PHYSICALREVIEWC,VOLUME61,034604 Photofission of actinide nuclei in the quasideuteron and lower part of the D energy region J. C. Sanabria,* B. L. Berman, C. Cetina, P. L. Cole,† and G. Feldman CenterforNuclearStudies,DepartmentofPhysics,TheGeorgeWashingtonUniversity,Washington,D.C.20052 N. R. Kolb, R. E. Pywell, and J. M. Vogt‡ SaskatchewanAcceleratorLaboratory,Saskatoon,Saskatchewan,CanadaS7N0W0 V. G. Nedorezov and A. S. Sudov InstituteforNuclearResearch,RussianAcademyofSciences,Moscow117312,Russia G. Ya. Kezerashvili§ BudkerInstituteofNuclearPhysics,Novosibirsk630090,Russia ~Received28May1999;revisedmanuscriptreceived25October1999;published11February2000! The total photofission cross sections for the actinide nuclei 232Th, 233U, 235U, 238U, and 237Np have been measuredfrom68to264MeVusingtaggedphotonsattheSaskatchewanAcceleratorLaboratory.Thefission fragmentsweredetectedusingparallel-plateavalanchedetectors.Theresultsshowthatthefissionprobability for 238Uis20%lowerthanthatfor 237Npand40%higherthanthatfor 232Th.Lesssignificantdifferenceswere also found among the individual uranium isotopes. These results contradict the assumption that the fission probability for 238U is approximately equal to unity in this energy range. It has also been observed that the fissionprobabilityasafunctionofenergyforalltheseisotopesisconstant,withtheexceptionofthatfor 232Th, whichincreaseswithenergy,althoughitseemstobereachingasaturationvalue.Comparisonbetweenthetotal photofissioncrosssectionfor 237Npandthephotoabsorptioncrosssectionsforlighternucleishowsabehavior consistentwithabroadeningoftheD resonancewithincreasingatomicmass. PACSnumber~s!: 25.20.2x,25.85.Jg,27.90.1b I.INTRODUCTION in part based on the assumption that the photofission prob- abilityofuraniumisclosetounity@4#.Eventhemostrecent A.Motivation results using monochromatic photons @5–7# show some dis- The total photofission cross section, in the case of the crepancies in the cross sections per nucleon for 235U and heavy actinides, has been thought to be a good approxima- 238U and between these isotopes and the so-called ‘‘univer- tion to the total photoabsorption cross section at photon en- sal curve’’ in the D-resonance region. ergies well above the giant dipole resonance region. This The most important discrepancy reported previously ap- allows one to study the effect of the nuclear medium on pears in the results of a measurement of the relative photo- processes such as baryon resonance formation and propaga- fission probability of 237Np compared with 238U from 60 tion within the interior of the nucleus. Specifically, for the MeV to 240 MeV photon energy @8#. In this measurement caseof 238U,experimentalmeasurementsandtheoreticalcal- the photofission probability of 237Np appears to be between culations have suggested that the photofission probability is 20% and 30% larger than that of 238U, so that the photofis- consistent with unity for photon energies larger than about sion probability for the latter isotope could be at most 0.8. 40 MeV @1,2#. Comparison of the total photofission cross Thisresulthasseriousimplicationsfortheinferredtotalpho- section per nucleon for the uranium nuclei with the total toabsorption cross-section strengths in the D-resonance re- photoabsorptioncrosssectionpernucleonfornucleifromBe gion, and needs to be verified. to Pb in the D-resonance region ~from approximately 200 Thelackofdirectmeasurementsofthetotalphotoabsorp- MeVto450MeVphotonenergy!showsasimilarshapeand tion cross sections for actinide isotopes, together with the strengthforthesecrosssections,indicatingthatthephotoab- discrepancies mentioned above, makes it very important to sorptionprocesscanbedescribedbyanincoherenttotalvol- measure very accurately the absolute and relative photofis- ume absorption mechanism @3#. However, this conclusion is sion cross section for 237Np and several uranium isotopes. The availability of high-duty-cycle electron accelerators liketheonesatMainz,SaskatchewanAcceleratorLaboratory *Present address: Departamento de F´ısica, Universidad de los ~SAL!,andJeffersonLaboratory,togetherwiththeirphoton- Andes,A.A.4976,Bogota´,Colombia. tagging facilities, allows one to measure the photofission †Presentaddress:DepartmentofPhysics,UniversityofTexas,El crosssectionsforactinidenucleiveryaccuratelyoverawide Paso,TX79968. range of photon energies. A simultaneous measurement of ‡Presentaddress:HypresInc.,Elmsford,NY10523. these cross sections under identical conditions determines §Deceased. their relative fissilities. 0556-2813/2000/61~3!/034604~14!/$15.00 61034604-1 ©2000TheAmericanPhysicalSociety J.C.SANABRIAetal. PHYSICALREVIEWC61034604 FIG. 2. Ratio of the 237Np to the 238U photofission cross sec- tions@8#. bedrawnduetothelargeerrorbarsofthetotalphotoneutron cross sections.1 In the D-resonance region ~from about 150 MeV to 500 FIG.1. Photofissioncrosssectionfor 235U~solidtriangles!,and MeV!, the total photofission cross sections for the actinides 238U~solidsquares!,comparedwiththe‘‘total’’photoneutroncross can be compared with the existing data on photoabsorption sectionfor 238U~opencircles!@1#. for nuclei from Li to Pb. These latter data have been ob- tained using a variety of experimental techniques, including thephotohadronictechnique,thetransmissiontechnique,and The accurate measurement of the total photofission cross themeasurementofneutronmultiplicities@14–17#.Thepho- sections for 232Th, 233U, 235U, 238U, and 237Np from 68 to tohadronictechniqueconsistsinmeasuringthephotoproduc- 264 MeV at SAL has provided us with explanations for the tion rate of events at large angles ~hadronic events!, while discrepancies mentioned above. rejecting the events whose products are emitted in the for- ward direction ~electromagnetic events!. The transmission B.Previousmeasurements techniqueconsistsinmeasuringthephotonbeamattenuation A careful and comprehensive inspection of the existing cross section and then subtracting the ~calculated! atomic data on photoabsorption and photofission for heavy nuclei is absorption cross section. The neutron-multiplicity technique necessary in order to determine the level of certainty in the consists in measuring the individual cross sections for the knowledgeofthefissionprobabilityofactinidenucleiandof evaporation of multiple neutrons and then adding them to the ‘‘universal’’ behavior of the photoabsorption cross sec- give the total cross section. tionpernucleon.Thephotonenergyspectrumcanbedivided The existing data on photofission of 235U and 238U come intoseveralregionsbasedonthedifferentabsorptionmecha- from tagged-photon experiments at Bonn, Frascati, and nisms: the giant dipole resonance, the quasideuteron, the D Mainz@5–7#.Frominspectionofallthesedata,onecancon- resonance, and the higher resonances. clude that the total photofission cross sections per nucleon Inthegiantdipoleresonanceregionthetotalphotofission for 238U and 235U are close to each other and to the total crosssectionsforseveralactinidenucleihavebeenmeasured photoabsorption cross section per nucleon for the light iso- very accurately using monoenergetic photons from the anni- topes, and they all follow the so-called ‘‘universal curve,’’ hilation in flight of fast positrons @9–12#. The fission prob- which implies that the total photabsorption cross section is ability measured in these experiments in the energy region proportional to the number of nucleons inside the nucleus near 10 MeV varies from 0.1 for the case of 232Th up to and hence to the nuclear volume. However, one can also about 0.6 for the case of 237Np. notice that the uncertainty in the shape of the ‘‘universal Inthequasideuteronregion~fromapproximately40MeV curve’’ is of the order of at least 10% ~not including the photon energy up to the pion threshold!, the quality of the systematic uncertainties of all the measurements!. This un- dataisnotasgoodasinthegiantdipoleresonanceregion.In certainty has taken on increased importance in view of the Fig.1thetotalphotofissioncrosssectionsfor 235Uand 238U results obtained at Novosibirsk on the photofission probabil- @1#arecomparedwiththeonlydataontotalphotoabsorption for 238U available in this region ~based on the measurement of the sum of the multiple @.1# photoneutron cross sec- 1Especially when one adds the (g,1n), (g,p), and possibly even tions! @13#. As one can see from this figure, no precise con- (g,a) cross sections not included in the s(g,xn) results of Ref. clusions concerning the value of the fission probabilities can @13#. 034604-2 PHOTOFISSIONOFACTINIDENUCLEIINTHE... PHYSICALREVIEWC61034604 ity of 237Np relative to that of 238U in the energy range C.Theory between50and240MeV@8#.Therelativefissionprobability The photofission of heavy nuclei, for photon energies of 237Np seems to be at least 20% higher than that of 238U, higher than about 40 MeV, is described as a two-stage pro- as shown in Fig. 2. This result thus contradicts the funda- cess. In the first stage ~the fast stage3!, the incident photon, mentalassumptionthatthefissionprobabilityforuraniumat with energy v, initiates an intranuclear cascade in which theseenergiesisapproximatelyequaltounityandthatthere- someoftheparticlesinvolvedescape,leavinganewresidual fore the photofission cross sections for these isotopes are nucleus in an excited state, the so-called compound nucleus, equivalent to the total photoabsorption cross sections. which is characterized by the property that the excitation Another interesting result in the photofission of actinide energyEx isdistributedoverallpossibledegreesoffreedom, nucleiistheverydifferentbehaviorof 232Thfromthatofthe such that thermodynamic equilibrium is established. The heavier actinides at intermediate to high photon energies. timescaleoftheformationofthecompoundnucleusisofthe Data below 100 MeV show that the fission probability ~rela- order of 10219 s. In a second stage ~the slow stage!, the tiveto 238U) increasessmoothlyfrom0.2to0.6withphoton compound nucleus disposes of the excitation energy by energy @2#. Data at higher energies ~from 250 to 1200 MeV! gamma-ray emission, particle emission, or fission. The indicatethatthefissionprobability~relativeto 238U),evenat gamma-ray emission is important only when the excitation energyisbelowthethresholdforparticleemissionorfission. quite large photon energies, does not exceed approximately Therefore, for highly excited heavy nuclei, the deexcitation 0.8andappearstosaturate@18#.Thishigh-energybehavioris processreducestoacompetitionbetweenfissionandparticle noteasytounderstand;inspiteofthescatterofthedata,the emission. The emitted particles are mostly neutrons, al- average fissility is substantially lower than 1. This might though emission of protons, deuterons, trinucleons, and a show a strong dependence of the fission probability for ac- particlesisalsopossible,butfarlessprobablebecauseofthe tinides on atomic number and mass, specifically on the fis- Coulomb barrier. The time scale of the deexcitation stage is sility parameter Z2/A of the target nucleus.2 Low-energy re- of the order of 10216 s. sults also show that there is a transition to very low fission The large difference between the time scales of the for- probability between thorium and radium @19#. Since the mationofthecompoundnucleusanditsdeexcitationjustifies number of neutrons and protons escaping the nucleus during the description of the photofission process as a two-stage theintranuclearcascadeincreaseswithincreasingphotonen- process. Based on this, one can write the cross section as ergy,andbecausefissionwilloccurwithincreasedprobabil- @21# ity after several neutrons have been evaporated, an extrapo- E lpartoiobnabiflirtoymoflotwhorieunmergisiesdiriesctlnyotreljautsetdifietod.thTohsee ffiosrsitohne sg,f~v!5( ( vsCN~AC,ZC;Ex!wf~AC,ZC;Ex!dvEx, uraniumandtransuranicisotopes.However,therearenodata AC ZC 0 ~1! in the energy region between about 100 and 250 MeV; it is very important to bridge this gap, and to verify the absolute wheres isthecrosssectionfortheformationofthecom- cross sections on both sides. CN pound nucleus with A nucleons, Z protons, and excitation At photon energies appreciably higher than the peak of C C energy E , and w is the probability that the compound the D resonance ~above 400 MeV!, the most recent data on x f nucleus will fission in the transition to the ground state. photofission and photoabsorption, from experiments at Fras- In Eq. ~1!, s summarizes the result of the fast stage of cati @6,18,17,20# and Mainz @4,5#, show no evidence of ex- CN thefissionprocess.ThesecondterminsidetheintegralinEq. citation of the baryon resonances D (1520) and F (1680), 13 15 ~1!,w , describesthecompetitionbetweenparticleemission which are clearly seen in the photon absorption cross sec- f and fission during the slow stage of the fission process. This tions for 1H and 2H as peaks at energies of ;0.7 and quantity can be written as ;1.0 GeV.Thisunexpectedbehaviorofthesenucleonreso- nances in nuclei is still an open theoretical and experimental ( problem.Itisimportanttohaveathoroughunderstandingof w 5 P ~A ,Z ;E !wf~A ,Z ;E !, ~2! f k C C x k C C x the relation between the total photoabsorption cross section k andthephotofissioncrosssectionforactinidenucleiatlower energies in order to understand these results. where Pk is the probability that the compound nucleus will In summary, the status of the previous data on photoab- emitkparticlesinthetransitiontothegroundstate,andwf is k sorption and photofission for the actinides shows the neces- the probability that fission will occur in one of the k evapo- sityforasimultaneousmeasurementofboththeabsoluteand rative steps. These two quantities can be expressed in terms the relative photofission cross sections for 235U, 238U, of the emission widths G for the set $j% of all possible emit- j 237Np, and 232Th below and in the D-resonance energy re- ted particles (j5n,p,d,t,3He,a) and the fission width G f gion. @21#. Therefore, based on this model, the determination of the fission probability of the compound nucleus reduces to the calculation of the partial decay widths G and G . j f 2The fissility parameter Z2/A results from the competition be- tweentheCoulombrepulsionandthesurfacetensionofthenucleus duringthefissionprocess. 3Thetimescaleofthedirectreactionisoftheorderof10223 s. 034604-3 J.C.SANABRIAetal. PHYSICALREVIEWC61034604 For high-Z nuclei, neutron emission is much more prob- II.EXPERIMENT ablethanemissionofchargedparticles,andthusw depends f The goal of the present experiment was to measure the strongly on the ratio G /G . An expression for this ratio is n f total photofission cross sections for the actinides 237Np, given in Ref. @19#: 238U, 235U, 233U,and 232Thwith3%statisticalprecisionand less than 5% systematic uncertainty. The experiment was Gn5 4A2/3af~E2Bn! exp@2a1/2~E2B !1/2 performed at the Saskatchewan Accelerator Laboratory, us- Gf K0an@2a1f/2~E2Ef!1/221# n n itniogneitds pishoottoopne-stagwgeinreg fiarrcaidliitayt.edThweittahrgmetsonoofcthhreomafaotricemreena-l 22a1/2~E2E !1/2#, ~3! photons, and for each photofission event one of the two re- f f sultingfissionfragmentswasdetected.Bycountingthenum- ber of events that resulted in fission, and dividing that quan- where B is the neutron binding energy, E is the fission n f titybythenumberofincidentphotons,thetotalphotofission barrier, a is a nuclear level density parameter, and K 0 514.39 MeV. The parameters a and a describe the cross sections were determined. change in level density in the saddfle point ~nlarge deforma- AtSALanelectronbeamincidentonanaluminumradia- tion! and in the ground state ~smaller deformation! @19#. torgeneratesabremsstrahlungphotonbeam.Aftertheradia- As one can see from Eq. ~3!, G /G has a strong depen- tor, the electron beam is deflected away from the beam line n f by a dipole magnet, and the energy of each individual elec- dence on a and a ; thus, an accurate knowledge of these n f tron is measured by an array of scintillators located in the parameters for several isotopes is necessary for the descrip- focalplaneofthedeflectingmagnet~thephotontagger!.The tion of the evaporative deexcitation and fission of a single photons continue downstream and traverse a target foil. target isotope. Unfortunately these are not very well known, When a photofission event occurs, two fission fragments are eithertheoreticallyorexperimentally,andonehastorelyon emitted in opposite directions ~the linear momentum trans- semiempiricalparametrizationsoftheirdependenceonAand ferred by a photon to a heavy nucleus is negligible in this Z. energy range; therefore, the center-of-mass and laboratory An expression for a proposed by Iljinov etal. includes corrections due to excitnation energy and shell effects @21#: frames are nearly the same!. If one of the emitted fission fragments goes through the fission-fragment detector, the event is registered. The energy of the photon is determined a 5~0.134A21.2131024A2! n F G from the difference in energy of the incident electron beam DM andtheelectrondetectedinthefocalplaneofthetagger.The 3 11~12e20.061Ex! MeV21, ~4! electronic coincidence between the signal coming from the E x fission-fragmentdetectorandoneofthephoton-taggerchan- nels determines the experimental trigger. where DM is the shell correction to the nuclear mass. As long as the target foils and the fission-fragment detec- As for af, the parametrizations usually are expressed in torsarenearlytransparenttothephotonbeam,onecanplace terms of r5af/an. By fitting Eq. ~3! to all the existing data several target-detector pairs along the beam line, in order to onGn/Gf forisotopesrangingfromSmuptoU,andusinga measure simultaneously the photofission cross sections for linear dependence of r on the fissility parameter Z2/A, Mar- different isotopes and/or to increase the statistics of the ex- tins etal. proposed the following expressions @22#: periment by having several targets per isotope. r5110.05917~Z2/A234.34!, Z2/A.34.90, ~5! A.SALelectronacceleratorandphotontagger TheelectronacceleratoratSALconsistsofalinearaccel- r5110.08334~Z2/A230.30!, 31.20,Z2/A<34.90, erator ~Linac! and a pulse stretcher ring ~PSR!. The PSR is ~6! designed to convert the pulsed electron beam of the Linac into an almost continuous beam. The energy range of the r51.28120.01842~Z2/A220.00!, 24.90<Z2/A<31.20. Linacgoesfrom50MeVuptoabout300MeV,withapeak ~7! current of 200 mA. The PSR stretches the pulsed beam into acontinuousbeam,increasingitsdutyfactortonearly100%. The complex dependence of G /G on A, Z, and E The range of energies of the PSR is the same as the Linac; n f x makes the results of these calculations very sensitive to the however, its maximum current is limited to 70 mA @24#. distributions DA(v)5A2A , DZ(v)5Z2Z , and E (v) In this experiment, the electron accelerator was operated C C x predicted by the intranuclear-cascade Monte Carlo models. at an energy of 287.7 MeV, so that the maximum photon While the distributions for DA and DZ are relatively sharp, energy coverage was achieved. During these measurements, the distribution for E becomes very broad as the photon themachinewasoperatedatarelativelylowcurrent,tolimit x energy increases, and its mean value tends to saturate for the number of accidental events coming from fission events photon energies above 120 MeV @8,23#. This, plus the in- induced by untagged low-energy photons @the photofission crease of the number of positive charges emitted by the cross section for actinide nuclei peaks in the giant dipole nucleus, could explain the saturation of the photofission resonance ~GDR! region, where the bremsstrahlung yield is probability above the giant dipole resonance region. also very high#. 034604-4 PHOTOFISSIONOFACTINIDENUCLEIINTHE... PHYSICALREVIEWC61034604 ground when the targets are actinide nuclei, they provide goodpulse-heightdiscriminationbetweentheaparticlesand the fission fragments. PPADs were also ideal for this experiment because they can be quite transparent to an intense high-energy photon beam, allowing the location of a sizable array of target- detector pairs along the beam line. This is very important in order to increase the statistics of the measurements, a com- mon problem for photofission experiments. It should also be pointedoutthatPPADshavealowproductioncost,ascom- pared, for example, with solid-state detectors, allowing us to build many of them. For this experiment, PPADs with anode wire grids were designed, built, and tested at The George Washington Uni- FIG. 3. Schematic diagram of the SAL photon tagger ~not to versity Nuclear Detector Laboratory. The use of a wire grid scale!. for the detector anode instead of the usual thin foil of con- TheSALphoton-taggingfacilityconsistsofaradiator,an ducting material resulted in a higher detection efficiency be- electronspectrometer,afocal-planedetectorarray,andacol- cause the fragments emitted from the target ~located 17.5 limator. In Fig. 3 a schematic view of the system is pre- mm away! were able to enter the active region of the PPAD sented. The focal-plane detector array consists of two rows without encountering any material ~with the exception of of plastic scintillators. The first row has 31 scintillators, and those fragments that collided with the wires!. Two collima- the second one has 32. The two rows are shifted so that the tors,onelocatedrightinfrontofthetargetandtheotherone scintillators in the first row overlap the scintillators in the located right in front of the detector, prevented fission frag- second one by 50% @25#. ments emitted at wide angles from being detected; in this Since the SAL photon tagger can tag photons only up to way the distance traveled by the detected fragments inside 223 MeV, a second spectrometer ~‘‘the end-point tagger’’!, the targets was minimized and the probability of reabsorp- located1.8mupstreamfromthefirstone,canbeusedtotag tion by the target material became negligible. At the same photonsupto264MeV,usingthesamefocal-planedetector time the collimators prevented a large fraction of the apar- array as the regular spectrometer ~after being relocated!. ticles emitted by the radioactive targets from reaching the In order to limit the spot size of the photon beam at the detectors, reducing significantly the main source of back- target,aleadcollimatorislocated2mdownstreamfromthe ground in the experiment. radiatoroftheregularspectrometer.Thediameterofthecol- An array of target-detector pairs was placed inside a re- limatorwaseither10or15mm,anditslengthwas130mm. action chamber. The thin aluminum windows of the reaction The reaction chamber for the fission-fragment detectors and chamberallowedthephotonbeamtopassthroughwithvery targets was located 2.5 m downstream from the collimator. littleattenuation.AtargetwasplacedinfrontofeachPPAD. Only a fraction of the photons being radiated reaches the The targets were aluminum foils nominally 100 mm thick, reaction chamber because of the collimator. This fraction is with a film of fissionable material deposited on one side. A measured during special runs ~tagging efficiency runs!, in detailed description of the detectors and experimental setup which a total-absorption lead-glass detector, located at the is presented elsewhere @34,35#. end of the photon beam line, is used to count the number of taggedphotonsinthecollimatedbeam.Theratioofthenum- C.Electronicreadoutanddataacquisition beroftaggedphotonscountedbythelead-glassdetectorand A schematic diagram of the experimental electronics is thenumberofelectronsdetectedbythechannelsofthefocal shown in Fig. 4. First, the signal from a PPAD is amplified plane detector determines the fraction of photons that pass and sent as input to a linear fanout module that splits it into throughthecollimator.Thesetaggingefficiencyrunsareper- two identical output signals. One of the output signals is formedregularlytoaccountforpossiblechangesintheelec- delayed and used as input for the analog-to-digital converter tron beam profile and location. During these runs the fission ~ADC!. detectors are removed from the photon beam line. TheexperimentaltriggerisdefinedbyalogicalORamong all the outputs of the constant-fraction discriminators B.Fission-fragmentdetectorsystem ~CFDs!.Inthisway,aslongasthereisasignalinanyofthe Forthisexperiment,thedetectorsofchoicewereparallel- PPADs,anX-triggersignalisproduced.IftheX-triggersig- plate avalanche detectors ~PPADs! because they are known nal is in coincidence with any of the tagger channels, the to be very efficient in detecting fission fragments. They are X-ref signal is sent back to the experimental electronics by also ideal for photofission experiments because they have thetaggerinterface,andisusedasgatefortheADCsandthe goodtimeresolution~allowingtheuseofthephoton-tagging coincidenceregisters.Itisalsousedasthestartsignalforthe technique!, and they are practically insensitive to neutrons, time-to-digital converters ~TDCs!. photons, and electrons @26–33#. And although PPADs are Every time that an X-ref signal is produced, the tagger sensitivetoaparticles,whichareacommonsourceofback- interface also sends a LAM ~look at me! signal to the data- 034604-5 J.C.SANABRIAetal. PHYSICALREVIEWC61034604 FIG.4. Experimentalelectron- ics.Fordetails,seetext. acquisition~DAQ!computer.InresponsetotheLAMsignal, The photon-tagging rate was kept at about 43106 pho- theDAQsystemreadsoutthecontentsofADCs,TDCs,and tonspersecond,andtheaveragetaggingefficiencywas45% coincidence registers. for the regular tagger ~collimator diameter 510 mm) and 40% for the end-point tagger ~collimator diameter 515 mm). The reason the latter is lower is that the end- D.Runsummary point tagger is further from the collimator. The pressure of The experiment took place at SAL during the month of thereactionchamberwas15Torr,andthevoltageappliedto January of 1997. Table I shows the photon energy ranges thedetectorswas750V.Duringtheexperimentthreetargets covered during the experiment. of 238U, 235U, 237Np, and 232Th, and four of 233U were Datawererecordedduringintervalsofapproximately2h. installed inside the reaction chamber. A 252Cf source and a Every8h,thereactionchamberwasremovedfromthebeam dedicated PPAD were also included, positioned out of the line and the total-absorption lead-glass detector was used to photon beam line, so that they could act as a monitor of the perform a tagging-efficiency run. These runs were divided performance of the other PPADs. Two aluminum targets of into two parts. In the first part, the lead-glass counter would different thicknesses were included in order to study the ef- count all the photons that were not absorbed by the collima- fect of this material present in the targets ~the fissionable tor, and whose signals were in coincidence with any tagger- material is deposited on 100-mm-thick aluminum foils!. For channel signal. In the second part, the radiator would be details, see Refs. @34,35#. removedfromthebeamlineandthenumberofcoincidences between the lead-glass detector and the tagger channels was III.DATAREDUCTION measured. The events recorded during the second part can Theexperimentallymeasuredtotalphotofissioncrosssec- only be explained as the result of background in both detec- tionsareproportionaltotheratiooffissioneventsandtagged torscausedbytheelectronbeam~thenumberoftheseevents photons. The number of fission events is determined from was always found to be negligible!. thecoincidencebetweensignalsfromPPADsandtaggerde- tectors ~having subtracted the background events resulting TABLE I. Photon energy ranges for the various tagger settings fromaccidentalcoincidencesandaparticles!.Correctionsto usedinthisexperiment. this number have to be made to account for the fraction of solidanglenotcoveredbythedetectorsandforthedetection Setting Tagger Egmin @MeV# Egmax @MeV# efficiency of the PPADs. The number of tagged photons is evaluatedbycountingthenumberofelectronsdetectedatthe 1 Regular 68 146 focal plane of the tagger, and then corrected for those pho- 2 Regular 128 189 tons absorbed by the collimator ~the tagging efficiency!. 3 Regular 161 206 Therefore, the photofission cross sections were determined 4 Regular 182 223 using the following equation: 5 Endpoint 218 247 S D 6 Endpoint 234 257 4p Al Yl 7 Endpoint 248 264 sl~Ek!5 N0 tleDl ETVDl ET NkekTkAG31030 @mb# ~8! 034604-6 PHOTOFISSIONOFACTINIDENUCLEIINTHE... PHYSICALREVIEWC61034604 FIG. 5. Tagger TDC spectrum for the target 237Np-2 before ADCcut. FIG.6. ADCpulse-heightspectrumforthetarget 237Np-2. wherelistheindexthatrunsovertargets,kistheindexthat as a narrow peak with a width of about 2.5 ns, which is the runs over tagger channels, E is the photon energy associated combinedresolutiontimeofthePPADsandtaggerchannels. with a tagger channel @MeV#, N is Avogadro’s number Figure 5 shows the TDC spectrum for one of the 237Np 0 @atoms/mol#, A is the target atomic weight @g/mol#, t is the targets. In this example the contributions of all the tagger target thickness @mg/cm2#, e is the detector efficiency, channels have been combined to enhance the statistics. DET V is the detection solid angle @sr#, eTAG is the tagging Thecoincidencepeakrestsontopofapedestalofrandom DET efficiency, Y is the yield of fission events induced by tagged coincidences. To subtract the background under the peak, photons, and N is the number of tagged photons. oneinterpolatesbetweenthebackgroundlevelsonbothsides The photon tagger covers only a fraction of the brems- of the peak. For this analysis, the level of background was strahlungspectrum.Bychangingthefieldofthemagnetand evaluated by taking the average of the number of counts in the position of the focal-plane detector array, different en- regions very close to the coincidence peak, in order to mini- ergy regions can be spanned. A first magnet ~‘‘regular tag- mize the effect of any structure in the distribution of back- ger’’! can tag photons up to ;223 MeV. In order to tag ground. There is an uncertainty associated with any back- higher-energyphotons,asecondmagnetisused~‘‘end-point groundsubtractionprocedure,butaswewillseelater,inthis tagger’’!withthesamefocal-planedetectorarray.InTableI case its magnitude is within the accuracy of our measure- the photon energy ranges used during the experiment are ments, and therefore there is no need to implement a more listed. For each tagger setting there are 62 energy channels. sophisticated procedure. The calibration of the energy covered by each channel was In principle, the TDC spectra are the only data needed to performed previously by the SAL staff. determinethenumberoftaggedfissionevents.However,the To determine the number of fission events induced by energy information stored in the PPAD ADC spectra can be tagged photons, the basic information comes from the TDC usedtoreducethelevelofbackground,andthustofacilitate spectrum of each tagger channel. The TDC is started by the its subtraction. In Fig. 6 a typical ADC spectrum is pre- X-trigger signal coming from a PPAD and it is stopped by a sented. The decaying distribution in the low-energy part is signal in the tagger channel. The event is accepted and read thecontributionoftheaparticles,andthelargecentralpeak outbytheDAQsystemifthetwosignalsarewithinagiven results from the fission fragments ~tagged and untagged!. resolvingtime~60nsinthiscase!.Theresolvingtimeshould Thetwodistributionsoverlapduetothewidecollimation.In be long enough so that a significant portion of the random spite of the overlap, one can still reject the events whose coincidence spectrum is recorded, studied, and subtracted. ADC channel number is low enough so that it cannot be the In many cases the signal in the PPAD is produced either result of a fission fragment, for example, in the case of Fig. byafissioneventinducedbyanuntaggedphotonorbyana 6,eventswhoseADCchannelnumberisbelow;75.Inthis particle. For such events ~accidental coincidences!, the fact way most of the aparticles are eliminated from the data. thatthePPADsignalmightbeincoincidencewiththesignal The number of tagged photons is equivalent to the num- from a tagger channel is an accident. However, there is no berofelectronsdetectedbythechannelsofthetaggerfocal- time correlation between those two signals, and these kinds planedetectorarray.Theeventrateperchannelisregistered of events therefore produce a flat background in the TDC by the scalers of the tagger electronics ~they are read out by spectrum. theDAQsystemevery10s!.Thesumofallthescalerread- When the PPAD signal is the result of a fission event ings over a period of time represents the number of tagged induced by a tagged photon, there is a time correlation be- photonsperchanneloverthatperiod.Ofthephotonsthatare tweenthissignalandthesignalproducedinoneofthetagger tagged,onlyabouthalfstrikethetargetsamples.Therestare channels by the corresponding electron. These kinds of absorbedinathickleadcollimatorwithanaperturediameter events ~true coincidences! will appear in the TDC spectrum of 10 mm, located 2 m downstream from the aluminum ra- 034604-7 J.C.SANABRIAetal. PHYSICALREVIEWC61034604 TABLEII. Thicknessesofthetargetsusedinthisexperiment,as measuredbytheiraactivities. Target Technique Thickness@mg/cm2# 232Th-1 Countingrate 1.2560.02 232Th-2 Countingrate 1.2360.02 232Th-3 Countingrate 1.2360.02 233U-1 Energyloss 0.6260.02 233U-2 Energyloss 0.6560.02 233U-3 Energyloss 0.5760.02 233U-4 Energyloss 0.5860.02 235U-1 Energyloss 0.8960.03 235U-2 Energyloss 0.9360.03 235U-3 Energyloss 1.0760.04 238U-1 Countingrate 1.4860.03 238U-2 Countingrate 1.6160.03 FIG.7. Resultofatypicaltaggingefficiencyrun. 238U-3 Countingrate 1.2360.02 237Np-1 Countingrate 0.9460.02 diator in the case of the regular tagger, and a 15-mm lead 237Np-2 Countingrate 0.9960.02 collimator located 3.9 m downstream from the aluminum 237Np-3 Countingrate 0.9360.02 radiator in the case of the end-point tagger ~both collimators were 130 mm thick!. All other sources of inefficiency, such as electronic noise in a focal-plane detector, were negligible ciency is approximately equal to the fraction of the area by comparison. The tagging efficiency was determined dur- blocked by the wires, which is 2%, resulting in a detection ing special runs in which the tagger was put in coincidence efficiencyof0.98.Sincetheacceptedfissionfragmentswere with the lead-glass photon detector positioned in the photon emittedinthedirectionnormaltotheplaneofthetarget~due beamline4mdownstreamfromthefissionchamber.Imme- to the collimation!, the distance that they traveled inside the diately after each of these normalization runs, the aluminum film of fissionable material was minimized. A detailed study radiator was removed from the beam line and background of the existing data on the range of typical fission fragments datawererecorded.Inthiswaythelevelofbackgroundcom- traveling through uranium showed that the fraction of frag- ingfromthetaggerspectrometerwasmeasuredandfoundto ments reabsorbed by the film is negligible. beminimal.Theefficiencyofeachofthe62channelsofthe The targets consist of thin layers of actinide isotopes de- photon tagger was evaluated using positedonaluminumfoils100 mmthick.Thedimensionsof thefoilsare12cmlongand6cmwide.Onlyacentralcircle YIN eTAG5 k , ~9! of 4 cm in diameter was exposed to the photon beam. In k NIN order to accurately determine the thickness and composition k of the film, a measurement of the rate and energy spectrum where N represents the number of electrons detected in the oftheaactivityforeachtargetwasperformedatSAL.Each k kth channel of the tagger, Y represents the number of those individual target was placed inside a vacuum chamber and k events detected in coincidence with the lead-glass detector, the aparticles were detected with a silicon surface barrier and the superscript ‘‘IN’’ represents the status of the alumi- detector.ScalerandADCinformationwasrecorded.Inorder num radiator with respect to the beam line. The result of a to get an energy calibration, data were also recorded for typical normalization run is shown in Fig. 7. 210Po, 241Am, and 252Cf sources. The energy resolution of In order to determine the geometrical acceptance of the thissetupwasapproximately30keV,allowingustoidentify target-detector pairs, a Monte Carlo calculation was per- the different isotopes present in the samples with no ambi- formed. This calculation included electron-beam spot size, guity.Bymeasuringtherateofaparticlesofagivenisotope the angular distribution for the bremsstrahlung photons, the emitted from a particular region of the foil in a particular effect of the photon-beam collimator, the isotropic distribu- solid angle, we were able to determine the amount of that tionoffissionfragmentsemittedbythetarget,andtheeffect isotopeinthesample.Withthistechniquewedeterminedthe of any collimators between the target and the PPAD. In this thicknesses of our 232Th, 238U, and 237Np films. For the way the size of the photon-beam spot at the position of each cases of 233U and 235U the presence of 234U and 238U con- target was accounted for. The effective solid angle covered tamination did not allow us to use the counting-rate tech- by the PPAD was then determined to be V 5 nique.Forthesetargetsweusedtheinformationprovidedby DET 0.3934p sr. the 232Th, 237Np, and 238U films ~whose thicknesses we had A PPAD is a very efficient fission-fragment detector. determinedalready!toestablisharelationbetweentheADC- Mostoftheinefficiencycomesfromthewiregridthatforms spectrum width and the film thicknesses for different iso- the anode plane. The collimators limit the angles at which topes, determining in this way the thickness of the 233U and the particles enter the active region; therefore, the ineffi- 235U samples. Table II summarizes the results of our mea- 034604-8 PHOTOFISSIONOFACTINIDENUCLEIINTHE... PHYSICALREVIEWC61034604 surements. A detailed description of the target-thickness measurements can be found elsewhere @35,36#. Asmentionedintheprevioussection,thefilmsoffission- able material were deposited on aluminum foils, 100 mm thick.Inordertomeasurethecontributionofthismaterialto the fission-fragment yield, aluminum foils were included in the reaction chamber. The reaction cross sections for these samples were measured using the same procedure as for the targets. The contribution of the aluminum substrate to the fission yields was determined to be of the order of 1023%; therefore its effect on the measured photofission cross sec- tions was negligible. The efficiency and the solid-angle acceptance of the de- tectors have been calculated numerically. As was mentioned above, these calculations include all the elements that could have an effect on the results. The uncertainties associated with these calculations have been estimated to be less than 1% for both quantities. The tagging efficiency has been measured to a statistical uncertainty of 1%. Any systematic uncertainty associated with the procedure comes from the fact that the accelerator electron current is decreased during the tagging efficiency runs. ~This is done in order to keep from overloading the lead-glass detector.! The physicists at SAL have performed this kind of measurement for several years and for many differentexperiments,andtheirobservationsindicatethatthe effect of the beam current reduction is small. We estimate theuncertaintyintheknowledgeofthetaggingefficiencyfor this experiment to be 2%. The uncertainty associated with the background- subtraction procedure is caused by the structure of the TDC distribution of random coincidences ~see Fig. 5!. If this dis- tribution were perfectly flat, there would be no systematic uncertainty in the interpolation that was done to determine the level of background under the coincidence peak. How- ever, the presence of time-dependent structure in the distri- FIG.8. Photofissioncrosssectionspernucleonfortheuranium bution makes the procedure somewhat less accurate. isotopesmeasuredinthisexperiment. The level of background under the coincidence peak in the TDC spectrum for each PPAD was determined by mea- tinideisotopesarepresented.Fromthesefiguresonecansee suring the average number of events per channel in the two thatallthecrosssectionsshowasimilarenergydependence. regions adjacent to the coincidence peak. The boundaries They are approximately constant up to 100 MeV photon en- defining the coincidence peak and the two adjacent back- ergy~inthequasideuteronregion!,andthentheriseoftheD ground regions were selected individually for each PPAD resonance is clearly observed. There is no clear evidence of andeachdataset~20PPADsand26datasets!.Bychoosing any other energy-dependent structure, especially above the different sets of boundaries, and studying the changes in the pion-production thresholds. The most important feature of background levels, the effect of this procedure on the final the results is the significantly higher cross section for 237Np cross sections was evaluated. The background subtraction when compared with 238U, and even with 235U. The cross studies showed that the change in the average cross sections section for 232Th is much lower than the ones for the other duetodifferentchoicesofbackgroundlevels~withinreason- actinide isotopes; this was expected from previous results at able limits! was at most 3%. lower and higher energies. Smaller differences among the uranium isotopes are also noticeable. In the next section IV.RESULTS these results will be analyzed in more detail, and compared with existing data and calculations. A.Totalphotofissioncrosssections To test the self-consistency of the results, the cross sec- Following the methods presented in the previous section, tions for the different targets and different tagger settings the total photofission cross sections for each target and each werecompared.Thedataforthedifferentsettingsalwaysfell tagger setting were measured. The cross sections for the tar- within the statistical uncertainties of the measurements, as gets of the same isotope were then combined. In Figs. 8 and didthedataforthedifferenttargetsofthesameisotope.The 9, our measured total photofission cross sections for the ac- datafromthetwotaggersusedintheexperiment,the‘‘regu- 034604-9 J.C.SANABRIAetal. PHYSICALREVIEWC61034604 FIG. 9. Photofission cross sections per nucleon for 232Th and 237Npmeasuredinthisexperiment. lar tagger’’ and the ‘‘end-point tagger,’’ also matched in the FIG. 10. Fissilities relative to 237Np measured by this experi- overlapping region. The data from the end-point tagger are ment. The ratios of the 237Np data relative to the fitted curve are slightlymorescattered,probablytheresultofthepresenceof shownaswell,sothatonecanassesstheirscatter. somewhathigherbackgroundsinthefocal-planedetectorar- rayforthisspectrometer.Todouble-checkthenormalization B.Relativefissilities among different targets of the same isotope, the average cross sections over the measured photon energy range were Traditionally the fissilities of the actinide isotopes have calculated and compared. From these results we notice that been presented relative to 238U by dividing the total photo- the 233U and 235U cross sections have a larger standard de- fission cross section of the isotope by the total photofission viationthantheothers.Thisdifferenceprobablyresultsfrom cross section of 238U. However, since the cross section for the uncertainty in the target-thickness determination, since 237Np is now seen to be higher, it makes more sense to for these two isotopes the target thickness was measured us- present the fissilities relative to this isotope. To do this, a ing the energy-loss technique, in contrast with the other iso- fifth-order polynomial was fitted to the 237Np cross section, topes, for which the thickness was determined using the a and then the other cross sections were divided by this func- activitytechnique.Inthecaseof 233U,itisalsoimportantto tion. Figure 10 shows the fissilities of the actinide isotopes note that the level of background was much higher than for relative to 237Np. the other isotopes due to its very high aactivity. The sys- Two important features of the results are the very weak tematic uncertainties for each cross section are listed in dependence of the fission probability on photon energy and Table III. its strong dependence on the fissility parameter Z2/A ~of the target isotope!. The differences in fissility between 237Np, 238U,and 232Thareverywellestablished~seeFig.10!.Less TABLEIII. Systematicuncertaintiesforthephotofissioncross- important differences are observed among the uranium iso- sectionmeasurements. topes: the fissility of 235U is highest and that of 238U is the lowest. Isotope Y eTAG V e tl Total k DET DET 232Th 1.5% 2% ,1% ,1% 2.0% 3.5% C.Cross-sectionfits 233U 2.5% 2% ,1% ,1% 3.5% 4.9% 235U 2.5% 2% ,1% ,1% 3.5% 4.9% Polynomial fits of fifth order to all the measured total 238U 2.0% 2% ,1% ,1% 2.0% 3.7% photofissioncrosssectionswereperformedinordertoassess 237Np 1.8% 2% ,1% ,1% 2.0% 3.6% differences in shape and absolute value of the cross sections for the different isotopes. The x2/N did not exceed 1.3 for df 034604-10
Description: