E . j24. ct I OF I ORNL P 1237 com . . . E + FEEFEEEE ni 1 MICROCOPY RESOLUTION TEST CHART NATIONAL BUREAU OF STANDARDS -1963 orna P-1235 .. MAY 17 1965 - CONF-65052/-/ RADIOACTIVATION ANALYSIS AS AN ANALYTICAL TOOL* W. S. Iyon Analytical Chemistry Division Oak Ridge National Laboratory . Oak Ridge, Tennessee Radioactivation analysis is one of the more beneficial results of "fall-out" from the U.S. Atomic Energy Program. A highly sensitive, speci- fic, often non-destructive technique, radioactivation analysis offers the analytical chemist an opportunity to perform a number of difficult determin- ations in a rather straight-forward manner. It is not a new technique:. Hevesy: and Levi (1) performed the first activation analysis in 1936. It is not an inexpensive technique: the most versatile neutron source is the expensive nuclear reactor, and unless one obtains the high fluxes of neutrons 1014 - 10+4 availa'yle from such reactors, the method has only limited sen- sitivity. It is not a universally applicable technique: for a number of elements the sensitivity is rather poor, and for some there is no sensitivity at all. Finally, it is not a method free from errors and interferences: the experience and judgement of a trained radiochemist is required in all but the most routine of application. In short, radioactivation analysis, like other analytical techniques, must be chosen discretely, applied intelli- gently, and its results interpreted judiciously. Almost all activation analyses performed today employ neutrons as the bombarding particles; probably 95% of all these neutron reactions invol.ve tae use of thermal neutrons. In this paper neutron will be understood to represent a thermal or 0.023 ev neutron unless otherwise stated. *Research sponsored by the U. S. Atomic Energy Commission under contract.. with the Union Carbide Corporation. PATENT CLEARANCE OBTAINED. RELEASE TO, THE PUBLIC IS APPROVED, PROCEDURES ARE ON EILE IN THE RECEIVING SECTION, .... - .. . . -- VER MARKET 2. : ry When any element, E, of atomic number 2 and atomic weight A is placed in a flux of neutrons, there is a finite probability that the element will capture a neutron to yield a new isotope of the same element but with a mass one unit heavier. Thus: .. " , " ELLE E + n-> ALLE + y ; AB[n, y)4+18 (12) N .'' L an example would be 2.00 +2 -> 6.900 + y ; 59co(n,)50cc (15) -- If, as is often the case, the newly formed nuclide is radioactive, one may determine its presercę by measurement of emitted radioactive particles or quanta. In the example (10) the bºco decays by beta emission followed by two quanta of energy (gamma rays). . In beta decay the radioactive nuclide emits a negative electron and an uncharged no-maes particle called a neutrino whose combined energies add up to the total decay energy of the transition. The beta emission results in the formation of an element of one atomic number higher; any additional energy is then usually emitted as gamma photons (uncharged). Thus be Co decays by beta emission with a half-life of 5.27 years: 6°Ni is stable so the decay chain ends; in some instances, however, several additional decay steps occur. Occasionally neutron capture results in production of two radioactive species of the same nuclide; one of these isomers usually decays by isomeric 3- transition - emission of gamma photons - to the other. Cobalt is an example of this. 59co(n,y)6 co(10 min) 59co(n,xj 60 co(5:27y) (3) Such isomer formation is relatively uncommon , and of little practical sigaificance in activation analysis. The production of a radioactive species by neutron (or particle) capture 18 given by the activation equation: . -0.60 A = 180(1 - -91693 €) A = Neo(1 - e where. A = disintegration rate of induced radionuclide N = atoms of element irradiated f = neutron or particle flux o = cross section of element t = irradiation time ' Ty = half-life of induced raiionuclide Inspection of (4) indicates that the amount of A produced (and hence the sensitivity of any activation analysis) is a linear function of the amount of element to be determined, the flux, and the cross section. The half-life, present in the exponential term determines the irradiation time: obviously activities with short half-lives require only short irradiations, whereas activities of very long half-lives may be impractical for use. Neutron fluxes presently available range from 10+- 1074n/cm2/sec. Cross sections of most thermal neutron capture reactions vary from 0.1 to 100 x 10-24 cm (0.1 to 100 barns). Based on a flux of 2015, irradiation time of 0.5 T, or 1 week, whichever is shorter, and a reasonable counting procedure, con- ; ! centrations of a number of elements determinable by activation analysis are.... shown in Figure 1.12) These values may be either raised or lowered, depending (Figure 1) on what interferences may be present in the sample. The important point is that the sensitivity is quite good for a number of important elements. Although absolute activation analysis could be used for some few ele. ments whose cross section and half-life are well known, uncertainties in the neutron flux spectrum, variation of the flux as a function of distance, and other experimental difficulties make this procedure difficult and uncertain. Most analyses are carried out through use of a comparator technique. Here: the sample and a known standard of the element ($) to be determined are irradiated and analyzed, and the amount of element in the sample found by comparison of the counting rates of standard and sample. There are, of course, interferences and errors in activation analysis. Inhomogeneities in the neutron flux seen by sample and standard have already been mentioned. Variation of neutron flux with distance in two reactor facilities is shown in Figure 2.' It is obvious that the flux pattern within . (Figure 2) onets Irradiation facility must be known. As is discussed in a later section, the presence and magnitude of a non-thermal neutron flux component may seri- ously affect analytical results obtained' by neutron activation analysis. Self- shielding is another source of error; it results from attenuation of the neutron flux seen by the sample due to the presence of large amounts of high cross section material. The determination of copper, which has a ~ 5 bazn cross section, in a matrix of europium (8000 barns) or cadmium (20,000 barns) would result in serious self-shielding. There are ways to correct for this effect, but one must be aware of it. An opposite effect is thermalization of the flux, which arises from matrices containing large amounts of hydrogeneous or other neutron moderating material; the presence of such material results in an enhancement of the thermal flux in the vicinity of the sample. Again corrections can be made for this effect. Some experimental measurements of . these effects are given by Reynolds and Mullins (3). So far only neutron capture reactions have been discussed, primarily because they are most important in activation analysis, or Thermal neutrons are the most abundant energy neutrons in a reactor, and thermal neutron cross sections are usually orders of magnitude larger than those of fast neutrons. Neutrons as produced by fission in a nuclear reactor exhibit a distribution of energies, but the presence of moderator material such as water or graphite reduces the energy of many of these neutrons to room temperature - thermal = 0.023 ev. Figure 3 presents the general features of a reactor neutron spectrum. (Figure 3) From this figure it is apparent that epi-thermal and fast neutrons are an appreciable fraction of the neutron environment. Elements capturing one . of these neutrons usually emit one or more particles rather than the gamma ray that follows thermal capture, Fast neutron interactions result in the following reactions : (n,r), (a,m), and (n,2n). Although cross sections for these reactions are usually only a few millibarns, fast neutron capture by the bulk matrix element may result in serious interference for the element to be determined. The mass and atomic number changes associated with these reactions and examples of each are shown below. A5 (,). 2B 60wa (n,p)60CO (n, as used 63cu(1,0)6OCO (1,2n) 4:23 59coln, 2n) 58cc :'(70) .. ... . . . . - - Fron equations (5) and (6) it is apparent that if one were attempting to determine cotalt in the presence of large quantities of nickel or copper, fast neutron reactions would produce the same 5 year 60 Co that thermal cap- ture by the impurity cobalt yields. Obviously this first order interference is a serious problem. It can be largely overcome, however, by repressing the last flux through use of suitable absorbers. Second order interference occurs when a product of a neutron capture undergoes an additional neutron capture. A . 45(1,9) attie Boys 4.23 (1,99472 585e(n,r)59pe nista 59co[n,716000 (76) In equation (76) 59Fe decays with a 15. day half-life to stable 59co. It is again apparent that the determination of cobalt ia Iron must take this effect into account. Usually, second order interference is unimportant except at high fluxes, long irradiations, or in determinations of extremely. low concentrations of trace elements. But, of course, it is for this latter 7. -7- experimental requiremeut that, activation analysis is so useful; hence, either correction must be made for the interference, or irradiation condi tions chosen so as to minimize or eliminate it. Ricci and Dyer 14) have calcu- lated second order interferences for a number of activation determinations. Probably in most activation analyses neither first nor second order inter- ference will be important, but the chemist musü be aware of the potential difficulties. Preparation of standard and sample in a similar manner with a similar matrix (if the matrix cross section is of importance), selection of proper irradiation facility and condition, and careful evaluation of the nuclear parameters associated with the problem will aid in avoiding most of the pitfalls mentioned above. . Up until the late 1950's neutron activation analysis was primarily a chemist's technique. The usual procedure was to irradiate sample and stan- dard, then dissolve both, add carriers, and perforin chemical separations for the desired elements. In the last step of the procedure the radioactivity of each separated and weighed precipitate ' would be counted, and the amount of sought element calculated by comparison of sample and the standard. With the . advent of modern NaI(Tl) gamma ray spectrometers equipped with reliable tran- sistorized multichannel analyzers, much of the chemistry has disappeared from the technique, which has become more and more nondestructive. It is now possible to determine a number of trace elements simultaneously by non- destructive gamma ray spectrometry, and it is this possibility, more than any other, that has given activation analysis its aura of glamour and pro- voked the flurry of rash predictions concerning its future usefulness. Present-day non-destructive activation analysis again begins with simultaneous irradiation of sample and standard. Following removal from the reactor, the sample is placed above a 3" x 3" NaI(Tl) crystal cemented to a ta photomultiplier tube. This detector is electronically coupled to a multi- channel analyzer, which is a device capable of sorting the various voltage pulses obtained from the detector system into energy increments. When elec- tromagnetic radiation (gamma and x-rays) interact with the NaI(TI) crystal, the interacting gamma ray may lose all, part, or none of its energy in the crystal. Interactions of the first type are called photoelectric events, and result in pulses of size proportional to the original gamma ray energy. Interactions of the second type are termed Compton interactions and result in a distribution of pulses up to some maximum that is less than the full gamma ray energy. Thus, in a multichannel analyzer system ; photoelectric events should appear as a single line spectrum; actually, statistical consider- ations cause the gamma ray photopeak to appear with a resolution (Full width of the peak at ha'if-peak height) of about 7.5% at 660 keV. This resolution slowly improves with increasing energy and becomes poorer as the gamma photopeak energy decreases. (Most gamma rays of interest in nuclear analysis range from 50 keV to 2500 keV (2.5 MeV):) Gamma ray interactions from the Compton process merely appear as a broad continuum; upon this continuum the photopeak is superimposed. As mentioned earlier, most radioactive nuclides prodiiced by neutron QV capture decay by emission of a beta particle followed by one or more gamma rays; the energies and half-life of these gamma rays uniquely characterize each radionuclide. Thus, if one is able to differentiate and quantitatively measure a gamma photopeak from a particular radionuclide, comparison with the measurement of a similar photopeak in a known standard will enable a quan- titative determination of the sought element to be made. It was mentioned earlier that Goco undergoes neutron capture to produce two products: 10 m Comco and 5.2y 60co. In the upper right hand corner . -9- of Figure 4 the decay scheme of this isomeric transition of Co is shown: (Figure 1) the decay is 997 % through a 59 keV gamma ray. The gamma ray spectrum of a sample of Oom co is shown to the left of this decay scheme. The 31 kev "escape peak" results from an Interaction in the crystal and is of no concern here. (For an explanation and discussion of excape peaks, backscatter peaks, sum peaks, and similar phenomena the reader is referred to reference 1). Observation of this gamma ráy, and confirmation of its half-life by decay measurement, would completely establish the presence of bonco in this spectrum. Figure 5 presents similar data for the 5.27y bo co : activity. The two gamma (Figure 5) ray photopeaks at 1.17 MeV and 1.33 MeV are characteristic of this nuclide. The Compton đi stribution accounts for the pulses seen from 0 - 300 pulse height units. In a mixture of radionuclides containing 6o co and (say) 237Cs, the photopeak of the 137cs (660 keV) would be superimposed on top of the Compton distribution from the co; such a' composite spectrum is shown in Figure 5. The relatively poor resolution of gamma ray photopeaks ; and the presence of an undesirable Compton distribution are the two main effects that limit the number of radionuclides determinable in any one mixture by non- destructive 7-ray spectrometry in activation analysis. Although there have been developed a number of techniques to minimize (anti-coincidence Compton shielding) or to correct (electronic y-ray resolution) this effect, accuracy falls off rapidly with increasing complexity of the spectrum. A new detector, the germanium diode, possesses extreme resolution (~ 7 kev), but its effi- ciency is so low that it is of little value for activation analysis at the present time. Because one must often obtain many spectra over a period of hours (decay measurements ) and remove the Compton . distribution from a number -10- of simultaneously measured radionuclides, modern computer techniques have been applied to these data handling. A typical spectrometer system complete with paper tape read-in read-out is shown in Figure 6. Data collected from the multichannel analyzer are read-out on the tape, which is later sent to an IBM 7090 computer for analysis. The computer searches out the desired photo- peaks, starting at the high energy end of the spectrum, and successively removes the Compton distribution contributed by the measured photopeak, before measuring the next highest one. Even so, after several subtractions precision and accuracy drop markedly. Probably no more than four or five separate deter- minations can be performed non-destructively on one sample at the present time. Often one or two simple chemical separations prior to measurement will enable a considerable number of nuclides to be determined by y-ray spectro- metry. Ross (5) has determined as many as 60 elements by a combination of chemical and physical measurement, Finally, mention should be made of non-reactor neutron sources and non- neutron activation analysis. Non-reactor sources at present include isotopic neutron sources - usually an alpha emitter mixed with beryllium - and the 14-MeV neutron generator now commercially available from a number of suppliers. Table I lists reasonable values for the fluxes possible from three different (Table I) types of neutron sources. From this table it is apparent that the extremely low neutron output of an isotopic source makes it useful only for very limited application e.8., determination of Hiivad Mn at 0.01% level, Al at the 0.2% level, and Si at 1%. Most other elements of interest would be detectable only at >1% concentration. The 14-MeV neutron generator has found increasing service - 1. as a tool for determination of oxygen. Oxygen can only be determined by thermal activation through a complicated and not altogether satisfactory secondary reaction technique: intimate mixing of the sample with lithium, irradiation in the reactor to produce the reactions : bril (, 31); 160(32,n) 28 18 (8) About 100 ug oxygen is the lower limit of detection (6). The fast neutron reaction: 1.6 160(n,p)26x B 16 (9) occurs with 14-MeV neutrons and enables oxygen to be determined by a non- destructive technique. Main difficulty experienced is the high background of oxygen in most sample container materials. Present sensitivity is limited to about 100 ug. A few other elements, notably nitrogen and fluorine, have been determined at comparable levels. Strain (7) in an excellent article discusses the uses and limitations of neutron generators. One other method of oxygen analysis has been recently proposed. Markowitz and Mahoney (8), noting that many He reactions are exoergic and hence require relatively low energies for the reacting particles, suggested that a relatively small cyclotron might be built to accelerate "He ions for use in oxygen and other low z element determinations. Ricci and Hahn (9) extending this work have studied the use of She as activating particles. They used a non-destructive measurement technique. The detectica limits they list for a number of elements using a reasonable beam current are shown in Table II. (Table II) THDAY J . - -12. . . A small (~9" radius) cyclotron such as that suggested by Markowitz and Mahony, though relatively expensive (10 dollars) has intriguing possibilities for use, not only with "He particles, but also as a source of neutrons. In this latter application one could irradiate a target continuously with Sre ions to produce neutrons at a constant rate, in contrast to the 14-MeV generator, which uses a ZrT target with a short (several hour) usage life time. Table III lists the results of their investigation of such neutron ; yields. (Table III) The technique appears promising. In conclusion, activation analysis may be thought of as a useful, spe- . cialized, and highly selective technique. Like other techniques, activation has interferences and limitations, but careful choice of conditions enables many of these difficulties to be minimized. For non-destructive, rapid, and extremely sensitive analyses, radioactivation offers the analytical chemist a tool of great utility. . REFERENCES 1. HEVESY, B. & H. LEVI, 1936. Kgl. Danske Vid. XIV: 5. 2. LYON, W. S. (ed.). 1964. Guide to Activation Analysis. Van Nostrand, Princeton, N.J. 3. REYNOLDS, S. A. & W. T. MULLINS. 1963. Intern. J. Appl. Radiation Isotopes 14: 421. 4. RICCI, E. & F.F. DYER. 1964. Nucleonics 22 No. 6: 45. 5. ROSS, W. J. 1964. Anal. Chem. 36: 1114. 6. BATE, L. C. 1963. Nucleonics 21 No. 7: 72. STRAIN, J. E. 1965. Use of Neutron Generators in Activation Analysis. In Progress in Nuclear Energy Series IV. H. ELION, Ed.: in press. Pergamon Press Ltd, Oxford, England. 8. MARKOWITZ, S. S. & J. D. MAĦONEY. 1962. Anal. Chem. 34: 329. 9. RICCI, E. & R. L. HAHN. 1965. Anal. Chem. 37: . Table I. Neutron Fluxes From Various sources Source Max. Thermal Neutron Flux n/cm2/sec. Max. Neutron Output n/sec. 226Ra - Be 2 x 107 1.8 x 105 1 x 107 14-MeV Generatore 1 x 2012 Reactor 208 - 2014 Table II. Thick-Target Detection Limits (ppb/100 dps) For Analyses by Activation of 160 in Zro, With 10-MeV 35e Particles Product 140 150 116 1 Product 86 Half-Life 72 ... 224.6 20.5 min 110 min Detection Limit 2120 1.190 .5.1 0.3 53.8 + 4.8 2.4 + 0.1 (3.4 + 0.2 for Sio) Table III. Neutron Yields from She-Induced Reactions in Low-Z Thick Targets Neutron Yield sec] Target 5 m.e.v. 3He Ions_ 10 m.e.v. He Ions VW lithium (LIH) beryllium 4.10 x 2020 5.59 x 1010 2.84 x 2010 boron 2.03 x 1021 3.78 x 1011 2.07 x 2011 3.66 x 1020 2.28 x 1020 carbon 3.90 x 109 6.49 x 108 oxygen (Zroz) : &calculated for a She-beam current of 100 amp. LA Figure 1. Sensitivities for Elements Using Activation Analysis. ! . . .. .over.. ....... ORNL-DWG. 65-3239A Sensitivities for Elements by Neutron Activation. 3 <0.002 mg VID. 0.02 - .002 kg W 0.2 - .02 kg 2 Li Be 1982 -0.2 kg B C N O F > 2 ug A Si P S CI Ar 4. K Ca Sc Ti V Cr Mn=Fe Co Ni Cu Zn Ga Ge As Se Br Kr 5 Rb Sr Y Zr Nb Mo Tc Ru Rh Pd AgEcd In Sn Sb Te I Xe 6 Cs Ba La Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Em 3 Na Mg DUDU } DIII UIT 6 Ce Pr Nd Pm Sm Eu Gd Tb DyEHOEr Tm Yb Lu£ Figure 2. Variation of Neutron Flux with Distance in Typical Reactor Facilities. ' 1 ORNL-LR-DWG. 79796 DISTANCE FROM REACTOR END OF FACILITY (OGR) P 무 ​후 ​우 ​우 ​- -- - - HOLE-71 (OGR) (in 'feet) RELATIVE THERMAL NEUTRON FLUX PNEUMATIC TUBE (ORR) (in inches) 3 h ñ 6 DISTANCE FROM REACTOR END OF FACILITY (ORR) 2 Figure 3. Reactor Neutron Spectrum. WIN ORNL-LR-DWG. 79797 1016 . NEUTRONS / cm2- sec-ev IIII .. - THERMAL EPI-THERMAL :- * NEUTRONS NEUTRONS FAST ~7 x 1013 ~3 x 1013 NEUTRONS n/cm2 /sec n/cm2 / sec ~2 x 1013 n/cm2 /sec 10-2 a 102 104 .. NEUTRON ENERGY, electron volts 10-4 Figure 4. Decay Scheme and y-Ray Spectrum of bomco. · -- -- · -... --- ORNL-LR-DWG. 36369 Co 60m (10 m) -0.059 Mev 99+% Co 60m 0.059 Mev Co 60 . :.. . . "ESCAPE PEAK" 0.031 Mev COUNTS (arbitrary units) MTTT -- 10°4 25 12 150 50 : 75 100 PULSE HEIGHT: ... .. .. Figure 5. Decay Scheme and y-Ray Spectrum of co. ORNL-LR-DWG. 36401A TT 4.17 Mev 11.33 Mev Co 60 (5.27y] :. 8 99+% 2.505 Mey Ir ; N760 . COUNTS (arbitrary units) 2.50 Mev "SUM" PEAK TT 1006 200 000 200 1000 PULSE HEIGHT DU . N . E .. .... . . Yunity 1: 3 2 : . ... . . .. - R . 12 H. TAT 5. . . .". . Y . _ 4 / 18 / 66 DATE FILMED END KA . ir i SEN .. .