National Academies Press: OpenBook

Detection and Measurement of Nuclear Radiation (1962)

Chapter: Scintillation Methods

« Previous: General Introduction
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 4
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 5
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 6
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 7
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 8
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 9
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 10
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 11
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 12
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 13
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 14
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 15
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 16
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 17
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 18
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 19
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 20
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 21
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 22
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 23
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 24
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 25
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 26
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 27
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 28
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 29
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 30
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 31
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 32
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 33
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 34
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 35
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 36
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 37
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 38
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 39
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 40
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 41
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 42
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 43
Suggested Citation:"Scintillation Methods." National Research Council. 1962. Detection and Measurement of Nuclear Radiation. Washington, DC: The National Academies Press. doi: 10.17226/18670.
×
Page 44

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

II. SCINTILLATION METHODS* 1. Introduction One of the earliest methods for detection of charged particles involved counting the scintillations produced in a phosphor screen. Such a device was employed by Rutherford and his collaborators in their famous study of alpha-particle scattering by nuclei. The scintillation method eventually gave way to electrical counters, which were more reliable and capa- ble of functioning at high rates. Modern scintillation counters followed closely the develop- ment of high-gain photomultiplier tubes. Combining various scintillating materials with a photomultiplier to count the scintillations has resulted in the most versatile detector available for nuclear research. Present-day scintillation counters possess a number of advantages over gas counters: They not only indicate the presence of a particle, but also may be used to record the rate of energy loss, or the energy if the scintillator is thick enough. With the proper combination of scintillator and photo- multipler tube, the detector is ideally suited to high counting- rate applications. The high density of solid or liquid scintil- lators has made the scintillation counter the most efficient gamma-ray detector available. A functional diagram of a scintillation detector is shown in Fig. 1. The energy of the incident radiation is converted to light in the scintillator. The reflector and optical coupling ensure that this light is transmitted efficiently to the photocathode, where the light energy is converted to a burst of photoelectrons. An electrostatic focusing electrode collects the photoelectrons and focuses them on the electron multiplier structure. The electron multiplier increases the number of electrons by secondary electron emission from a cascade of elements called dynodes. The multiplication of a typical dynode is about 4; hence, the overall current gain for the standard, 10-stage multiplier is (4)10, or about 106. A variety of For general references on scintillation detectors and appli- cations, see Bell, Crouthamel, Mott and Sutton, Murray, and O'Kelley.5

ANODE OUT PHOTOMULTIPLIER TUBE 10 DYNODES OPTICAL COUPLING FOCUS ELECTRODE-O- ^ PHOTOCATHODE TRANSPARENT, FACEPLATE INCIDENT RADIATION REFLECTOR SCINTILLATOR Fig. 1. Diagram of a scintillation counter, illustrating schematically the way in which light from the scintillator is coupled to a photomultiplier tube. A typical wiring diagram is shown for the 10-stage photomultiplier operated with a positive high-voltage supply. 10-stage photomultiplier tubes are available, with overall gains of (0.6 to 5.0) x 106. When a burst of electrons arrive at the anode, the current which flows through R. yields a voltage drop which is coupled j-j to the measuring equipment through the blocking capacitor C . c This negative output pulse will generally have an amplitude of a few millivolts to perhaps a few volts. The rise time of the pulse, that is, the time for the pulse to rise from 10% to 90% of its maximum height, is determined by the lifetime of the excited state in the scintillator which emits the light, and by the time spread introduced by the multiplier. The time for the signal pulse to return to zero is determined by the product of

the net anode load resistance (RL in parallel with the input resistance of the amplifier) and the capacitance of the signal lead to ground. 2. Electron Detection and Spectrometry The scintillation method is very well suited to the de- tection of electrons. This subject will be considered at the outset, not only because of its intrinsic interest, but also as a necessary preliminary to a discussion of gamma-ray detection, in which photons are detected by the secondary electrons produced. A. Scintillators. Organic scintillators are best for spectrometry and counting of electrons and beta distributions, largely because of their low effective atomic number. A low atomic number reduces the probability of backscattering, in which an electron incident on a scintillator may scatter out, leaving only a fraction of its original energy in the scintil- lator. This effect is worst at low energies and for high- atomic-number scintillators. The organic scintillators have an additional advantage in that their gamma-ray sensitivity is low, so beta particles may be counted in the presence of moderate gamma-ray fields. To reduce gamma-ray interference, the mini- mum thickness of scintillator required for electron detection should be used. Further remarks on scattering and correction for gamma-ray background will be found below. It is convenient to divide organic scintillators into two classes: single crystals and solutions. Anthracene is typical of the single crystals, and because of its early popularity in the formative years of scintillation spectroscopy, it has become the standard against which other scintillators are usually compared. Characteristics of anthracene and a number of other typical organic scintillators are shown in Table 1. It will be seen that although anthracene yields the largest light output of any organic scintillator, its fluorescence decay is the slowest listed; therefore, applications which demand very rapid pulse rise times may require a scintillator with faster response at a sacrifice in pulse height. Although much useful work has been performed using scintillating crystals, the development of liquid and solid (plastic) solution scintillators has endowed scintillation

TABLE 1. ORGANIC CRYSTAL SCINTILLATORS (From Mott and Sutton, reference 3) Material Density, g/cm3 Relative Pulse Height for jS Excitation Wavelength T , of Maximum Emisgion, nsec A Anthracene 1.25 100 23 to 38 4450 trans-Stilbene 1.16 1.23 46 30 < 3 to 4. 8.2 3850 5 4000 p-Terphenyl p-Quaterphenyl •• 94 4. 2 4350 ar = time for decay to 1/e of initial light intensity; one nsec = 10~9 sec. counting with new scope. The liquid scintillators have the advantage of easy fabrication in almost unlimited volumes. Plastic scintillators are mechanically rugged and can be readily machined. Both liquid and plastic solutions can be obtained with fluorescence lifetimes as short as the best crystalline scintillators. Some examples of solution scintil- lators will be found in Table 2. TABLE 2. ORGANIC SOLUTION SCINTILLATORS (From Hayes, Ott, and Kerr, reference 6) Wavelength of Maximum Emisgion, A Solvent Primary Solute* (g/D Secondary Solutea (g/1) Relative Pulse. Height0 Toluene PPO(4)C POPOP(O.l) 61 - Toluene TP(4) POPOP(0. 1) 61 4320 Polyvinyltoluene TP(36) POPOP(l) 51 4300 Polyvinyltoluene TP(36) DPS(0.9) 52 3800 Solute abbreviations: PPO = 2,5-diphenyloxazole; TP = />-terphenyl; POPOP = l,4-di(2-(5-phenyloxazolyl))-benzene; DPS = />-/>'-diphenylstilbene. Pulse height for electron excitation, relative to anthracene pulse height as 100. "PPO is preferred for low-temperature applications because of the poor solubility of TP in cold toluene.

Solution scintillators may be made from two, or more often three, components. The bulk material is the solvent, in which a scintillating substance termed the primary solute is dis- solved; another scintillator, called the secondary solute or wavelength shifter, is also usually included. It is generally accepted that the incident particle first dissipates its energy by producing free electrons and ionized and excited molecules of the solvent. Then, most of the energy is transferred by nonradiative processes to the primary solute. The amount of energy so transferred depends on the overlap between the emission spectrum of the solvent and the absorption spectrum of the primary solute. Many of the primary solutes which may be used with the common hydrocarbon solvents fluoresce at such a short wave length that a conventional photomultiplier tube cannot efficiently make use of the light; for this reason, the second- ary solute is added to the solution. The absorption band of this latter solute should overlap the emission band of the primary solute. The final emission should be shifted to a longer wavelength which falls within the photomultiplier response and for which the absorption of light by the solution is small. The solute concentration in solution scintillators is rather low. As will be seen in Table 2, the primary solute concentration ranges from a few per cent for plastics down to less than one per cent for liquids. The required concentration of secondary solute is only a few per cent of that of the primary. It is a familiar and often distressing fact that the addition of certain impurities to liquid scintillators, even in small concentrations, decreases the light output enormously. Clearly, any additive which interferes with the energy transfer sequence will tend to quench the fluorescence. Such quenching will occur if the foreign substance has an absorption band at the emission wavelength for the solvent or one of the solutes and if the energy thus transmitted is dissipated by processes which do not give rise to useful light. The mechanism of quenching has been under study for some time in several labora- 789 tories and has been discussed in a number of review articles. ' ' The choice of a solvent for a solution scintillator depends upon the application. For liquid solutions xylene is 8

usually preferred, because it yields the greatest pulse height for a particular primary solute. On the other hand, toluene exhibits a much smaller absorption of the fluorescent light and is recommended where large volumes are required. Also, toluene does not react with some of the common light reflectors used in scintillation counters to the extent that xylene does. Other solvents are used in cases where it is necessary to introduce materials which are insoluble in toluene or xylene. For example, water is only slightly soluble in the usual scintil- lator solution, and it has a strong quenching effect. Aqueous solutions can be introduced by using />-dioxane as the solvent and then reducing the quenching effect with naphthalene. The details of liquid scintillator preparation, together with appli- cations to various problems, will be found in several reviews. ' Except for the fact that they are solid solutions, the composition and fluorescence properties of plastic scintillators are similar to those of liquid scintillators. The only important base plastics (solvents) in use are polystyrene and polyvinyl- toluene; some examples are given in Table 2. From the stand- point of convenience in machining and polishing, a plastic is to be preferred over anthracene; however, it should be borne in mind that the pulse height from a plastic is only about half that of anthracene, which leads to poorer resolution in electron spectroscopy. In counting applications and in spectroscopy at high electron energies, the plastics are highly recommended because of their convenience. B. Detector Arrangements. Several ways in which organic scintillators may be used are sketched in Fig. 2. In (a) is shown the simplest arrangement, a cylindrical crystal optically coupled to a photomultiplier tube. While this detector may be used for counting beta particles or electrons, it is not suit- able for measurement of low-energy spectra because of the back- scattering effect already described. Above about 1.5 Mev the backscattering probability is reduced to a point such that beta- ray spectra can be determined accurately with this simple detector, provided that the shape of the low-energy part of the spectrum is of no concern. The scattering problem can be circumvented in a number of ways. A helpful technique is to collimate the electron beam striking a flat scintillator surface to insure that the

PHOTOMULTIPLIER -SOURCE- ALUMINIZED PLASTIC FILM SCINTILLATOR REFLECTOR - /COLLIMATOR AND SUPPORT* PTICAL JOINT- PHOTOMULTIPLIER (a) (b) .SOURCE I ^-REFLECTOR xSCINTILLATORS OPTICAL JOINT— PHOTOMULTIPLIER LIQUID SCINTILLATOR CELL PHOTOMULTIPLIER (C) (d) Fig. 2. Mounting arrangements for electron detectors using organic scintillators. (a) Flat scintillator. (b) "Hollow-crystal" spectrometer, with electrons collimated into a well. (c) "Split-crystal" spectrometer, in which the electron source is sandwiched between two scintillators. (d) A counter using a liquid scintillator, in which the beta emitter is dissolved. 10

particles will enter the surface near normal incidence and so will tend to penetrate deeply into the crystal before scatter- ing. The "hollow-crystal" detector [see Fig. 2(b)] proposed by Bell reduces the backscattering contribution by collimating the electrons into a conical hole in the scintillator, from which the probability of escape is low for the scattered electrons. Hollow-crystal spectrometers have for several years been used to measure beta spectra, and they have consistently given improved performance over a flat scintillator, both as regards energy resolution and backscattering. ' A set of plastic scintillators machined and polished for use in hollow- crystal detectors is available commercially. Because all organic scintillators are somewhat gamma- sensitive, it is necessary to correct any beta spectrum data for the gamma-ray background, if the source is gamma radio- active. For the detectors sketched in Fig. 2(a), (b), this correction is obtained by interposing a beta absorber between source and detector and using the resultant gamma-induced spectrum for a "gamma background." Such a correction is only approximate, because the shape of the background spectrum obtained in this way is distorted by the contributions from bremsstrahlung and scattered photons produced in the absorber. Another method for reducing the consequences of scattering is to surround the source with scintillator so that no scattered electrons are permitted to escape. This may be achieved in the "split-crystal" detector of Ketelle,13 shown in Fig. 2(c). Here, the source is located between two scintillators arranged so that electrons scattered by one crystal are detected by the other. The nearly 4n geometry makes it difficult to measure the gamma-ray background with an absorber and leads to a high probability that gamma-induced counts may sum with pulses from coincident beta particles to yield a very confusing spectrum. Excellent spectra of "inner" beta groups have been measured with a split-crystal detector in coincidence with 14 gamma rays by use of beta-gamma coincidence techniques dis- cussed in Section VI.7. At low beta-particle energies (below about 200 kev), all of the methods described so far become rather difficult to Nuclear Enterprises, Ltd., 550 Berry Street, Winnipeg 21, Manitoba, Canada. 11

apply, because the response is somewhat sensitive to the treat- ment of the scintillator surfaces, and the source must be very thin to be free from scattering and absorption effects. These obstacles may be overcome by adding the radioactivity directly to an organic liquid scintillator, and thus making the source an integral part of the scintillator [Fig. 2(d)]. This is the basis of liquid scintillation counting, a method which has found widespread use in chemistry, particularly in tracer experiments using low-energy beta emitters such as C14, S35, Ca45, and H3. In some experiments it may be difficult to find a chemical form of the radioactive material which will not also quench the fluorescence. Liquid scintillation counting, in- cluding such special topics as the use of suspension and gels, is a rather specialized technique, and will not be treated further here; however, the interested reader will find useful information in some of the review articles ' ' and in literature published by manufacturers of liquid scintillation counting equipment. In addition to the general arrangements discussed above, there are a few other counting techniques which make use of the versatile plastic scintillators for special applications. One such technique for the assay of solutions is the use of scintillator beads proposed by Steinberg. The detector is similar to that of Fig. 2(d), except the container is filled with small beads of plastic scintillator, instead of a liquid scintillator. The solution whose radioactivity is to be determined is poured into the container; the liquid fills the interstices between the beads, and thus puts the liquid and solid phases in intimate contact. A suitable solution must be transparent to the light from the scintillator and must not attack the plastic beads; although these solution requirements somewhat restrict the use of the technique, the preparation of a sample can be extremely rapid, and the beads can be washed and used repeatedly. Plastic scintillators may be conveniently fashioned into many other shapes. Plastic scintillator dishes for containing radioactive liquids may be mounted directly on a photomultiplier Suitable beads are: "B-Beads," manufactured by Pilot Chemical Co., 39 Pleasant Street, Watertown 72, Massachusetts; or NE 102 Spheres, obtainable from Nuclear Enterprises, Lts., 550 Berry Street, Winnipeg 21, Manitoba, Canada.

tube for counting. Capillary tubing made from plastic scintil- lator can be wound into a spiral and attached to a photomulti- plier to make a very simple flow counter for beta-radioactive gases and liquids. Devices of this sort are available com- mercially. C. Electron and Beta Spectrometry. In addition to their use as counters, these organic scintillators are useful for determining electron and beta-ray energies. The consensus of the available experimental information indicates a linear pulse height-energy curve down to a low energy of -100 kev; below this energy the response is also nearly linear, but with a slightly different slope.17 The response of an organic scintillator to monoenergetic electrons is mainly a gaussian peak whose width varies inversely with the square root of the energy. This energy dependence is predicted from the statistical variation in the number of photo- electrons at the photocathode and the electron multiplication processes within the photomultiplier tube; hence, the conti- bution to the peak width by the scintillator itself is small. For an anthracene hollow-crystal spectrometer a resolution (full width at half-maximum counting rate) of about 10% can be achieved at 624 kev, but if a plastic scintillator is used the resolution is only about 14%, because of the lower light output of the plastic. Although the resolution of the electron scintillation spec- trometer is poor compared with that of a magnetic spectrometer, the scintillation method has some appealing features. Used with a multichannel pulse-height analyzer (£f~. , Section VI.6.B., below) to display the distribution of pulse height (« energy), it is possible to record the entire beta spectrum in a single counting interval. On the other hand, the magnetic spectrometer is a single-channel instrument, which can record only a single point on the spectrum at one time. The advantage of the scintillation spectrometer in studies of rapidly decaying sources is obvious. Further, the required scintillation de- tector is relatively inexpensive. Such simple spectrometers should find increased use in the analysis of mixtures of pure beta emitters and in distinguishing between tracers which possess similar gamma-ray spectra. Nuclear Enterprises, Ltd. 550 Berry Street, Winnipeg 21, Manitoba, Canada. 13

Before a careful analysis of the beta spectrum shape can be made, the pulse-height distribution must be corrected for finite instrumental resolution. This is especially important for beta-ray endpoints below about 1 Mev. Corrections for finite resolution can be made by using the method of Owen and 18 Primakoff, and assuming that the scintillator response is a gaussian whose width varies as E"1/2. It was shown by 19 Freedman, et al., that such a procedure corrected the spectrum near the maximum beta energy but did not account for the excess of events at low energies. This excess counting rate arises from scintillator backscattering, which is never completely eliminated in a low-geometry arrangement [for example, the detectors of Fig. 2(a) and (b)]. The typical response of a flat anthracene spectrometer to the electrons and beta rays from a Cs137 source is shown in Fig. 3, which includes a spectrum due to the internal conversion peak alone. It is seen that the backscattering "tail" is almost flat, 19 and is about 6% of the peak height. Freedman, et al., developed an iterative method to correct the experimental data for both backscattering and resolution effects. A comparison is made in Fig. 4 between two Fermi plots of the low-energy beta group of Cs13 7, The upper curve shows the result obtained when the resolution distortion correction alone is applied, and the lower curve shows the improvement in the quality of the low-energy data when corrections are made for both resolution and backscattering. The conventional correction procedures break down for beta groups below about 100 kev. This is because the resolution width becomes so large that most of the counts which appear to arise from events in the high-energy portion of the spectrum are actually due to low-energy electrons which fall within the detector resolution. Some idea of the resolution width at low energies can be gained by recalling that a good organic scintil- lation spectrometer with a resolution of 14% at 624 kev (Ba137m conversion electrons) will exhibit a resolution of about 50% at 50 kev. These difficulties make it advisable to employ empirical corrections derived from data on low-energy beta emitters whose energies and spectral shapes are well known. 14

100 200 400 600 800 PULSE HEIGHT 1000 1200 Fig. 3. Spectrum of a Cs1 3 7 source, measured on a flat anthracene crystal. The internal-conversion electron line at 624 kev and the continuum from the 523-kev beta group are shown. When the beta rays and electrons are stopped in an absorber, the background spectrum from the 662-kev gamma ray is obtained. A coincidence between 624-kev electrons and their associated X rays excludes the beta and gamma spectra, and leaves only the electron line and its scintillator back- scattering spectrum. 15

CORRECTIONS TO SCINTILLATION SPECTRUM OF Cs137 o o CORRECTED FOR 0 RESOLUTION ONLY • CORRECTED FOR RESOLUTION AND SCINTILLATOR BACKSCATTERING 0.1 0.2 0.3 ENERGY (Mev) 0.4 0.5 Fig. 4. Fermi plots of the low-energy beta-ray group of Cs1 3 7 . The upper curve was corrected for the unique spectrum shape and the scintillation spectrometer resolution. The lower curve shows the improvement obtained when the correction for scintillator backscattering is included (Gardner and 3. Gamma-Ray Counting and Spectrometry A most important contribution of the modern scintillation technique has been made in the field of gamma-ray detection. The much higher density of solid gamma-ray scintillators gives them a stopping power (i.e., detection efficiency) for photons far greater than gas-filled counters. It is now quite feasible to prepare scintillating crystals large enough to stop com- pletely a sizeable fraction of incident gamma rays; thus it is possible not only to count gamma events, but also to measure 16

energy spectra and gamma-ray intensities as well. The reali- zation of these possibilities has put in the hands of the research chemist a versatile, precise tool. A. Scintillator Considerations. To be effective for gamma-ray detection, a scintillator should be of high density and high atomic number; these requirements are best satisfied by the inorganic scintillators. Although there are many scintillating inorganic materials, only the activated alkali halides can be grown in single crystals of sufficient size and yet possess the required transparency to their emitted light. Sodium iodide, activated with 0.1% Til, is the only alkali halide scintillator in routine use. It has the high density of the alkali halides and has a moderately high effective atomic number. The light output in Nal(Tl) per Mev is the largest of any known scintillator and is about twice that of anthracene. Large single crystals of Nal(Tl) are readily obtainable and are highly transparent to their own fluorescent light, which is o o emitted in a band about 800 A wide, centered at 4100 A. This wavelength is quite compatible with the response of standard photomultipliers having an S-ll response. (See the discussion of photomultipliers in reference 3.) The fluorescence decay time is 0.25 psec, comparatively short for an inorganic crystal. Because of its higher effective atomic number, thallium- activated cesium iodide has been investigated as a gatnma-ray scintillator. At present, crystals of CsI(Tl) are far more expensive to manufacture than crystals of Nal(Tl). Further, although moderate pulse-height resolution can be obtained, the usable light output is only about 40-45% that of Nal(Tl). This lower apparent output may arise because the fluorescent light o is emitted at longer wavelengths, namely 4200-5700 A, and conse- quently cannot be measured efficiently by an S-ll photomulti- plier tube. Improved pulse height could probably be attained by using a photomultiplier tube with better response in the red, such as the low-noise, multialkali-cathode tubes. The decay time of the fluorescent light from CsI(Tl) is 1.2 usec, which is rather long for many applications. B. Mounting Sodium Iodide Crystals. The method chosen for mounting a Nal(Tl) crystal on its photomultiplier tube involves a consideration of the deliquescence of the crystal, its opti- cal properties, and the necessity for avoiding gamma-ray scattering. For some uses the two latter considerations may be 17

relatively unimportant, but in all cases it is necessary to take great care that the crystal surfaces are not exposed to moisture. The usual procedure is to prepare a crystal for mounting in a dry-atmosphere box; once mounted, the crystal enclosure should contain a dry atmosphere or else be evacuated. Whenever the largest possible light output is required, as in gamma-ray spectrometry, it becomes very important to gather the light and transmit it to the photomultiplier tube with the least possible attenuation. This is especially difficult when Nal(Tl) crystals of refractive index 1.77 must be optically coupled to glass faceplates with a refractive index of 1.5, because, if the surfaces are polished, much of the light tends to be critically reflected back into the crystal. The use of diffuse reflection at the crystal surfaces has been found to give higher and more uniform light output than specu- lar reflection, since the probability is increased for light to be reflected onto the exit face of the crystal within the critical angle. The diffuse reflector surface is formed on the crystal by grinding all crystal surfaces, including the exit face. Any light escaping from the other crystal surfaces should also be returned by a diffuse reflector such as magnesium oxide or ,.-ulumina If the Nal(Tl) detector is to be used as a spectrometer, it is important that any material surrounding the crystal be very thin. Otherwise, the gamma-ray spectrum will be distorted by Compton electrons and degraded gamma rays. A crystal mount which used an enclosure made of 5-mil aluminum, coated on the inside with a thin optical reflector of cr-alumina, was devised by Bell and co-workers for use where crystal and photomultiplier were of approximately the same diameter. Detailed procedures for fabricating the metal can and applying the reflector have been published. Figure 5 shows a crystal mount which is more easily mass-produced, and uses magnesium oxide powder as the diffuse optical reflector. Similar crystal-photomultiplier assemblies with good resolution are also available commercially. The light transmission from the crystal to the photo- multiplier should be as efficient as possible; for this reason Harshaw Chemical Company, 1945 East 97th Street, Cleveland 6, Ohio. 18

^- '/^ in. PACKED MgO 3X 3-in. NoI(Ti) CRYSTAL SEE DETAIL (I BELOW Al FOIL MgO 5-mil Al FOIL CAPSULE in. PACKED MgO DC 200, 35 XI06c«ntijtoK«s Fig. 5. Integral crystal mounting arrangement for a 3 x 3-inch Nal(Tl) crystal and 3-inch photomultiplier tube (V. A. McKay, Oak Ridge National Laboratory). 19

the crystal generally should be attached to the phototube by a single optical joint. In noncritical applications, such as simple counting, the loss of light from an additional optical seal and window may not be important. Here, an assembly is often used consisting of a crystal enclosed with its reflector in a thin can and optically coupled to a transparent window. Various crystal assemblies may then be attached to the same photomultiplier tube as required. Examples typical of such crystal mounts are shown in Fig. 6. Fig. 6. Sealed Nal(Tl) crystal assemblies for a variety of applications. Transparent windows are provided for optical cou- pling to photo-multiplier tubes (Harshaw Chemical Company). Mounting Nal(Tl) crystals larger than about 3 inches in diameter calls for a more elaborate technique. Such large crystals are very heavy, and the mechanical problems involved in constructing a large detector make it necessary to relax somewhat the requirements of a thin container. The response of a large crystal is not as sensitive to scattering from the container walls as are the smaller crystals, for which the mass ratio of Nal(Tl) to cladding material is much less favorable. C. Special Counting Problems. Whenever gamma-emitting nuclides are used in radlochemistry some variation of the versatile Nal(Tl) scintillation counter is nearly always used. It will be the purpose of this section to set down a few of the uses to which these counters have been put. Perhaps the most generally useful configuration is the well counter (see Fig. 7). Various sizes are available; an 20

10-mil Al WELL 15-mil Al CAP PACKED MgO Al CAPSULE CRYSTAL NP-185 EPOXY NP-475 EPOXY (SOFT) BLACK TAPE R-313 EPOXY DC 200. 2.5 X I06 centistokes PHOTOTUBE SEALER RING 0.082-in. COPPER EVACUATION TUBE Fig. 7. An integral mounting arrangement for a 2 x 2-inch Nal(Tl) "well-type" crystal on a 2-inch photomultiplier tube (V. A. McKay, Oak Ridge National Laboratory). 21

inexpensive, yet efficient, detector is a Nal(Tl) crystal, at least 2x2 inches, with a 3/4-inch diameter well about 1-1/4 inches deep. Once the crystal has been mounted in a thin aluminum enclosure and a protective liner has been inserted to avoid the chance of permanent contamination of the crystal can, about 1/2 inch of the original diameter will be left for insertion of samples. The well counter is the most sensitive gamma-ray detector available and certainly one of the most convenient. To prepare a sample may involve only the transfer of a few ml of solution to a small test tube. Because of the penetrating nature of gamma rays, self-absorption in such sources is small, and further treatement of the sample to reduce the mass is usually not necessary. Commerical well-crystal counters are available which provide automatic sample changing and count recording for many samples. In certain cases it may be desirable to assay large volumes of solution directly. A variation of the once-popular immersion counter is sketched in Fig. 8; if a 3 x 3-inch crystal is surrounded by solution as shown, samples of several liters can be accommodated. A more compact assembly uses a 3/4-inch diameter Nal(Tl) crystal, 3 inches long. The manu- facturer states that the efficiency of this detector is one- fifth that of a typical well-crystal with 5-ml capacity, but has a sample volume 30 times greater. The increasing need to make measurements of flow systems has led to several modifications of the scintillation counter for use with both liquids and gases. An excellent flow counter can be made from a length of plastic tubing either laid across the face of a Nal(Tl) detector, wrapped around it, or the tubing may be doubled into a U and inserted into a well crystal. Crystal packages are also available with a hole drilled through along a diameter, so that a tube passing through the crystal is surrounded by Nal(Tl); an example is included in Fig. 6. Other special methods which involve spectrometry rather than integral counting will be discussed in the following section. D. Gamma-Ray Spectrometry. One of the most important applications of the Nal(Tl) scintillation detector is in the Atomic Accessories, Inc., 813 West Merrick Road, Valley Stream, New York. 2.2.

.SAMPLE ( LIQUID OR POWDER) isWK^iWli^ -:^Kft;W::%'^5S:v:::;;-::: yvY --:::;.,- S SK'-A^vftyifS;; 3x3-in. NoI(TI) CRYSTAL PHOTOMULTIPLIER TUBE Fig. 8. Illustration of a way to obtain high counting efficiency for a scintillation counter used with large counting volumes. field of gamma-ray spectrometry. Now that large, clear crystals of Nal(Tl) and photomultiplier tubes with uniform high-efficiency photocathodes are available, it is possible to make a spectrometer which will not only measure energies of gamma rays to high precision but also yield their intensities. Much of the popularity of the Nal(Tl) scintillation spectro- meter in radiochemistry lies in its ability to differentiate between various gamma-ray components; hence, the presence of a particular nuclide in the spectrum of a mixture can be established by the characteristic energies observed, and the 23

amount of the nuclide can be determined quantitatively from the appropriate gamma-ray intensities. Interaction of Gamma Rays in Nal(Tl). It should be recalled that gamma rays as such are not detected, but rather it is the secondary electrons produced by the interaction between the gamma rays and the crystal which give rise to the fluorescent light. Thus it is appropriate to discuss briefly the three processes (photoelectric effect, Compton effect, and pair pro- duction) by which gamma rays interact, in terms of their effects on the response of the scintillation spectrometer. Partial absorption coefficients in Nal for these processes are shown in Fig. 9. Below about 100 kev, the photoelectric .1 o PHOTOELECTRIC COMPTON TPAIR PRODUCTION Multiply Ordinate by 100 PHOTOELECTRIC PHOTOELECTRIC PAIR PRODUCTION I I i I I I I I I i I I I 10' ENERGY, Kev Fig. 9. Gamma-ray absorption coefficients in Nal(Tl) for various gamma-ray energies (Bell1). 24

* SMALL CRYSTAL (1V2 xlin.)- LARGE CRYSTAL (3x3in.)- Fig. 10. Schematic representation of gamma-ray inter- interactions within Nal(Tl) crystals of two sizes.

effect is by far the most probable; however, the photoelectric effect shows such a rapid decrease in absorption coefficient with increasing energy, that the Compton effect is left as the dominant process in the intermediate-energy region. Pair pro- duction, which sets in at 1.02 Mev and increases rapidly in probability thereafter, becomes the most important of the processes at very high energies. In all of these processes, it is predominantly the iodine of Nal which, because of its high Z, interacts with the gamma rays. An attempt to demonstrate qualitatively the practical importance of the effects listed above is shown in Fig. 10. The low-energy gamma ray yi undergoes a photoelectric encounter within the first 1/8-inch of material through which it passes. The photoelectron, which has a very short range, is stopped in the Nal(Tl) and gives up its energy to the crystal. The iodine atom which released the electron is left with a vacancy, most probably in its K shell; the act of filling this vacancy yields a K X-ray. Usually the iodine X-ray is captured by the crystal, for its energy is only about 28 kev; however, since low-energy gamma rays are always stopped near the crystal surface, there exists a small but significant chance that the X-ray may escape the crystal surface entirely, and give a peak whose energy corresponds to the gamma energy minus the 28 kev of the X-ray. Above a few hundred kev, multiple processes play an important role, and so it becomes necessary to take into account the crystal size in addition to the various absorption coefficients. In Fig. 10, y2 illustrates a Compton scatter by an intermediate-energy gamma ray. The Compton electron e is stopped and yields a light pulse proportional to the electron's kinetic energy; on the other hand, if the crystal is small the scattered photon y2' may not be stopped, and its energy will then be lost. The figure shows that in a larger crystal, further Comptron processes may occur until the energy of the scattered photon is reduced to an energy so low that a photo- electric event finally transfers the remaining gamma energy to the crystal. It is important to bear in mind that the stepwise process just described occurs very rapidly, compared with the speed of present-day electronic instruments; therefore, the interactions of y2 in the 3 x 3-inch crystal of Fig. 10 would give rise to a single electrical pulse whose height would correspond to the total energy of y2. 26

Pair production introduces a very complicated response, as illustrated by the case of ya in Fig. 10. A high-energy gamma ray forms a positive and negative electron pair which carry off as kinetic energy the original gamma-ray energy, minus the 1.02 Mev (two electron rest masses) required to create them. The two short-range electrons stop, and their kinetic energy is acquired by the crystal. The positron annihilates, forming two photons, each with an energy of 0.51 Mev (m0c2) and correlated at 180 . Once pair production occurs, the response depends on the probability that the annihilation photons will be captured. The example in Fig. 10 shows that in a small crystal, the probability is greatest for the escape of both photons; in a larger crystal, it is more likely that at least one of the photons will be stopped. To summarize, then, the pair-pro- duction response leads to three peaks in the pulse-height distribution: the full-energy peak, which corresponds to the capture of all the incident gamma-ray energy by the multiple processes; the "single-escape" peak, which signals the loss of one annihilation photon; and the "double-escape" peak, which indicates the loss of both annihilation photons. Typical Gamma-Ray Spectra. The effects just described are illustrated by some representative gamma-ray spectra in Figs. 11-13. More detailed explanations will be found in references 1, 2, 5, and 20. Because the light output from Nal(Tl) is very nearly linear with respect to the energy deposited, the distribution in the height of the pulses from various Nal(Tl) scintillation spectrometers will be treated as energy spectra. All of the gamma-ray spectra described below will have semilogarithmic intensity scales and linear pulse- height (or energy) scales. This has the decided advantage that the wide range of counting rates encountered in gamma spectro- metry can be easily accommodated; further, spectral shapes can be compared by superimposing two spectra plotted on the same log paper with identical energy scales, even though the absolute heights of the peaks may be very different. A typical spectrum from a low-energy gamma ray is shown in Fig. 11. Although a large peak is present, arising from 22-kev X rays in the sample, let us direct our attention to the gamma-ray peak at 88.5 kev. Nearly all of the events in this peak are from the photoelectric effect near the crystal surface (cf~., Y1 °f Fig. 10); a peak about 28-kev lower is due to 27

3x3-in. Nal(TI) h = 10cm Iodine X-Ray Escape 40 50 60 70 PULSE HEIGHT 90 100 Fig. 11. Spectrum of 87.5-kev gamma rays and 22-kev X rays from a Cd109 source, illustrating the phenomenon of X-ray escape following detection of 87.5-kev gamma rays.

o o LJ or LiJ Q- o o 0.662 Mev BACKSCATTER PEAK l*!n.DIa.x1 in. Nal(TI) -2in.x 2 in. Nal(TI) COMPTON DISTRIBUTION^ -3 in.x 3in.- :NaI(TI) 200 400 600 800 1000 PULSE HEIGHT 1200 with Fig. 12. Spectra obtained by measuring a Cs137 source Nal(Tl) spectrometers of three crystal sizes (Heath ) escape of the iodine K X-rays. As the gamma-ray energy increases, the photons penetrate more deeply into the crystal before they undergo photoelectric absorption, and so the proba- bility for X-ray escape diminishes. In addition, because the energy separation between the full-energy peak and the X-ray escape peak is a very small fraction of the gamma energy, the X-ray escape phenomenon is not observed above about 170 kev.

A comparison is sketched in Fig. 12, which shows the spectra obtained at 0.662 Mev with Nal(Tl) crystals of differ- ent sizes. All of the spectra are normalized at the maximum of the full-energy peak. The smallest crystal yields a charac- teristic distribution below the main peak which results from an event in which a Compton-scattered photon is lost and the Compton electron is captured (cf~., Fig. 10). It will be noted that as the crystal size increases, the probability for multi- ple processes also increases; this is manifested in an increase in the fraction of events falling within the main peak. Figure 12 shows that the ratio of the height of the full-energy peak to that of the Compton distribution is nearly twice as great for a 3 x 3-inch as for a 1-1/2 x 1-inch crystal. Of course the more nearly the response approximates a single peak for a single gamma-ray energy, the more useful the spectrometer becomes. The complexity of the spectrum when pair production is involved may be seen in the spectrum of Na24 shown in Fig. 13. Two gamma rays are present in this source at 1.38 and 2.76 Mev. The full-energy peak and the two pair peaks stand out clearly in the high-energy portion of the spectrum. Note that in the smaller crystal there are relatively few multiple events lead- ing to counts in the full-energy peak; in fact, the double- escape peak is the most intense of the three. This response may be contrasted with that of the 3 x 3-inch crystal, in which the contribution from multiple events has made the full-energy peak the most intense. Further, the probability of double escape is quite low. It may be of interest to note that the spectrum of the 1.38-Mev gamma ray shows no evidence of pair peaks; in practice, the effect of pair production is not detectable below about 1.5 Mev, even though the threshold falls at 1.02 Mev. Environmental Effects. The gamma-ray spectra measured in a given situation will be complicated, in addition to the ele- mentary interactions just described, by several important environmental effects. While space does not permit a complete treatment of such spurious responses here, a report by Heath includes a valuable analysis of various experimental factors; some of Heath's findings as well as other related data will also be found in reference 5. In the discussion which follows it will be convenient to 30

n-SINGLE ESCAPE 1/2 Xl-in. Nal (Tl) AT 2.5 cm 3X3-in. Nal (Tl) AT 9.3 cm Na24 GAMMA SPECTRA 200 400 600 800 1000 1200 Fig. 13. Gamma-ray spectra of Na24, using 1-1/2 x 1-inch and 3 x 3-inch Nal(Tl) spectrometers. 31

refer to Fig. 14, which shows a typical arrangement for a 3 x 3-inch Nal(Tl) detector situated in a Pb shield. One of the most persistent experimental difficulties is scattering, whose consequences may take various forms. It is easy to show that, for Compton scattering at large angles, the energy of the scattered photon is nearly independent of the incident gamma-ray energy, and attains an almost constant value around 200 kev. Thus, large-angle scattering from shield walls, source holder, or other matter in the vicinity of the source (cf~., Fig. 14) will be manifested as a peak at about 200 kev. This peak is generally called the backscatter peak. It may be reduced by making the inside dimensions of the shield very large, thus decreasing the geometry between the detector and shield walls. Heath demonstrated that a shield made from lead yielded a much smaller backscatter peak than one made from iron. Another form of scattering arises from the beta-ray absorber which is usually placed between the source and the detector to stop beta particles or electrons from the gamma- ray source. Because of the geometry involved, the scattering is restricted to small angles, and so the scattered photons detected in the Nal(Tl) crystal are only slightly reduced in energy. This has the effect on the gamma spectrum of filling in the "valley" between the Compton-electron distribution and the full-energy peak. Secondary radiation from the shield walls may cause serious complications. If bare lead walls are used in a spectrometer shield and a source of low-energy gamma rays is inserted, then fluorescent lead X-rays are emitted from the walls and are detected by the Nal(Tl) crystal, causing a spuri- ous 72-kev peak in the gamma spectrum. The mechanism for X-ray production is similar to that discussed above for X-ray escape from Nal(Tl) crystals. The best way to remedy this situation is to cover the lead surfaces with a sufficient thickness of a medium atomic number material, usually Cd, to attenuate the Pb X-rays to a negligible level; the Cd is covered in turn by a thin veneer of Cu to absorb any fluorescent radiation from the Cd. When very intense high-energy gamma rays are present in the source, it is common to observe a peak at 0.511 Mev in the gamma-ray spectrum. This peak is due to pair production in 32

3-in. x 3-in. NoKTl) CRYSTAL AND PHOTOMULTIPLIER Fig. 14. Cross-section of a typical scintillation spectrometer installation, showing the 3 x 3-inch Nal(Tl) detector assembly, the lead shielding with "graded" liner, and the use of a low-mass support for the source and beta absorber. The origin of scattered photons is illustrated. 33

the Pb shield walls, with escape of annihilation radiation. Just as in the case of environmental scattering, the secondary radiation from the photoelectric effect and from pair pro- duction may be markedly reduced by increasing the separation between the shield walls and the source-detector combination. Internal bremsstrahlung produced in the source and external bremsstrahlung emitted when beta rays are stopped in the absorber will be detected just as any other electromagnetic radiation. Therefore, when the number of gamma rays per beta disintegration is low, a prominent contribution from bremsstrah- lung will be noted. Such an effect is shown as an upturn at low energies, with much the same shape as a decreasing expo- nential function added to the gamma-ray response (see Fig. 15). Analysis of Gamma-Ray Spectra. The gamma-ray scintillation spectrometer has proved to be an important tool for the quanti- tative determination of gamma intensities. Since the true shape of the Compton electron distribution for a single gamma- ray energy is so obscured by the spurious effects which have just been described, the area of the full-energy peak is generally chosen as the basis for intensity measurements. The spectrum exhibited by a source which emits gamma rays of sever- al energies will be a summation of the responses to the indi- vidual gamma rays. The process by which accurate intensities may be obtained involves first breaking down the gross spectrum into its components ("spectral decomposition"), from which the areas of the full-energy peaks may then be extracted. The above discussion demonstrates that the pulse-height distribution of a Nal(Tl) spectrometer arising from the inter- action of a single incident gamma-ray energy contains not just a full-energy peak, but in addition a complicated spectrum down to zero energy. In the course of performing a spectral decompo- sition, it is essential that the complete spectrum shape be used, and not just the full-energy peak. The detailed shape of the spectrum from a single gamma ray, or from a particular sample, is often called the response function. The decomposition is relatively straightforward if the gamma-ray spectrum in question happens to be made up of gamma- ray components whose spectra can be determined individually. In such cases, the procedure to be followed simply involves normalizing the response function for the most energetic gamma 34

100 UJ 5 cr CD O o 0.1 91 58-day Y 3"x3" No I Absorber 1.34g/cm^ Source Dist. 10cm -BREMSSTRAHLUNG 0 200 400 600 800 PULSE HEIGHT 1000 1200 Fig. 15. Gamma-ray spectrum of 58-day Y91, showing the bremsstrahlung spectrum characteristic of a source for which the beta-to-gamma intensity ratio is very large. (From Heath .) 35

ray to the experimental points at the full-energy peak; the response function is then drawn in and subtracted from the experimental data. The most energetic peak in the residue is fitted to the response function for that energy, and the sub- traction process is repeated until all components have been "peeled off." An example of this process is shown in Fig. 16. It should be emphasized that the response functions must be determined under conditions identical with those under which the unknown was measured, so the response function used in the analysis will include the same spectral distortions, such as backscatter peaks, which affected the unknown. A frequently overlooked effect is the variation in gain with counting rate, which may occur in certain photomultiplier tubes and multi- channel analyzers; for this reason it may be necessary to adjust the energy as well as intensity scales before attempting a point-by-point subtraction. There are commercial devices which permit an adjustable fraction of a standard spectrum to be sub- tracted from a pulse-height distribution stored in a multi- channel analyzer memory. When using a system such as this, it is particularly important that no serious gain shifts occur with changes in counting rate. It may not be possible to measure directly the response functions for the component gamma rays of an unknown spectrum; in this case, it is necessary to synthesize the required function from a measurement of gamma-ray standards over the energy range of interest. The full-energy peak is nearly gaussian, except on its low-energy side where the Compton spectrum contributes a slight distortion. By the use of standard spectra, the width parameter for the full-energy peak can be plotted as a function of energy, and values for the unknown can be evaluated. Other features of the spectrum can be constructed by interpolating on plots which correlate the coordinates (pulse height and counting rate relative to that for the full-energy peak) of various "key" points of the spectrum with the gamma-ray energy. Some of the key points which may be used are: the backscatter peak; the level, peak, and inflection of the Compton distribution; and the valley below the full-energy peak. If pair peaks are involved, their vital statistics must, of course,be included. Once a particular full-energy peak has been resolved from the spectrum, the area under the peak P(y) may be obtained by summing the counting rates of the channels which contain the 36

peak, or by means of the equation 0.56(AE) ' where a is the half -width of the gaussian peak at h/e; h is the peak height in the same units as P(y) ; and AE is the channel width. Both a and AE must have the same units and typically may be expressed in pulse-height divisions, channels, or kev. Use of Computers in Gamma-Ray Analysis. The decomposition of a very complex gamma-ray spectrum by the process just described is extremely tedious if all subtractions are performed manually, point by point. Although much useful work can be done in this way, hand calculations are limited to a rather small volume of data, and so digital computers promise to be very useful in the decomposition of complex spectra. A situation in which the computer can, in principle, be most readily applied to the spectral decomposition problem is that for which the experimentally measured spectrum is made up of components whose spectra may be determined individually. This condition is often met in radioactivation analysis, and Kuykendall and Wainerdi ' have investigated the use of a digital computer in an automatic activation analysis system. Their programs require a library of standard spectra and will not identify any gamma-ray peak whose response function does not appear in the library. Two computer programs have been written: (1) The half -life and gamma-ray energy are used to identify each statistically significant peak in the spectrum. Starting at the high-energy end of the spectrum, the ordinate scale of each appropriate standard response function is normal- ized at the full-energy peak and subtracted from the total spectrum in sequence. (2) Another program compares the unknown spectrum with the automatically selected library spectra in a simultaneous matrix solution. Although programs using a spectrum library appear straight- forward, they suffer from two defects. First, they cannot correct for calibration shifts, as has already been mentioned. Also, certain nuclides (e.g., many fission products) always occur in mixtures, and so the needed individual spectra cannot be determined. Therefore, a versatile gamma-ray unscrambling program must be capable of generating the required response 37

OMPOSITE SAMPLE Sc47-Be7- 10 u UJ to «s. O UJ 5 - o-composfte sample •-subtracted points — individual sample data 10 PULSE HEIGHT j i 100 800 1000 1200 Fig. 16. Decomposition of a composite gamma-ray spectrum into its components by successive subtraction of standard spectral shapes 38

O IT) SiNnoo 39

functions by using analytical expressions derived from standard spectra. Mathematical techniques for generating gamma-ray response functions have been described by West and Johnson, Chester, ' '' and Heath. An example of a computed response function which accurately reproduces the experimental data is shown in Fig. 17. The analytical methods for generating response functions 23 27 have been used by West and Johnson, ' Strickfaden and Kloepper, and Heath in their computer programs, which closely simulate the manual "peel-off" technique. These programs first make a gaussian fit to the highest-energy peak; the appropriate response function is calculated and subtracted from the total spectrum, and the process is repeated with the next highest energy peak. Continuous spectra are not suited to this type of analysis. Some workers ' ' ' have investigated the use of the so-called incremental methods, which are capable of analyzing both line spectra and continua. The method consists of con- structing a response matrix whose k column is the response function if the full-energy peak is centered in the k channel. The input data from the gamma-ray spectrometer are multiplied by the smoothed inverse of the response matrix. The smoothing of the inverse matrix is needed to prevent large oscillations in the output data arising from small fluctuations in the input 2 8 data. When properly applied, the matrix inversion method should yield a smoothed version of the true gamma-ray spectrum; that is, a single gamma-ray component would appear in the output, not as a line, but rather as a single gaussian curve. Although the programs just described are but a beginning, the results are encouraging. In view of the great need for practical solutions to the data analysis problem, it is to be hoped that more versatile computer programs will be available soon. Determination of Gamma-Ray Intensities. To obtain the intensity of gamma radiation emitted from the source, it is necessary to know the probability that a gamma ray from the source will strike the crystal, and the probability that an incident gamma ray will cause an event in the full-energy peak. The former probability is just the solid angle fl for the particular geo- metry, and the latter is often called the "intrinsic peak efficiency" ep(y). Although it is very difficult to compute 40

€p(y) exactly because of the multiple processes occurring within a large crystal, it is easy to compute the total intrinsic efficiency eT(y), which is simply the probability that a gamma ray denoted by y will produce a count once it strikes the crystal. The fraction of all counts in the spectrum which con- tribute to the full-energy peak is called the "peak-to-total" ratio, or "photofraction," and may be denoted by f (see Fig. 18). 20 UJ »- < o: o ID O o BACKSCATTER PEAK 200 400 600 800 PULSE HEIGHT 1000 Fig. 18. Illustration of a measurement of the "peak-to- total" ratio f. Note that the backscatter peak is excluded from the total area. 41

Then, ep(y) = f eT(y) • It is very important that in the experi- mental determination of f, spectra be measured under conditions which reduce the effects of scattering as much as possible; otherwise, environmental scattering will add to the Compton electron spectrum and yield a high value for the area of the total spectrum. Earlier it was mentioned that the full-energy peak was chosen for use in obtaining intensities because its area was free of spurious responses arising from the environment. Thus, once ep(y) has been determined by the process just described, it may be applied to experiments having considerable differences in energy resolution and scattering conditions, just so the crystal dimensions and source-to-crystal distance remain the same. Figure 19 presents values of ep(y) for Nal(Tl) from the work of Lazar, et al. , ' for 1-1/2 x 1-in. and 3 x 3-in. cylin- ders and a 3 x 3-in. cylinder with the top beveled at 45°, 1/2 in. from the edge. Values of eT(y) computed at Oak Ridge and the values of n eT(y) computed by Wolicki, Jastrow, and Brooks are compiled in a review by Mott and Sutton. Heath and Vegors, et al. , " extended these calculations of fi €T(y) to include point, line, and disk sources located on the axis of several sizes of Nal(Tl) cylinders; they also included measure- ments of f for use in the calculation of ep. The number of gamma rays of a given energy emitted from the source N(y) may be obtained from where P(y) , ep(y) , and fl have the same meaning as above; the factor A corrects for gamma absorption in any material between source and detector, and in the absence of experimental cor- rections it may be approximated by A = e^ . It often happens that the gamma ray of interest is coinci dent with another gamma ray; in this instance the full-energy peak area will be decreased by coincident summing. This situ- ation has been treated by Lazar and Klema, who derived the following equation for the emission rate: nep(y,) [1 - n eT<y2)W(o°)q 42

o -z. UJ o t UJ O cn 1.U<J A ^^n •^"-^ s^ x.1** s^ ^ \ s^' •a v ^~!n V y AT 3 cm/ \ s S 0.50 s s v \ v \ 3x3-in / * ^ y AT 9. 3 cm / ' , X \ s / /BEVELED 3x3-in AT 9.3 cm V ^ 0.20 0.10 0.05 17X 1-in AT AT 7cm- 2.5 c / ^ S ^ v \ \ ^ \ /BEVELED 3x3 -in / AT 3cm 4- 1-in m / X \ \ ^ V *t y ^ijs y \ ^ 5 s \^ \ S \ \ ^ ^ ^ \^ S \ V " •—7 cm s \ \ > 3x 3-in AT 9.3cm/ s , - 0.02 \ \ nm 0.1 0.2 0.3 0.40.5 1.0 2.0 3.0 4.05.0 GAMMA-RAY ENERGY, Mev L-l/2 x 1-inch and 10.0 3x3 Fig. 19. Intrinsic peak efficiency eg 1-1/ 1-inch Nal(Tl) crystals (Lazar, etal.30). Here, yl denotes the gamma ray of interest; q is the number of y2 in coincidence with y1 ; W(0°) is the angular distribution function of the two gamma rays integrated over the surface of the crystal. Since the correction fteT(y2)W(0°)q is small under most conditions, only a small error results fr6m setting W(0°) = 1, if the spin changes of the gamma transitions needed for the exact calculation of W(0 ) are not known. Coincident summing of gamma rays leads to another important experimental implication. The gamma rays which are lost to their respective full-energy peaks appear in a "sum peak," whose apparent energy is the total energy of both gamma 43

rays. Since the crossover transition for a simple gamma cas- cade will also be detected at this energy, such a peak must be corrected for the area of the sum peak, which is given by: P(y,) nep(y2)w(o°)q2 , N „ = - „ '— + N,, . (4) CS 1 - n eT(y2)W(0°)q2 , The notation is the same as above, and the term N represents the contribution from random summing within the resolving time of the electronic system: Nr - 2Tp(y,)P(y2) , (5) where P(y,) and P(y2) are the areas of the full-energy peaks due to yt and y2. The resolving time 2r is usually about 1 usec, and with the complexity of modern multichannel ana- lyzer systems this quantity is very difficult to compute. It may be determined quite easily by measuring the random coinci- dence peak from a source which contains no true coincidences. An example of such a source is Mn5*, which emits only a single gamma ray at 0.838 Mev; consequently, the random sum peak will be found at about 1.7 Mev. From Eq. 5, Nr : lPCO.838))* ' Once 2r has been determined by this method it can be used for any experiments with the same electronic system. Note also that 2r, N , and P(y) must have the same time units. 4. Detection of Heavy Charged Particles As has been mentioned, the earliest application of the scintillation method was to the detection of alpha particles. In its modern form the scintillation counter has found extensive use in counting and spectrometry of other heavy charged particles as well. Because of the short range of alpha particles, the scintillator may be made very thin. This also insures that the response to more penetrating radiations such as electrons and 44

Next: Ionization Chambers »
Detection and Measurement of Nuclear Radiation Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF
  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!