10 shows that the deduced RSF changes as a func- are expected (these fluctuations are not included in the error tionofE,whenlowexcitationenergyispopulatedaftertheγ- i bars). However,theoverallgoodresemblanceencouragesus emission. Thelower,leftpanel,wheretwoconsecutiveRSFs to study the detailed evolution of the RSFs as a function of are plotted together, illustrates this. Here, the bumps seen at initial excitation energies using the approximations (14) and high γ-ray energies are the artifacts of the decay to specific 8 E = 5.4-5.9 MeV E = 6.0-6.5 MeV i i 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 E = 6.6-7.1 MeEgV (MeV) E = 7.2-7.7 MeEgV (MeV) i i 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 -3)V Ei = 7.8-8.3 MeEgV (MeV) Ei = 8.4-8.9 MeEgV (MeV) e F (M 10-8 10-8 S R 0 2 4 6 8 10 0 2 4 6 8 10 Ei = 9.0-9.4 MEge (VMeV) Ei = 9.6-10.0 MEge (MVeV) 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 Two last RSFsEg (MeV) All RSFs Eg (MeV) 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 Eg (MeV) Eg (MeV) FIG. 10: (Color online) Deduced RSFs from various initial excitation bins Ei. The RSFs are evaluated from the ratio P(Ei,Eγ)/ρ(Ef) as describedinthetext. isolated levels below 3 MeV of excitation. This is also the andthechangesbetweenthevariousdeducedRSFsarelarge reason why apparently the fluctuations are typically a factor duetoPorter-Thomasfluctuations. of 2−3 in the panel where all deduced RSFs are plotted to- InordertoshowthesimilarityofthededucedRSFsinthe gether (lower, right panel). In the plot of all deduced RSFs, case of strongly suppressed Porter-Thomas fluctuations, we we see a minimum at about 4−6 MeV and some interest- havecomparedthetwouppermostexcitationenergygatesE i ingstructuresatlowγ-rayenergies. Thesestructuresandthe and E in the lower, left panels of Figs. 10 and 11, respec- f minimumareindependentoftheuncertaintyinthelog-slope tively. The deduced RSFs are here extracted for γ decay be- oftheleveldensity,asshowninFig.7. tween states in quasi-continuum, except for the data points Thevarious f(Ef,Eγ)plotsinFig.11aredifficulttocom- with Eγ >7 MeV in Fig. 10 where the final excitation en- pare due to different limits appearing at both low and high ergyis<3MeV.ThededucedRSFsextractedfromthequasi- γ-energies for the various E regions. For example, the first continuumregionbehavesimilartothededucedRSFfromthe f three spectra do not reveal the region of low γ-energy en- all-overfitdisplayedinFig.7. hancement because Eγ > 5.5MeV−Ef > 3.7 MeV. These The good agreement between the deduced RSFs at higher spectrarepresentthedecaytothe0+,2+and4+groundband energies is consistent with the expectation of suppressed states,respectively,wheretheexperimentalenergyresolution Porter-Thomas fluctuations due to more initial and final lev- makessomeoverlapbetweenthesestates. Ingeneralthevar- elsintheevaluationoftheγ strength. Theseresultsstrongly ious spectra show strong fluctuations in the deduced RSFs indicatethattheconceptofanRSF,whichisindependenton when the γ emission ends up at low excitation energy, typi- excitationenergy,isvalidalreadyatrelativelylowexcitation cally Ef <3 MeV. The panel with all deduced RSFs plotted energies in 46Ti. At the same time, the results give us con- together (lower, right panel) shows approximately the same fidence that Oslo method works reasonably. The differences scatteringofdatapointsasinFig.10. indeducedRSFsfromlowerenergiesindicatethatthePorter- It is thus clear that the experimentally deduced RSFs for Thomasfluctuationsaresopoorlysuppressedthatitisdifficult which states below 3 MeV of excitation energy are involved topredicttheshapeoftheuniversalRSF.Thisisnotverysur- in the γ decay are very different. Figure 5 shows that this prising and the differences in the deduced RSFs seem to be excitationregioncoincideswitharegionoffewandwellsep- consistentwiththeexpectedfluctuations. aratedlow-lyinglevelswithspecificstructures. Thebumpsin In order to display the small differences in the deduced thededucedRSFsarespecifictothelowenergylevelscheme, RSFs obtained from regions ofhigher level density, we have 9 E = 0.0-0.6 MeV E = 0.6-1.2 MeV f f 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 E = 1.2-1.8 MEeg V(MeV) E = 1.8-2.4 MeEgV (MeV) f f 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 -3)V Ef = 2.4-3.0 MEeg V(MeV) Ef = 3.0-3.6 MeEgV (MeV) e F (M 10-8 10-8 S R 0 2 4 6 8 10 0 2 4 6 8 10 Ef = 3.6-4.2 MEge (VMeV) Ef = 4.2-4.8 MeEgV (MeV) 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 Two last RSFsEg (MeV) All RSFs Eg (MeV) 10-8 10-8 0 2 4 6 8 10 0 2 4 6 8 10 Eg (MeV) Eg (MeV) FIG. 11: (Color online) Deduced RSFs populating various final excitation bins E . The RSFs are evaluated from the ratio P(E + f f Eγ,Eγ)/ρ(Ef) as described in the text. It should be noted that there are no final states at Ef =1.2−1.8 MeV, however the experimental resolution(seeFig.5)isresponsibleforincludingthe2+and4+groundbandstatesinthisgate. taken the four highest gates shown in Fig. 10 and only con- difficulttodeterminesincetheexperimentalstatisticalerrors sideredγ energiesupto5.1MeV.Thus,thesestatisticallyin- (seeerrorbars)arealsoofthesameorder. dependentRSFsareevaluatedinquasi-continuumwithinitial andfinalexcitationenergiesofroughlyE =8−10MeVand i E =3−7 MeV, respectively. The deduced RSFs are pre- f V. CONCLUSIONS sentedintheupperpanelofFig.12. Forthisdatasetweevaluatedtherelativedeviationsby The level density and radiative strength function for 46Ti fij−(cid:104)fi(cid:105) have been determined using the Oslo method. Similar level r = , (17) ij (cid:104)f(cid:105) densityfunctionshavebeenextractedfromstatisticallyinde- i pendentdatasetscoveringdifferentexcitationenergies. This where the index i represents the γ energy and j is the initial gives confidence to the Oslo method, since the disentangle- excitation energy. The average strength at each γ energy is mentoftheleveldensitybyFermi’sgoldenrulepredictsone estimatedfromthefourindividualRFSs andonlyoneuniqueleveldensity,independentofthedataset. 1 ThededucedRSFdisplaysanenhancementatlowγ-rayen- (cid:104)f(cid:105)= ∑f . (18) i ij n ergywhereweseeabumparound3MeVandanotherstruc- j j tureatenergiesnear2MeV.Asimilarenhancementhasbeen InthelowerpanelofFig.12therijvaluesareplottedshowing seen in several other light mass nuclei and is still not ac- therelativefluctuationsfromthemeanvalueateachγ energy. countedforbypresenttheories. The average ratio for all data points (ni =28 and nj =4) is AmethodtostudytheevolutionofthededucedRSFsasa takenas functionofinitialandfinalenergyregionshasbeendescribed. 1 The deduced RSFs are found to display strong variations for r= ∑|r |, (19) nn ij different initial and final excitation energies if transitions to i j ij thelowestexcitationsareinvolved. Thereasonfortheviolent givingr∼6%. ThedifferencesinthededucedRSFscaneas- fluctuationsofafactorof2−3isthatonlyafewisolatedlev- ily be interpreted as remnants of the Porter-Thomas fluctua- els are present at low excitation energies with E <3 MeV. f tions. However, a quantitative estimate of the fluctuations is The differences in the RSFs obtained from a few number of 10 · 10-9 transitions,thatcanbeexplainedasaconsequenceofPorter- 16 Thomas fluctuations of individual intensities, show that this E = 7.8-8.3 MeV Ei = 8.4-8.9 MeV (a) energyregioncannotbeusedfordeterminationoftheuniver- 14 i E = 9.0-9.4 MeV salRSF.Eventhough,thededucedRSFsbasedonarestricted i E = 9.6-10.0 MeV 3) 12 i numberoftransitionsstillindicatethatthedecayisgoverned -V byauniversalRSF. e 10 M 9 -0 8 1 ( F 6 However, the present work shows that it is possible to get S R 4 morepreciseexperimentalinformationontheuniversalRSF. By imposing restrictions on the initial and final excitation 2 energies, the RSFs for the decay between states in quasi- 0 continuumcanbeextracted(i.e.forEf >∼3MeV).Theresults from this selected data set show that the decay is consistent ) 20 Eg (MeV) (b) withanRSFwhichisidependentofexcitationenergywithin % ( lessthan∼6%alreadyattheserelativelylowexcitations. F S 10 R n e i 0 Insummary, theobservationofalmostidenticallevelden- c n e -10 sities and RSFs extracted from different data sets in quasi- r e continuum, indicates that the Oslo method works well. Pro- f Dif -20 videdthatweusedatafromthequasi-continuum,auniversal RSF in the light mass region of the 46Ti nucleus can be ex- tracted. 2 2.5 3 3.5 4 4.5 5 Acknowledgments 2 2.5 3 E (3M.5eV) 4 4.5 5 g FIG.12: (Coloronline)(a)Radiativestrengthfunctionsforγ transi- tionsbetweenstatesinquasi-continuum. Datafromthefourhighest excitationenergygatesofFigs.10havebeenchosen. (b)Ratiosof TheauthorswishtothankE.A.OlsenandJ.Wikneforex- thedeviationfromtheaverageRSFateachγenergy,seetext. cellentexperimentalconditions. Thisworkwassupportedby theResearchCouncilofNorway(NFR). [1] P.A.M.Dirac,Proc.R.Soc.LondonA114,243-265(1927). Phys.Rev.C79,024316(2009). [2] E. Fermi, Nuclear Physics (University of Chicago Press, [13] A. Gilbert and A. G. W. Cameron, Can. J. Phys. 43, 1446 Chicago,1950). (1965). [3] DataextractedusingtheNNDCOn-LineDataServicefromthe [14] T. von Egidy and D. Bucurescu, Phys. Rev. C 72, 044311 ENSDFdatabase.[http://www.nndcnbnl.gov./ensdf/] (2005);Phys.Rev.C73,049901(E)(2006). [4] S.S.DietrichandB.L.Berman,At.DataNucl.DataTables38, [15] S. Goriely, S. Hilaire, and A.J. Koning,, Phys. Rev. C 78, 199(1988). 064307(2008). [5] A.Schiller,L.Bergholt,M.Guttormsen,E.Melby,J.Rekstad, [16] Kristine Wikan, Master thesis in Physics, De- S.Siem,Nucl.Instrum.MethodsPhys.Res.A447,498(2000). partment of Physics, University of Oslo, 2010, [6] M.Guttormsen,A.Atac,G.Løvhøiden,S.Messelt,T.Ramsøy, http://www.duo.uio.no/publ/fysikk/2010/105030/Master.pdf J.Rekstad,T.F.Thorsteinsen,T.S.Tveter,andZ.Zelazny,Phys. [17] N.U.H.Syed,A.C.Larsen,A.u¨rger,M.Guttormsen,S.Haris- Scr.T32,54(1990). sopulos, M. Kmiecik, T. Konstantinopoulos, M. Krtic˘ka, A. [7] M.Guttormsen, T.S.Tveter, L.Bergholt, F.Ingebretsen, and Lagoyannis, T.Lo¨nnroth, K.Mazurek4, M.Norby, H.T.Ny- J. Rekstad, Nucl. Instrum. Methods Phys. Res. A 374, 371 hus, G.Perdikakis, S.Siem, andA.Spyrou, Phys.Rev.C80, (1996). 044309(2009). [8] M. Guttormsen, T. Ramsøy, and J. Rekstad, Nucl. Instrum. [18] Merico Salas-Bacci, Steven M. Grimes, Thomas N. Massey, MethodsPhys.Res.A255,518(1987). Yannis Parpottas, Raymond T. Wheller, and James E. Olden- [9] D.M.Brink,Ph.D.thesis,OxfordUniversity,1955. dick,Phys.Rev.C70,024311(2004). [10] A. Schiller and M. Thoennessen, At. Data Nucl. Data Tables [19] J. Bardeen, L.N. Cooper, and J.R. Schrieffer, Phys. Rev. 108, 93,549(2007). 1175(1957). [11] C.E.PorterandR.G.Thomas,Phys.Rev.104,483(1956). [20] J.KopeckyandM.Uhl,Phys.Rev.C411941(1990). [12] N.U.H. Syed, M. Guttormsen, F. Ingebretsen, A. C. Larsen, [21] A.Voinov, M.Guttormsen, E.Melby, J.Rekstad, A.Schiller, T. Lo¨nnroth, J. Rekstad, A. Schiller, S. Siem, and A. Voinov, andS.Siem,Phys.Rev.C63,044313(2001).

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.