180 BANHAM et al. : GERMANIUM DIODES FOR GAMMA-SPECTROMETRIC [Analyst, Vol. 91 The Use of Lithium-drifted Germanium Diodes for the -spectrometric Determination of Radioactive Fission-product Nuclides BY M. F. BANHAM, A. J. FUDGE AND J. H. HOWES (Chemistry and Electronics Divisions, U. K. Atomic Energy Research Establishment, Harwell, Didcot, Berks.) The superiority of a y-spectrometer incorporating a germanium - lithium diode detector and field-eff ect transistor head amplifier over the conventional sodium iodide - thallium system, for the resolution of most of the difficult determinations encountered in fission-product radiochemistry, is demonstrated. DURING the past 6 years, semiconductor solid-state detectors have become generally available for the detection of heavy charged particles.Lately, the discovery of the lithium drift- process, first in silicon1 and then in germanium,2 has led to the publication of several papers on their use, initially for conversion electron investigations and most recently for y-spectro- metry.3s4s6 Because of its higher atomic number germanium has a much better photo-electric response than silicon, and is therefore a much more attractive detector for y-spectrometry, particularly in the region up to about 1.5 MeV. In the drift process, lithium is diffused to depths of several millimetres in a piece of p-type germanium to produce a layer of n+-type material, a central layer of intrinsic material and a layer of the original p-type germanium. The whole makes a p-i-n detector. Incident y-radiation is absorbed by photo-electric or Compton processes, and the ions produced are collected by means of a potential difference applied across the detector.Lithium ions are highly mobile at normal temperatures and so, to achieve stability and a low noise level in the detector, it must be operated at low temperatures. It is usual to mount the detector in a cryostat operated at liquid-nitrogen temperature (77" K). y-Spectrometry is one of the most useful techniques available to radiochemists for the analysis of mixed fission products in irradiated nuclear fuel specimens. However, the best sodium iodide detector has a resolution of not less than 50 keV (the full width of the photopeak at half maximum height) a t 662 keV (caesium-137). This has been inadequate for the determination of some individual nuclides in fission-product mixtures.Several difficult problems in fission-product analysis have been overcome by radio- chemical separation, or where this is not possible, by mathematical analysis of the y-spectrum. There are, in particular, four instances where an improvement in resolution, such as is claimed for lithium-drifted germanium diodes, would enable direct measurement of the activities to be made. These determinations are as follows- (a) zirconium-95 in the presence of niobium-95 daughter, (b) ruthenium-106 in the presence of ruthenium-103, (c) caesium-137 in the presence of caesium-134, ( d ) cerium-144 in the presence of cerium-141. The three fission-product nuclides most frequently used for the determination of burn-up (the total number of fissions) of irradiated nuclear fuel specimens are zirconium-95, caesium-137 and cerium-144.A preliminary investigation has been carried out to test the feasibility of using the high resolving power of a germanium - lithium detector for the determination of these nuclides in mixed fission products, with a view to extending its application to non- destructive y-scanning of nuclear fuel. The results of the first experiments are presented in this paper.March, 19661 DETERMINATION OF RADIOACTIVE FISSION-PRODUCT NUCLIDES 181 EXPERIMENTAL APPARATUS- The apparatus used for this experimental work is standard Hanvell2000 series equipment, with the exception of the head amplifier and distribution unit which are experimental and the subject of another paper.6 The schematic layout is shown in Fig.1. The input stage of Main Distribution Head unit amplifier amplifier 2153 A Cooled with ~~;~~~~~ Pulse-height Print-out jTH--t-iil analyzer unit 45 germanium detector FiFl generator Fig. 1. Schematic diagram of y-spectrometer the amplifier is an n-channel field-effect transistor* which is cooled to about 100" K together with the detector, a lithium-drifted germanium crystal, as shown in Fig. 2. The germanium detector is located on a copper rod which is immersed in liquid nitrogen, and the input stage of the head amplifier is in close proximity with the detector. The source is placed directly on the cold chamber envelope if suitable. F 7- /I H A = Liquid nitrogen B = Base plate C = Head amplifier D = Thermocouple E = Multiway cable F = Field-effect transistor G = Lid H = Vent I = Lid core j = Lid seal K = Cold chamber L = Crystal M = Platform N = Cap 0 = Tag-board assembly holding input components P = Dewar flask Q = Thermal shield R = Thermal rod S = Outer case Fig.2. Lithium-drifted germanium y-detector * Type 2N2346, made by Amelco Semiconductors Ltd., P.O. Box 1030, Mountain View, California, U.S.A.182 BANHAM et al. : GERMANIUM DIODES FOR GAMMA-SPECTROMETRIC [Analyst, Vol. 91 The advantage of operating the field-effect transistor at low temperature is that the input capacitance of the system can be reduced to a minimum by locating the input stage close to the detector. This reduction in capacitance in conjunction with lower capacity detectors can afford an improvement in resolution.A significant reduction in input noise is achieved by the fact that the thermal noise and gate leakage current noise are reduced to a minimum at this temperature. The noise performance of the head amplifier with single R.C. differentiation and integra- tion with 10-pF total input capacitance is 160 r.m.s. electronic charges. IVhen used in conjunction with a germanium - lithium detector (2 cm in diameter and 4 to 5-mm depletion depth') with a total input capacitance of 48 pF, the full width of the photopeak at half maximum value for cobalt-57 (122 keV and 136 keV) is less than 3 keV. This performance has been repeated over several months. A limitation in achieving a higher resolution is the over-all drift of the system. It has been shown that a spectrum stabiliser (type 2149) would achieve an improvement for counting periods greater than 15 minutes dependent on the over-all drift of the system.APPLICATION OF THE APPARATUS TO FISSION-PRODUCT ANALYSIS- The apparatus described above has been used to resolve the major problem of fission- product analysis, namely the interference of one nuclide with another, which conventional methods, with thallium activated sodium iodide detectors, are incapable of doing. DETERMINATION OF ZIRCONIUM-95 IN THE PRESENCE OF NIOBIUM-95- Zirconium-95 is used extensively for the determination of burn-ups in short irradiated samples. For the best results it is necessary to separate the zirconium chemically from the main interference, viz., the daughter-product niobium-95.The energies of the y-rays emitted by these two nuclides differ by only 40 keV and they are not even partially resolved by a sodium iodide detector. Zirconium-95, in fact, emits two y-rays at 724 and 756 keV, respec- tively, whereas niobium-95 emits only one y-ray at 764 keV. The spectra obtained with the conventional method (sodium iodide detector) and the lithium-difted germanium detector are shown in Fig. 3. In the spectrum obtained with the sodium iodide - thallium detector I I Fig. 3. y-Spectrum of niobium-95 + zirconium-96; curve A , sodium iodidc detector; curvc R, germanium detector only one photopeak is observed which is composed of the three photopeaks expected. These three peaks are clearly seen in the spectrum obtained with the germanium - lithium detector.The two zirconium-95 y-rays are completely resolved from each other, and the peak for niobium-95 at 764 keV is partially resolved from the peak for zirconium-95 at 756 keV. This spectrum demonstrates that it is thus possible to determine zirconium-95 in the presence of its daughter, niobium-95, without recourse to mathematical methods to calculate the proportion of each nuclide present. It is, in fact, possible to calculate the ratio of niobium-95March, 19661 DETERMIEATION OF RADIOACTIVE FISSION-PRODUCT NUCLIDES 183 to zirconium-95 from this spectrum, provided that the decay scheme is known. By using the decay scheme shown and the simple arithmetic which follows, this ratio was calculated. DECAY SCHEME OF ZIRCONIUM-% AND NIOBIUM-% 65 days MeV MeV MeV 0.764 MeV -7- Let- C, be the number of counts in the photopeak for zirconium-95 a t 724 keV, A, be the absolute abundance of the 724-keV y-ray, El be the efficiency of detection a t 724 keV.Then, if A, and E, are the corresponding absolute abundance and efficiency of detection for the 756-keV y-ray of zirconium-95, the number of counts in this photopeak, C,, is given by- If C, is the total count in the combined photopeak for zirconium-95 a t 756 keV plus niobiiim-96 a t 764 key, then the true count of the peak for niobium-96 a t 764 keV, C,, assuming that the absolute abundance of this y-ray is 100 per cent., is given by- If this photopeak is counted with an efficiency of E,, then the ratio of niobium-95 to zirconium-96 is given by- c, = c3 - c, C, x A2 x E, C, x E, R =- or The calculated value was 2-22, which compares well with the theoretical value of 2.18 for an equilibrium mixture, of which Fig.3 is the spectrum. The accuracy of this calculation is dependent on the accuracy of the decay scheme and efficiency of the detector for y-rays of different energies. MIXTURES OF RUTHEKIUM-103 AND RUTHENIUM-106- Ruthenium isotopes are rarely used as burn-up monitors although both ruthenium-103 and ruthenium-106 are potentially useful. The chemical properties of ruthenium are such that dissolution of a fuel specimen often results in the loss, by volatilisation, of an unknown amount of this element from the sample solution. Determination of the absolute amount of ruthenium is therefore of no value. However, it may still be useful to determine the ratio of ruthenium-103 to ruthenium-106 which will remain undisturbed although the total amount of ruthenium may be decreased.The fission yield of ruthenium-106 varies by a factor of more than 10 for the two fissile nuclides, uranium-235 and plutonium-239, whereas that of ruthenium-103 differs by about 2. It should be possible, therefore, to determine the relative proportions to the total number of fissions in specimens containing mixtures of184 [,4nalyst, Vol. 91 fissile nuclides, e g . , plutonium dioxide - uranium dioxide and plutonium carbide - uranium carbide fuels from the measured ratio of ruthenium-103 to ruthenium-106. Another case of interest is highly-burned-up low-enriched uranium, where significant quantities of pluton- ium-239 are produced from uranium-238 by neutron capture, and then undergo fission.BANHAM et al. : GERMANIUM DIODES FOR GAMMA-SPECTROMETRIC / I I I I - / I I I I I - / I I I I I - 1 I I I I I / I r z I 5 38 lteV L I \ \ \\A \ \ \ \ \ \ \ B ‘\ \ A50 500 550 600 650 I( Fig. 4. y-Spectrum of ruthenium-103; curve A, sodium iodide detector; curve B, germanium detector J I I / / I I I I I I / I / / --.’ -4 5 I 3 lteV I I I I 1 -150 100 550 6CO 650 I< V Fig. 5. y-Spectrum of ruthenium-106 -rho- dium-106; curve A, sodium iodide detector; curve B, germanium detector Portions of the spectra of ruthenium-103 and ruthenium-106 - rhodium-106 are shown in Figs. 4 and 5, respectively. The spectrum of a mixture of ruthenium-103 and ruthenium-106 - rhodium-106 is shown in Fig. 6.From the spectra of the pure nuclides, the ratios of the counts in the peaks for ruthenium-103 at 498 keV and 610 keV, and in the peaks for ruthen- ium-106 at 513 keV and 621 keV can be calculated. With this knowledge it is possible to determine the amount of each nuclide present in a mixture by mathematical treatment as follows- If R, is the ratio of the counts in the peak a t 498 keV to the counts in the peak for ruthcnium-103 R, is the ratio of the counts in the peak at 513 kcV to the counts in the peak for ruthenium-106 C5 is the total count in the mixed peak for ruthenium-103 a t 498 keV and ruthcnium-106 at 513 keV; C, is the total count in the mixed peak for ruthenium-103 at 610 keV and ruthenium-106 at 621 keV, then it can be derived that the number of counts due to the 498-lreV y-ray of ruthenium-103 in the mixed peak is given by- a t 610 keV; at 621 keV; 103 - R3 (C5 - c, x’ R,) From the spectrum shown in Fig.6 the value of Ci:: was calculated, and was within These values were 21,209 and 20,487, respectively. The R, - ri, c,,, - 3-5 per cent, of the expected value. ratio of ruthenium-106 to ruthenium-103 was 9.5. MIXTURES OF CAESIUM-134 AND CAESIUM-137- Caesium-137, like zirconium-95, is used extensively for the determination of burn-up, and has many advantages. One of these, its migration at relatively low temperatures, excludes its use with samples where the temperature of the sample is known to have been greater than 650” C during the irradiation. Where this is not the case, a second disadvantage may still cause severe limitation to its usefulness.Several isotopes of caesium are produced in high yield in fission, including the stable isotope caesium-133. This isotope has an appreciable neutron-capture cross-section which may be greatly enhanced under suitable reactor conditions by a very large resonance It does, however, suffer from two major disadvantages.March, 19661 DETERMINATION OF RADIOACTIVE FISSION-PRODUCT NUCLIDES 185 in its capture cross-section spectrum at about 6 eV. The product of this (n,y) reaction is caesium-134 (half-life of 2.3 years). This nuclide has a complex y-spectrum with y-rays of high abundance a t 605 keV and 796 keV that cause significant interference with the 662-keV y-ray of caesium-137 when the spectrum of a mixture is obtained by the conventional sodium iodide detector. The extent of this interference can only be determined with any degree of accuracy by involved mathematical treatment of the spectrum and it is, therefore, of interest to be able to resolve these two nuclides.Fig. 7 shows the y-spectrum of a mixture of caesium- 134 and caesium-137, in which the ratio of caesium-134 to caesium-137 is 2-75. The three peaks in question are all well separated from each other, and the determination of caesium-137 from this spectrum should present no difficulty. aJ 3 u c .- E L aJ Q VI c) K 3 0 V ' 0 3 R u Ru 3 keV 621 IkeV i I I I I I 450 500 550 600 650 I Fig. 6. y-Spectrum of mixed ruthenium- 103 and ruthenium-106 - rhodium-106 aJ 3 c u - E a L a, VI LI r 3 0 V / cs 134 106 keV , \ \ I / / ,/ 134 .cs 796 IkeV I I I 1 Pig. 7. y-Spectrum of mixed caesium-134 and cacsium-137; curve A,. sodium iodide detec- tor; curve B, germanium detector MIXTURES OF CERIUM-141 AND CERIUM-144- The two major y-active isotopes of cerium produced in fission are cerium-141 (half- life of 32 days), which emits a y-ray at 145 keV, and cerium-144 (half-life of 285 days) which emits a y-ray at 134 keV. The chemical properties of cerium-144, together with its half-life, combine to make it an attractive nuclide to use for the determination of bum-up. The presence of cerium-141 in specimens that have not been allowed to decay for long periods, however, complicates the determination to a considerable extent. It is possible to use the 2.18 MeV y-ray of its daughter praseodymium-144, which quickly reaches equilibrium with its parent cerium-144, but the low abundance of this y-ray, combined with the low efficiency of y-ray detectors at this high energy, makes its measurement difficult. Fig.8 shows the spectrum of a mixture of cerium-141 and cerium-144 obtained with a germanium - lithium detector. For comparison, spectra of the same mixture and of pure cerium-144 obtained with sodium iodide - thallium detector are also shown. The advantage of the gennanium - lithium detector over the sodium iodide - thallium detector is clearly demonstrated by the two completely resolved and almost perfectly symmetrical peaks which are obtained by this means. The resolution at 134 keV is 3.3 keV (the full width of the photopeak at half maximum height).186 BANHAM et d.: GERMANIUM DIODES FOR GAMMA-SPECTROMETRIC [AfldySt, VOl. 91 MIXED FISSION PRODUCTS- The preceding paragraphs have shown that the apparatus described is capable of resolving many of the problems in fission-product analysis that are difficult to solve by using conven- tional sodium iodide - thallium detectors. All the sources of activity used were separated chem- ically, and each contained only one interfering nuclide after separation. It would be of value to be able to resolve the nuclides of interest (zirconium-95, ruthenium-103, ruthenium-106, caes- ium-137 and cerium-144) from fission-product solutions without previous chemical separation. Three samples, representing different conditions of irradiation and decay, were examined in experiments to determine the feasibility of this approach.The spectra are shown in Figs. 9, 10 and 11. Eight peaks from isotopes of caesium, cerium, ruthenium, niobium and zirconium can be identified between 550 and 800 keV. Solutions were examined in all three cases, and the small amounts of ruthenium present resulted from the loss of this element by volatili- sation during dissolution of the specimen. In nearly all instances the photopeaks are well resolved and superimposed on a background that can be estimated with a good degree of certainty. It should therefore be possible to carry out fission-product analysis directly on fission-product solutions without any chemical separation, for a wider range of nuclides and irradiation and decay conditions than has hitherto been possible.I I I I ~ I0 120 130 140 I50 I< Fig. 8. y-Spectrum of mixed cerium-141 +- cerium-144 - praseodymium-144; curve -4, sodium iodide detector; curve B, germanium detector THE EFFICIENCY OF THE APPARATUS- 600 650 700 753 9’Nb 764 keV HOO keV Fig. 9. y-Spectrum of mixed fissicn pro- ducts; $ 2 days’ irradiation, 150 days’ decay The efficiency of the apparatus is defined by the equation- E = (% - CB:) x 100 per cent, A x D where C z is the total number of counts per second in a photopeak contained in channels m to n, CB; is the total number of counts per second background in channels m to n, A is the abundance of the y-ray in the decay scheme of the emitting nuclide, D is the disintegration rate of the source. It has been showng that the efficiency is strongly dependent on y-energy.The efficiency of the apparatus described was determined at 134 keV for cerium-144 and 662 keV for caesium-137 with standard sources. The values obtained were 0.16 per cent. and 0.0069 per cent., respectively, when the sources were placed as near to the detector as possible, i.e.,March, 19661 DETERMINATIOX OF RADIOACTIVE FISSION-PRODUCT NUCLIDES 187 directly on the cold-chamber envelope. These values are much lower, and the graph of efficiency versus energy of incident y-ray is much flatter than reported by other workers.6 These facts are probably attributable to a low geometry factor, due partly to distance and partly to the fact that the detector was inverted, thereby causing the y-rays to be incident on the back of the detector.The precision of these measurements is the same as the precision of counting, i e . , 1 per cent. on lo4 events recorded in the photopeak. The degree of accuracy, however, will be less good due to the errors in the absolute calibration of the standard sources (about 2 1 to 2 per cent.) and in the branching ratios in the decay schemes of the nuclides used (about k5 per cent.). Most of the sources used had been prepared previously for measurement on a conventional sodium iodide - thallium y-spectrometer with an efficiency of 10 per cent. at 134 keV, and were in the range 1 to 10 pC in activity. Although counting times with the germanium - lithium detector were usually 1 to 16 hours to accumulate about lo4 counts in <he principal photopeak, the apparatus, kven with this low efficiency, is great value to the radiochemist.of W c Y - 2 - E L W a Lo u C 3 0 U '06 Ru 621 keV cs 62 lteV I 3 7 95 Zr \ 1,756 keV' 695 keV 95 Nb 764 IkeV '3"s 756 keV I I I 650 700 750 800 I< Fig. 10. y-Spectrum of mixed fission pro- ducts; 193 days' irradiation, 750 days' decay W c Y - - E a - aJ Lo Y C 7 0 v 600 650 700 750 800 IkeV Fig. 11. y-Spectrum of mixed fission pro- ducts; 760 days' irradiation, 150 days' decay CONCLUSIONS An apparatus incorporating a lithium-drifted germanium detector, and designed primarily for studying low-noise amplifiers, has been used for a preliminary examination of the capabilities of this type of detector for y-spectrometry. The results presented here show that, even with poor geometry and low efficiency, the germanium - lithium detector is far superior to the best sodium iodide detector for resolving complex mixtures of y-ray emitting nuclides such as exist in fission-product mixtures. This high resolving power, coupled with the ability by electronic means to examine small parts of the y-spectrum in detail, should enable direct quantitative determinations of individual nuclides in intact or dissolved fuel specimens to be made.It should also be possible to scan irradiated fuel specimens, e.g. , plates or rods, for individual fission-product nuclides with188 BANHAM, FUDGE AND HOWES [Analyst, Vol. 91 little or no interference from other emitters present. This in turn should promote the study of, for example, the migration of fission products in irradiated fuel specimens in greater detail than has hitherto been possible. The study of the applications of germanium - lithium detectors to fission-product analysis is being continued with a second apparatus with improved geometry and efficiency, whilst the apparatus already described will continue to be in use for investigations of low- noise amplification systems. REFERENCES 1. 2. 3. 4. 5. 6. 7, 8. 9. Pell, E. M., J . AppZ. Phys., 1960, 31, 291. Freck, D. V., and Wakefield, J., Nature, 1962, 193, 669. Webb, P. P., and Williams, R. L., Nucl. Instrum. Meth., 1963, 22, 361. Tavendale, A. J., and Ewan, G. T., Ibid., 1963, 25, 185. Ewan, G. T., and Tsvendale, A. J., Can. J . Phys., 1964, 42, 2286. Howes, J. H., in preparation. Owen, R. B., and Gibbons, P. E., U.K. Atomic Energy Authority Refioyt, Harwell, AERE-M 1602, Banham, M. F., and Fudge, A. J., “The Determination of Burn Up in Nuclear Fuel Test Specimens Heath, R. L., and Cline, J. E., Report of Phillips Petroleum Company, Atomic Energy Division, Received September 132h, 1965 1965. using Zirconium-95,” to be published. Idaho Falls, Idaho, IDO-17050, 1964.