IEEE Transactions on Nuclear Science, Vol. NS-27, No. 1, February 1980
EVALUATION OF A PHOSWICH DETECTOR FOR THEIN SITU ANALYSIS OF 90Sr
N. A. WogmanR. L. Brodzinski
D. P. Brown
Battelle, Pacific Northwest LaboratoryRichland, Washington 99352
SUMMARY
A Phoswich detector system which employs a 3.45mm thick by 12.5 cm diameter CaF2(Eu) crystal coupledto a 6.4 cm thick NaI(Tl) crystal was evaluated forthe detection of 90Sr(90Y) in the presence of 106Ru(l06Rh) and 137Cs and for its ability to measure238pu, 239pu, 241Am and 244Cm. The radioactivespecies were studied, both as surface contaminationand contamination distributed in sediments, as afunction of overburden thickness for different mate-rials. The CaF2(Eu) crystal detects beta or lowenergy photon emissions while the NaI(Tl) crystal onlydetects higher energy photons from radionuclide de-cay or bremsstrahlung radiation. As little as 9d/m/cm2 of 90Sr as surface contamination can bemeasured in a 1000 second counting period in the pres-ence of comparable concentrations of '37Cs and 106Ru.The Phoswich system is specifically designed for themeasurement of 90Sr(90Y) and does not detect thetransuranic species well. The operating parametersand use of the Phoswich system for identification ofspecific radionuclides in shallow land waste burialenvironments are discussed.
INTRODUCTION
Strontium-90 produced in the nuclear weapons andcommercial nuclear reactor programs must be measuredin a variety of field situations including nuclearwaste storage areas and reactor sites. Strontium-90and its daughter 90Y are pure beta emitters, and canonly be measured by direct analysis of the beta emis-sions, by the analysis of bremsstrahlung radiation,or by neutron activation analysis.
In recent years significant advances have beenmade in the development of low-level radionuclideanalysis systems. For low-level gamma-ray spectrom-etry, large volume NaI(Tl) crystal and Ge(Li) diodegamma-ray spectrometer systems have been developedand applied to the nondestructive measurement of com-plex radionuclide spectra. An increased emphasison environmental monitoring has inspired improvedtechniques for beta and alpha analysis. These im-provements are generally aimed at increasing the sen-sitivity or selectivity of a system for a specificradioisotope in a selected environment. The followingexamples detail some problem areas where 90Sr measure-ment is required or 90Sr inhibits the analysis ofother radioactive species.
* The measurement of effluents from evampyatorsassociated with nuclear burial sites is)irequired for the determination of decontam-ination factors and process control infor-mation.
e In routine surveillance operations the mea-surement of radionuclides in air from anoperating nuclear facility involves entrain-
ment of the radionuclides on selective filteror absorptive media before discharge and adetermination of radioactivity on the filtermaterial. A more sensitive instrumental tech-nique for determination of these radionuclideswithout extensive chemical separations(2is desirable.
e The necessity for in-situ determination ofalpha emitting contaminants in soil motivatedthe development of a ZnS scintillation systemand a Phoswich detector as field surveyinstruments. 3)
o Intrinsic germanium and Ge(Li) detectors havebeen adapted to the in-situ determination ofsubterranean 90Sr concentrations.(4) Thesesystems were designed for migration ratestudies in and around shallow land burialgrounds.
o The measurement of surfaces contaminated byradionuclides emitting low energy photons hasprompte v5valuation of a variety of differentprobes.
The following discussion centers on the use of aPhoswich detector for the measurement of 90Sr in thepresence of the typical nuclear waste radioisotopesMRu, 137Cs, 238Pu, 239pu, 241Am, and 244Cm.
INSTRUMENTATION
A Phoswich detector system designed primarily forthe measurement of 90Sr is shown in Figure 1. It con-sists of a primary CaF?(Eu) beta detector 12.5 cm indiameter by 3.45 mm thick covered with a 0.006 mmaluminized mylar window. This thin window allowsalpha particles as well as beta particles to enter theCaF2(Eu) crystal. A 3.2 mm thick quartz light pipe/absorber is sandwiched between the CaF2(Eu) beta de-tector and a 6.4 cm thick NaI(Tl) gamma-ray detector.The assembly is housed in stainless steel and coupledto an RCA 8055 photomultiplier tube with an attachedlow background tube base. System electronics consistof a high voltage power supply for the photomultipliertube, a preamplifier, a pulse shape amplifier/analyzer,and two scaler-timers. A 10 cm thick lead shield wasused to eliminate the majority of background events.In this evaluation, one scaler recorded the pulses fromthe CaF (Eu) crystal while the other recorded the pho-tons detected by the NaI(Tl) crystal. In normalPhoswich operation, the gamma detector is operated inanticoincidence with the CaF2(Eu) detector. However,in this evaluation, both beta and gamma-rays wereanalyzed; thus the pulse shape analyzer was electron-ically optimized for this purpose. The system wasadjusted by exposing to to M37Cs and 90Sr(90Y) radia-tions. A time output oscilloscope was used to observe
the amplitude and width of the pulse height distri-bution time spectrum. This spectrum contained thetwo time distributions for the beta particle andgamma-ray interactions in their respective detectors[beta in CaF2(Eu) and gamma-ray in NaI(Tl)]. By re-moving one or the other of these sources the timespectrum belonging to each was identified. Time con-trols were adjusted to maximize the separation of thetwo time amplitude spectra. Hence, the system wasoptimized for splitting the beta and gamma-ray pulseswhich, as will be seen later, significantly reducedthe overall efficiency of the beta detecting crystal.
STANDARDS AND ABSORBERS
Sources representing surface contamination wereprepared by uniformly distributing known quantitiesof the appropriate radionuclides on a saran sheetwhich was mounted on a 15 cm diameter plastic ringto provide rigidity. Two sets of standards were pre-pared in this fashion; one set contained no coveringmaterial while the second set was covered lightlywith Krylon to insure integrity. No difference incounting characteristics was found between these twosource configurations. Standards without Kryloncoating were also made for the four transuranic radio-isotopes 238PU, 239pu, 241Am and 244Cm.
Sources representing contamination uniformly dis-tributed in sediments were also prepared for eachradionuclide using sieved Hanford sand. Each isotopewas spiked into 686 grams of sand and was dried andmixed by quartering and coning to insure homogeneity.These sources were then poured into a 2.5 cm deep by15 cm diameter plastic container and covered withsaran. Table I lists the radionuclide sources andtheir concentrations used in this study.
Aluminum, steel, polyvinyl chloride (PVC),. fiber-glass, wood, dunite sand, and Ni-CaSO4 absorbers wereprepared in a range of thicknesses varying from 10-10,000 mg/cm2 to study the effects of various over-burden materials commonly found in field situations.The dunite sand and Ni-CaSO4 absorbers were preparedin the form of plates by "casting" in plaster-of-paris and were intended to represent various min-eralogical overburdens. The dunite sand was mixedwith small amounts of CaSO4 such that the density ofthe cast plates varied from 1.72 to 2.83. The Ni-CaSO4 plates varied in density from 2.97 to 5.08.
RESULTS
In normal Phoswich detector operation, thesecondary crystal [NaI(Tl) in this system] serves asan anticoincidence shield for the primary crystal[CaF2(Eu) in this system]. However, in this investi-gation, both crystal outputs were independently re-
corded such that a measure of the events in each crys-tal could be studied from a variety of sources. Theeffects of various absorber materials on the responseof each crystal to specific radionuclides were alsomeasured.
The radioisotopes 238Pu, 239pu, 241Am, and 244Cmwere investigated first to determine whether theywould interfere with the detection of 90Sr(90Y) and toevaluate the use of this Phoswich system for the fieldmeasurement of transuranic activities. The efficien-cies determined for the massless "surface contami-nation" standards were 0.0059%, 0.0031%, 0.0045%,and 0.0024% in the beta detector and 0.14%, 0.030%,0.57%, and 0.15% in the gamma-ray detector for 238Pu,
239Pu, 241Am, and 244Cm,respectively. The correspon-ding efficiencies for 90Sr(90Y) were measured to be4.2% and 0.12%. Therefore, these transuranics will notinterfere with the detection of 90Sr in this Phoswichsystem unless their concentrations are more than 1000times higher than 90Sr.
A Phoswich detector system which had been develop-ed to specifically identify transuranics in field sit-uations used a 1 mm thick NaI(Tl) crystal in t4 frontand a 38 mm thick CsI(Na) crystal in the back. t) Thedetector used in the current work was designed specif-ically for the detection of 90Sr(90Y) beta particles.The data from these two studies indicate that Phoswichdetectors can be successfully engineered to enhancetheir sensitivity for any particular species desired.
Since the transuranic isotopes would not interferewith the detection of 90Sr, major emphasis was placedon the interferences created by 137Cs and 106Ru4 Aspreviously mentioned, the Phoswich system describedherein was electronically adjusted to maximize theseparation between signals from the CaF2(Eu) andNaI(Tl) crystals or, in effect, between those frombeta particles and those from gamma-rays. Since90Sr(90Y) is essentially a pure beta emitter, it is notsurprising that the "beta" efficiency is 34 times ashigh as the "gamma" efficiency which is due primarilyto the bremsstrahlung interaction of 90Sr(90Y) betaparticles with their environment. Conversely, 137Cs,a high yield gamma-ray emitter with only a low energybeta emission was detected with 2.2% efficiency in theCaF2(Eu) crystal and 6.4% efficiency in the NaI(Tl)crystal.
These efficiencies are reversed from those of90Sr(90Y) and provide a separation factor of 99 forthe measurement of 90Sr in the presence of 137Cs. Ru-thenium-106 (rhodium-106), another radioisotope pairnormally present in nuclear wastes, yields copiousquantities of both high energy beta particles andgamma radiation and was found to be 7.9% efficient inthe beta detecting crystal and 2.4% efficient in thegamma-ray detecting crystal. Therefore, the separationfactor for detection of 90Sr in the presence of 106Ruis only 11, and large quantities of M06Ru could sig-nificantly interfere with the measurement of 90Sr.
If the system were to be maximized for beta de-tection only with the NaI(Tl) detector in anticoin-cidence to the CaFZ(Eu) crystal, 90Sr could be detectedat up to 45% efficiency. The response of the Phoswichsystem to a massless (surface contamination) and a 2.5cm thick homogeneous sand (contaminated sediments) 90Sr(90Y) source is shown in Figure 2 as a function ofoverburden (absorber) composition and thickness. Thebeta curves for both 90Sr sources decrease systemat-ically with increasing absorber thickness independentof the composition. However, the gamma-ray response iswidely scattered and uncorrelated. Since this resultwas not anticipated, the experiments were repeated ata later date, and each individual data point was ex-actly reproduced as in the earlier "shotgun" pattern.
To investigate this variable but reproduciblegamma-ray response to 90Sr a series of experiments were
conducted on bremsstrahlung production in variousabsorbers. Energy spectral data were accumulated witha Ge(Li) spectrometer viewing a 0.4 mCi 90Sr sourceshielded by dunite, fiberglass, aluminum, wood, andPVC absorbers an average of 822 mg/cm2 thick. A spec-trum was also obtained for a slightly thicker stainlesssteel absorber. The purpose of these experiments was
to determine the difference in the shape of the brems-strahlung spectra due to differences in the atomic
734
number of the absorber material. As can be seen inFigure 3, no significant difference in curve shapeis seen. The bremsstrahlung production in thesevarious absorber materials is nearly identical. Thetotal number of photons above 40 keV in each of thesespectra is also constant within ± 6% as shown inTable II. Hence, these data indicate that a uniformresponse should have been recorded by the NaI(Tl)crystal of the Phoswich assembly; thus the scatterin the gamma-ray data is as yet unexplained.
Table II also contains information describingthe gama-ray detector response to 90Sr(90Y) as afunction of aluminum absorber thickness. The datashow that the bremsstrahlung attenuation in Al variesby a factor of a across absorber thicknesses from 41to 10,000 mg/cm2. These data correspond to the ex-pected attenuation factors integrated over the entirebremsstrahlung spectrum.
The beta crystal response was only 4 timesgreater than the gamma-ray crystal response for thethick 90Sr(90Y) sand source which correlates with theexpected attenuation by the bulk sample.
The beta and gamma response to the 137Cs standardsources is shown in Figure 4. Smooth curves were ob-tained for both sources as a function of absorberthickness for both the beta and gamma crystal outputs.This makes the uncorrelated data obtained from thegamma detector for the 90Sr source even more difficultto explain.
The response curves to the 106Ru sources areshown in Figure 5. Again, the curves are smoothfunctions of absorber thickness and display trendssimilar to those of 137Cs.
Figures 6, 7, and 8 show the response of thePhoswich system to the mixed radioisotopic sources -90Sr plus 137Cs, 90Sr plus 106Ru, and 90Sr plus 106Ruplus 137Cs,respectively. The experimental data inthese figures coincide precisely with the sum of theindividual radioisotopic source data shown in Figures2, 4, and 5. Although this was not unexpected, it wasgratifying to see that the mixed radioisotope curveswere not perturbed by the additional energy depositionor dead time losses in the Phoswich detector system.Even the mixed gamma-ray response curves are smoothand not adversely affected by the variable 90Sr com-ponent which becomes a negligibly small contributionto the total response.
DISCUSSION
Since the data obtained from mixingradioisotopes are strictly additive, the curveshapes for complex radioisotopic mixtures can be pre-dicted from individual radioisotopic standards. Thisis most easily accomplished using computer techniques.The absolute efficiencies for any absorber thicknessfor a given mixture of radioisotopic sources can besynthetically generated by the computer and comparedto data obtained in field environments. In normaloperation the quantity of 90Sr will be measured usingsimultaneous equations for the various species presentbased on the data in this communication. Althoughthis system has not as yet been used in a field situa-tion, as little as 9 disintegrations per minute percm2 of 90Sr as surface contamination can be measuredin a 1000 second counting period in the presence ofcomparable concentrations of 137Cs and 106Ru.
However, for 905r homogeneously mixed in sandysoils typical of the Hanford, Washington nuclear wasteareas the detection limit is 1.5 nCi/cm3 in a 1000second counting period if 137Cs and 106Ru are present
at a concentration of 0.1 nCi/cm3. This can be com-pared to a 90Sr detection limit of 0.03 nCi/cm3 underthe same conditions utilizing a bremsstrahlung detec-tion techniquea4nd an intrinsic Ge diode gamma-rayspectrometer.t4) Hence, although this Phoswich systemprovides adequate sensitivity for measuring 90Sr as asurface contamination, superior techniques are avail-able for measuring 90Sr interstitially mixed in sedi-ments.
Research to define the erratic gamma-ray responsefrom 90Sr is underway. In addition, attempts are beingmade to determine the photon energy spectrum from theNaI(Tl) crystal output. If this can be accomplished,spectral data from the crystal will allow photonemitting species to be determined directly by typicalgamma-ray spectroscopy. Species emitting multiplephotons can be used to define their average distancefrom the detector and hence the ave aqe thickness ofoverburden using gamma peak ratios-.(7) Isotopeidentification and attenuation data will greatly assistin the computer analysis of 90Sr concentrations anddistribution patterns.
ACKNOWLEDGMENTS
The authors would like to acknowledge the assist-ance of Don Edwards and Rush Campbell who rapidly andaccurately prepared the sources and calibrated thedetector with continual persistence to detail.
REFERENCES
1. R. L. Blanchard, B. M. Montgomery, H. E. Kolde,and G. L. Gels, Environmental Protection Agency,Montgomery, AL, "Supplementary RadiologicalMeasurements at the Maxey Flats Radioactive WasteBurial Site," 1976-1977. Report No. EPA/520/5--78-011.
2. H. L. Krieger, E. R. Martin, and G. W. Frishkorn."Sequential Radiochemical Analysis for Ru, Sr,and Cs in Environmental Air." Health Physics 30,#6:465-470, 1976.
3. A. J. Ahlquist, C. J. Umbarger, and A. K. Stoker."Recent Developments for Field Monitoring of AlphaEmitting Contaminants in the Environment." HealthPhysics 34:486-489, 1978.
4. R. L. Brodzinski and H. L. Nielson, "A WellLogging Technique for the In-Situ Determination of90Sr," and PNL-SA-7246, 1979. Submitted toNuclear Instruments and Methods for publication.
5. W. J. Iles, P. H. Burges, B. F. White. "Measure-ment of Activity of Surfaces Contaminated by BetaEmitting Nuclides and 55Fe." National Radio-logical Protection Board, Harwell, United Kingdom.Report #NRPB-R--60. NTISPC A02/MF AO1, February1977.
6. C. J. Umbarger and M. A. Wolfe. "A Battery Oper-ated Portable Phoswich Detector for Field Monitor-ing of Low Level Transuranic Contaminants."Nuclear Instruments and Methods 155:453-457, 1978
7. H. L. Nielson, N. A. Wogman, R. L. Brodzinski."In-Situ Subterranean Gamma-Ray Spectroscopy."Nuclear Instruments and Methods 143:385-389, 1977.
735
TABLE I
RADIONUCLIDE SOURCES USED TO EVALUATE A PHOSWICH DETECTOR
Source* Radionuclide dpm**
A 137Cs 1.11 x 105B 90Sr 1.30 x 105C 106Ru 1.02 x 105
S90r 1.30 x 1050 137Cs 1.11 x 105
90Sr 1.30 x 105106Ru 1.02 x 105
90Sr 1.30 x 105F 106Ru 1.02 x 105
137Cs 1.11 x 1051 06Ru 1.02 x 105
G 137Cs 1.11 x 105
H 239Pu 9.8 x 104I 238UPu1.0 x 105
i 241Am 1.32 x 105
K 244Cm 9.95 x 104
*Both massless and 2.54 cm thick sand sources 15 cm diameter were made.
**disintegrations per minute.
TABLE II
BREMSSTRAHLUNG CREATION AS A FUNCTION OF ABSORBER TYPEAND THICKNESS FROM A 0.4 mCI 90Sr SOURCE
