RADIOACTIVATION ANALYSIS By E. N. JENKINS M.Sc. A.R.I.C. and A. A. SMALES B.Sc. F.R.I.C. (ATOMIC ENERGY RESEARCH ESTABLISHMENT HARWELL NR. DIDCOT BERKS.) Introduction IN radioactivation analysis the weight of the required element in a sample is determined by measuring the intensity of induced radioactivity rather than by measuring say the optical density of a coloured solution the weight of a precipitate or the volume of reagent consumed during a titration. The intensity of induced radiation is directly proportional other things being equal to the weight of the required element and is of course independent of the state of chemical combination of the element. The assay of certain ores for uranium and of fertilisers for potassium are common industrial examples of quantitative chemical analysis by measurement of radio- activity.The a- p- and y-radiation from 238U and its decay products and the p- and y-radiation from 40K (of 0.012~0 abundance in natural potassium) are examples of natural or spontaneous radioactivity only twelve elements of atomic number 92 or less show any such appreciable spontaneous radioactivity in the naturally occurriog mixture of isotopes. Most of the remaining elements can however be converted into artificially radioactive isotopes by appropriate nuclear bombardment or " activation ". The process of nuclear bombardment of a weighed sample normally together with a standard followed by measurement of the intensity of the induced radiation constitutes radioactivation analysis. Normally the method in- cludes the isolation and purification of the required artificially radioactive isotope or mixture of isotopes by specific chemical operations often per- formed in the presence of milligram amounts of an inactive isotopic carrier added after the activation in favourable cases this isolation is not required and the final measurement is made on the intact sample.In either case the mass of the required constituent Y in the sample is finally calculated by the simple equation Mass of Y in sample = Radiation intensity from Y in sample Radiation intensity from Y in standard Mass of Y in standard x Radioactivation analysis is applicable in principle to almost all the elements though a few of them present great difficulty. In general the distinctive features of the method are its extreme sensitivity its freedom from contamination by reagents the ease with which both the sample and the standard may be activated at the same time and the possibility in certain cases of preserving the samples intact through the analysis.83 84 QUARTERLY REVIEWS Experimental methods The methods and techniques involved in radioactivation analysis will be discussed with particular reference to a specific and important example the determination of small quantities of arsenic in biological material.1 Sampling and Activation.-A feature of radioactivation is the minimum pretreatment received by the sample before activation. Chemical treat- ment say for concentration of the arsenic or for the removal of interfer- ing elements or for the oxidation of organic matter is rarely necessary a t this stage. The physical preparation of the sample consists only of weighing it into a suitable irradiation container such as a quartz or Polythene ampoule which is then sealed.Refractory minerals may require preliminary crushing and grinding to aid their eventual dissolution Just as in any other sensitive analytical method extreme care must be taken to avoid contamination of the sample at this stage. I n particular the laboratory and balance used when weighing out samples must be segregated from the main chemical laboratory where radiochemical operations will be carried out involving macroquantities e.g. of arsenic as carriers. The most widely used method of activating the sample and standard is to place them side by side within the high flux of thermal and fast neutrons and y-rays provided by a nuclear reactor. This process is usually called a " pile irradiation ".Most of the nuclear reactions useful in radioactivation analyses can be induced by the thermal neutrons occasional use can be made of the fast-neutron component of the pile flux. Alternative methods of providing an intense neutron flux or bombardment with deuterons or other charged particles will be mentioned later. In the example we have chosen the activation of arsenic by pile irradiation the appropriate nuclear reaction is 75As(n,y)76As (t4 = 26.8 hr. p- and y-active) and it proceeds with a cross-section (for thermal neutrons) of 4.1 barns. An important feature of nuclear reactors particularly those of the graphite moderated type (e.g. B.E.P.O. Harwell) is their capacity for the simul- taneous irradiation of several hundreds of small samples (including standards if required).The practical details of tlhe irradiation of samples in B.E.P.O. have been discussed fully in a review by Smales.2 The irradiation facilities are available to outside workers and information can be obtained fkom the Isotope Division A.E.R.E. H a r ~ e l l . ~ Given adequate facilities for the irradiation of analytical samples the analyst will next be interested in the quantitative laws governing the growth of induced activity in the sample and the standard and its decay on removal from the irradiation device. Generally speaking it seems fair to remark that activation of a minor constituent by thermal neutrons is influenced far less by the nature of the solid or liquid matrix than is the case with Smales and Pate Analyst 1952 '77 196. Smales Atomics 1953 4 No.3 55. " Radioactive Materials and Stable Isotopes " Catalogue No. 3 Isotope Division Atomic Energy Research Establishment Harwell near Didcot Berks. 1954. JENKINS AND SMALES RADIOACTIVATION ANALYSIS 85 spectrometric arc or spark excitation. This implies that radioactivation analysis stands in much less need of specially prepared standards than direct emission or mass spectrometry. Normally the sample may be irradiated side by side with a primary chemical standard say a pure metal or oxide or a small volume of a standard solution. However it is not always per- missible to assume that the effective neutron flux is uniform throughout the combination of sample and standard. Apart from possible local variations in pile flux (which could be tested for by irradiating a number of specimens simultaneously) a variation in flux may be introduced through neutron absorption (self-shielding) by the required constituent or by others.For- tunately many common matrix materials (Si 0 C H Al) have low total absorption cross-sections for neutrons of thermal or of intermediate energy. The absorption by a minor const'ituent would not normally be excessive and the standard could be appropriately diluted to an equally low concentra- tion of absorbing centres. Where tlhe matrix material or a major constituent of it is opaque to thermal neutrons then the sample weight must be restricted to the minimum. The complete calculation of the self-shielding faqtor for a given matrix in a given shape integrated over the whole spectrum of pile neutrons would be a formidable task and only approximate solutions have been put f ~ r w a r d .~ A useful preliminary assessment can be made using the simple exponential equation f = foewNuT where fo f are the incident and attenuated fluxes N the number of absorbing centres per cm.3 r the radius of a spherical sample and cr the total atomic neutron absorption cross-section for the complete spectrum of pile neutrons a t the point of irradiation (not merely the cross-section for thermal neutrons). Attention has recently been drawn again by Plumb and Lewis to the effect of neutrons of ( ( resonance " energies. Any contribution by neutrons of intermediate energy to the tot'al self-shielding effect will be reduced if the irradiations are carried out (at considerable loss of sensitivity) in the thermal column of the reactor.I n all cases the practical effect of self-shielding in a given sample and/or standard could be judged by simultaneous irradiation of a number of samples of different weights or dilutions. The growth of induced activity during irradiation follows the law Io (dis. min.-l) where IoB = induced radioactivity due to product B t / t = ratio of irradia- tion time to half-life of B WA = mass of target element A 8 = fractional abundance of a specific isotope 6 = activation cross-section (in barns) for that specific isotope MA = atomic weight of the element A andf = activat- ing flux of neutrons per cm.2 per sec. The expression [l - exp (- 0.693t/tj)] is often called the (' growth = 60Wg.0.6.02.1023.0,.10-24.f[l - eXp (- 0*693t/'t:)]/nf~ Hughes and Harvey " Neutron Cross Sections " B.N.L.325 U.S. Govt. Printing 5 Keyes U.S. Atomic Energy Commission Unclassified Report No. A.E.C.D.-3@@0 6 Plumb and Lewis Nucleonics 1955 33 No. 8 42. Office Washington 25 D.C. 1955. 1950. 86 QUARTERLY REVIEWS factor ”. It has a value of 0-5 for an irradiation time equal to the half- life of the product ; for longer periods of irradiation it rises more slowly to a limiting value of 1.0. Practically an irradiation for 5 half-lives induces almost a saturation activity of magnitude controlled solely by the factors WA 8 oa MA andf. Under these conditions and with the values WA = 10-6 g. f = 1012 cm.-2 sec-1 (appropriate for irradiations in B.E.P.O.) the previous equation reduces to This equation is convenient in making a rapid preliminary assessment of activation possibilities.The decay of the induced activity starting as soon as the irradiation ceases follows the law I d = IoB exp (- 0.693d/t4) where Idg is the residual radioactivity of B after a time d. In the specific case of arsenic irradiation almost to saturation would require a t least 5 days ; a t the end of this period the induced activity due to 76As from lop6 g. of arsenic in the sample may be calculated as follows Io = 3.6 x lo7 x 4.1 x 1.00/75 = 1.96 x lo6 dis. min.-l pg.-I Twenty-four hours after the end of the irradiation the residual activity would still be about 1 x lo6 dis. min.-l. Radiochemical Purification.-Immediately after irradiation the sample is normally subjected to radiochemical purification in order to isolate the required product e.g. 76A~ free from extraneous radioactivity.For example the irradiation of lop6 g. of arsenic in 1 ml. of blood for 5 days would yield not only 2 x lo6 dis. min.-l of 76As but also 4 x lo7 dis. min.-l of 37-min. 38Cl and 2 x lo8 dis. min.-l of 14.8-hr. 24Na. Rather than undertake the labour of the quantitative separation and purification of 10-6 g. of arsenic with the attendant risks of losses such as those due to absorption on precipitates the analyst adds 50 mg. of inactive arsenic as a carrier for the radioactive 76As. Provided that the mixture can now be treated so as to bring all arsenic atoms normal and radioactive into the same chemical form-say as arsenate-the subsequent chemistry need not be quantitative. g. of arsenic (or it could well be loblo g. for radioactivation analysis has been applied a t this low level) is replaced by that of purifying 50 mg.of arsenic with quantitative determination of the chemical yield that is the percentage recovery of the added arsenic carrier. In the specific example of the deter- mination of arsenic in biological material the arsenic in the sample might be present as As02-l A s O ~ ~ - As3+ or as an organic compound The carrier was added as sodium arsenite followed by a wet oxidation of the sample and added carrier with hydrogen peroxide nitric acid sulphuric acid and perchloric acid. The radiochemical purification of the induced activity in the presence The problem of the quantitative recovery of JENKINS AND SMALES RADIOACTIVATION ANALYSIS $7 of the carrier involves normal chemical procedures such as precipitation distillation solvent extraction ion exchange and cknomatography.An operation frequently of use consists of " scavenging " traces of unwanfed elements from solution by forming in it a strongly adsorptive precipitate e.g. of ferric hydroxide antimony sulphide or barium sulphate. It is of course restricted to conditions such that the carrier element remains in solution. The purification need not be too time-consuming ; the operations may be carried out on a semimicro-scale and permit the use of a clinical centrifuge as a rapid alternative to filtration. Again small losses due to incomplete transference of solutions or to incomplete precipitation are adjusted by final measurement of the chemical yield. It has been found possible to purify the relatively short-lived isotope 233Th (half-life 22.1 min.) from very high initial extraneous ,&activities due to uranium zirconium yttrium and rare-earth elements within two hours of the irradiation and dissolution of the ample.^ The purification of arsenic in irradiated biological material or germanium oxide included several precipitations as the metal with ammonium hypophosphite distillation under oxidising conditions to remove germanium while the arsenic remains in the acid mixture and distillation of arsenic as arsenious chloride under reducing conditions.Useful accounts of the radiochemical purification of a variety of elements have been published by Meinke,* by Coryell and S~garman,~ and by Klein- berg. 10 During the radiochemical purification of the induced activity from the sample the standard will require similar but possibly less extensive treatment.At the very minimum the standard must be taken into solution together with carrier and then precipitated and weighed for determination of chemical yield in exactly the same chemical form as the end-product of the purification of the sample. In our example arsenic was finally precipitated from acid solution with ammonium hypophosphite washed with water and dried under a radiant heater. In the example followed above inactive 75As was converted by neutron capture into the p- and y-active 76A~ which was purified after admixture with arsenic carrier by chemical procedures specific for arsenic. It should be pointed out that cases arise where the product of pile irradiation is not identical chemically with the target element-cf. oxygen and uranium among others in the Table (p.94). In the determination of uranium utilising the fission of 235U it is convenient to add barium carrier to the irradiated sample and to carry out chemical procedures which are specific for barium not for uranium.ll Measurement of Source.-The mounting and counting of the carrier precipitates bearing radiochemically pure induced activity demand only 7 Jenkins Analyst 1955 80 301. 8 Meinke U.S. Atomic Energy Commission Unclassified Reports Nos. A.E.C.D.-2738 9 Coryell and Sugarman " The Fission Products Book 3 " National Nuclear Energy 10 Kleinberg U.S. Atomic Energy Commission Unclassified Reports No. LA-1566 11 Smales Analyst 1952 77 778 ; Seyfang and Smales ibid. 1953 78 394. 1949 and -3084 1951. Series Vol. IV-9 McGraw-Hill New York 1951. 1953 and LA-1721 1954. 88 QUARTERLY REVIEWS an elementary knowledge of radiochemistry.12 It is usually convenient to make the precipitates into a slurry on circular aluminium trays about an inch in diameter.Considerable attention must be paid to obtaining a uniform thickness over the tray ; in the most precise work the thickness should be so far as possible similar for the sample and the standard. With the counting equipment a t present commercially available in this country the ultimate sensitivity of radioactivation analysis is best attained by counting /?-particles rather than y-photons. Simple p-counting equip- ment comprising a lead-shielded end-window Geiger tube a power pack to give up to 1800 v a pre-amplifier and a scaling unit costs only a few hundred pounds and is becoming widely available. Doubtless y -counting will in future assume greater prominence with the improvement of tech- niques-possibly this may include a reduction in the present high background of scintillation crystals and the provision of inexpensive and stable multi- channel pulse-analysers.At the moment the chief interest in y-measure- ments lies in y-spectroscopy l3 usually with moderately intense sources to select individual y-energies and avoid much of the necessity for chemical separations. Examples of this approach will be given below. Measurement of the /3-activity of the sample and standard is best regarded as measurement of a ratio between two sources of similar composition and thickness at identical counting geometry. In this way many of the potential errors of absolute /?-counting (variations in self-absorption self-scattering back-scattering external absorption) are avoided./%Particles of maximum energy below 0.2 Mev require special mounting and counting techniques and in some cases a gas-flow proportional counter is used with the sample mounted internally with 272 geometry. Most of the elements listed in the Table with favourable sensitivity can be determined by making the final measurements with an end-window Geiger tube preferably of the mica- window EHM2 type. One fundamental aspect of measuring radioactivity must be stressed- the statistical nature of the radioactive process causes variations to appear between successive measurements on the same sample (even where decay loss is insignificant over the duration of the experiment). The standard deviation of a series of repeated measurements is an index of this variability and is numerically equal to the square root of the average number of particles or photons recorded during each measurement (not the square root of the count rate).The statistical fluctuations in measurements of radioactivity need not introduce large errors provided that the counting periods are prolonged so as to include 10,000 counts; the standard deviation is then 100 i.e. the coefficient of variation is 1%. The lower limits of reasonably precise measurement are governed by the efficiency of the counter and by the natural background of the equipment- owing to radioactive elements in the constructional materials and to cosmic l3 Cook and Duncan " Modern Radiochemical Practice " Oxford Unix-. Press l3 Owen Atorrzics 1953 4 No. 1 5 ; No.2 34 ; Connally and Lebmuf ,4naZ9t. Oxford 1952. Chent. 1953 25 1095; Peirson Natztre 1954 173 990. JENKINS AND SMALES RADIOACTIVATION ANALYSIS 89 radiation. Reasonable limits are 100 dis. min.-l for /3-emitters where the maximum /3-particle energy exceeds 0.2 MeV 1000 dis. min.-l for isotopes emitting softer ,&radiation and 1000 photons min.-l for y-emitters. The upper limit of precise measurement can be extended at will by taking only a small known fraction of the purified carrier for the actual measurement of radioactivity . In the above discussion we have assumed that the chemical treatment of the sample after addition of the carrier isolated the required activity com- pletely free from extraneous activities. Much of the reliability of radio- activation analysis arises from the ease with which this assumption can usually be checked.The radiochemical purity of the final precipitates isolated from sample and standard respectively may in the first instance be checked by a series of decay measurements. Aft,er appropriate corrections the count rates at various intervals are plotted on a logarithmic scale against time on a linear scale. Successive count rates obtained during the decay of a single radiochemically pure isotope will lie on a straight line of slope related to the half-life. Alternatively the penetrating power of the /3-radia- tion may be measured through a series of aluminium absorbers of increasing thickness with correction for any concomitant decay. The corrected count rates are again plotted on a logarithmic scale but against the absorber thickness on a linear scale.The shape of the absorption curve is char- acteristic of the p-spectrum of a particular isotope and may be closely compared with standard curves recorded with especially purified isotopes under the same conditions. The penetrating power of any y-radiation emitted by the isolated radioactive isotope may be measured through a series of lead absorbers a scintillation counter would normally be used as a detector. Finally the characteristic y-spectrum of the isotope can be recorded by means of a sodium iodide scintillation spectrometer l3 (see Figure) provided that the source is of reasonable intensity (with present equipment including a single-channel pulse analyser at least 1000 photons min.-l are required to obtain reasonable results). The record may be compared with standards.In certain cases the purified induced activity may be a mixture of radio- active isotopes e.g. the irradiation of cerium under the conditions defined in the footnotes to the Table would give 141Ce (33 days 0.58 Mev ,6,0.145 Mev y ) together with 14We (33 hr. 1-39 Mev @ 0.035 0.126 0.160 0.289 0.356 0.660 and 0-720 Mev y). The test of radiochemical purity remains that of comparing the sample and the standard under similar conditions. It is not expected that each of the three parameters half-life shape of p-absorption curve and shape of y-spectrometer record can in every case be precisely determined for example several activities listed in the Table decay only comparatively slowly. In general enough characteristic results may be gathered to establish the required degree of purity.The practical steps in a normal radioactivation analysis may now be summarised 1. Simultaneous activation of weighed quantities of sample and standard. 2. Dissolution of sample and standard each in the presence of carrier. 90 QUARTERLY REVIEWS 3. Treatment of sample and standard to bring the induced activity and 4. Chemical purification of carrier in the solution from the sample (and the carrier into the same chemical form in a homogeneous mixture. y-Spectrum of 7sAs. A Photoelectric y-capture peak at 1.21 Mev. B 9 9 9 , , 0.65 M ~ v . c D Compton continuum due to y of 0.55 MeV. E Back-scatter peak due to y of 0.55 Mev. 9 ¶ ¶ , , 0-55 M ~ v . possibly also from the standard) until the induced activity is radiochemically pure. 5. Determination of the chemical yields of carrier for the sample and standard severally.6. Comparison of the intensity of radioactivity of suitably mounted sources prepared from sample and standard under identical counting con- ditions with appropriate corrections for chemical yield decay dead-time losses in the counter etc. 7. Confirmation of the radiochemical purity of the isolated induced activities by measuring (a) the rate of decay ( b ) the shape of the p- and possibly the y-absorption curves and (c) in suitable cases the characteristic y-spectrometer record. JENKINS AND SMAI;ES RADIOACTIVATION ANALYSIS 91 Limitations caused by conflicting nuclear processes In the preceding discussion it was assumed that 76A~ the illustrative example could be formed in the sample as in the standard only by thermal neutron activation of the required element arsenic.However in certain rather special cases the required activity may be induced by neutron bombardment of some element other than arsenic. This hardly ever occurs except when the required element is accompanied in the sample by a large excess of an element differing from it by only one or two units of atomic number. Nuclear reactions of the (n,p) or (n,cc) type may then induce an activity in the foreign element identical with that of the product of the (n,y) reaction of the target element. A more subtle case arises when a foreign element of atomic number one unit less than that of the required element is actually transmuted into the latter by the succession of an (n,y) reaction and /3-decay. Examples of each of the above type of interference are given below.(a) During determination of traces of sodium in aluminium the fast neutron component of the pile flux induces the reaction 27Al(n,~t)2~Na [cf. 23Na(n,y)24Na]. Measurements made by Salmon 14 show that the spurious sodium activity introduced in this way is equivalent to 81 p.p.m. of sodium for irradiations near the centre of B.E.P.O. (1954). (b) Le Claire Gregory and Smales l5 attempted the estimation of traces of 40A in sylvite (KCI) minerals by neutron activation to 41A (this determina- tion would be significant in estimating the ages of such minerals). The estimation is subject to prohibitive interference from the reaction ( c ) Smales and Pate l6 have shown that the following sequence of reactions occurs during the activation of germanium 74Ge(n,y)75Ge(tg 1-37 hr.) ; 75Ge (/3 decay) 75As (stable) ; 75As(n,y)76As(tg 26.8 hr.) However insufficient arsenic is produced by transmutation from the germanium to introduce appreciable error into the determination of arsenic in germanium dioxide at the 0.05 p.p.m.level when the irradiations are made for 15 hours a t a flux of 10l2 neutrons sec.-l. It must be emphasised that the limitations described in this section become apparent only in rather special cases where one is trying t o deter- mine traces of an element in the presence of a very large excess of an element which adjoins it in the Periodic Table the cross-sections for (n,p) ( n ~ ) reactions or the yield of the combined (n,y ; #I) reaction being normally extremely low. The ( n g ) and the (n,a) reaction are normally due to the fast-neutron componept of the pile flux and can often be avoided if the irradiations are carried out in the thermal column.This irradiation facility provides a substantially lower flux than that near the centre of the pile and the analyses are correspondingly less sensitive. A.E.R.E. C/R 1324 1954. Each case relates to activation using pile neutrons. lK( n,p ) lA [ cf. 4OA( n y ) 41A]. 1 4 Salmon United Kingdom Atomic Energy Authority Unclassified Report No. 15 Le Claire Gregory and Smales personal communication. 16 Smales and Pate Analyt. Chem. 1952 24 717. 92 QUARTERLY REVIEWS The production of spurious 76As from germanium during the determina- tion of traces of arsenic in 1 g. samples of germanium dioxide depends on the time of irradiation and on the square of the neutron flux although the error is + 0.08 p.p.m.for 75 hours’ irradiation at 2 x 10l2 neutrons sec.-l it falls to + 0.0001 p.p.m. for 10 hours’ irradiation a t 2 x loll neutrons cm.-2 sec.-l. At this reduced flux and irradiation time the method retains adequate sensitivity for the analysis of 0.001 p.p.m. of arsenic. General discussion of radioactivation analysis for trace elements Irradiation of lowg g. of arsenic under standard conditions has been calculated (p. 84) to yield 2 x lo6 dis. min.-l. 100 dis. min.-l can be measured with reasonable precision when normal @-counting equipment is used. After allowance for minor losses due to decay incomplete chemical recovery etc. the lower limit of satisfactory analysis for arsenic may be set at 10-10 g. and Smales and Pate l6 have measured arsenic a t this level in distilled water.It is well known that the measurement of radioactivity permits the detection of extremely minute quantities of the natural radio- active series the analyst may now extend the scope of highly sensitive radioactivity measurements to include species which are themselves inactive but which acquire induced activity on neutron activation. By reason of this intrinsic sensitivity radioactivation analysis is particularly suited to the determination of trace elements i.e. elements present in a sample a t a concentration of 100 p.p.m. or less. The utility of an analytical method can be assessed by answering the questions (a) What property is measured ? ( b ) What elements can be determined and t o what lower limits ? (c) Is the element determined identified with certainty ? (d) Can several elements be determined simultaneously ? (e) How reliable are the quantitative results ? (a) What Property is measured ?-Radioactivation like emission or iiiztss spectrometry is a method of determining elements irrespective of their state of combination.Chemical bonds are frequently broken during neutron and y-ray bombardment. That is radioactivation does not disclose the chemical form in which an element occurs in a sample and it cannot be used for determination of specific compounds. It has just been stated that radioactivation is a method of determining elements but strictly it determines a specific isotope of the element that is radioactivation analysis for say dysprosium proceeds through a measure- ment of the induced activity due to the product of the reaction l64Dy(n,y)lg5Dy.Now the abundance of 164Dy in naturally occurring dysprosium is 28.47% the other important stable isotopes being those of mass 156 158 160 161 162 and 163. None of those isotopes is directly measured. Radioactivation has of course been used as a method of isotopic- abundance analysis e.g. by measuring the ratio 235U total uranium for enriched and for depleted uranium metal. l 1 These are favourable cases, JENKINS AND SMALES RADIOACTIVATION ANALYSIS 93 however and in general radioactivation is very much less versatile than inass spectrometry for isotopic analysis. (b) What Elements can be determined and to what Lower Limits?- On p. 92 the limit of sensitivity of radioactivation analysis for arsenic was discussed on the assumption that the sample was irradiated to satura- tion at a flux of 1OI2 neutrons cm.-2 {ec.-l and that 100 dis.min.-1 of induced activity due to 76As can be isolated and measured. It is in- structive to extend this discussion over the entire range of chemical elements. A practical definition of sensitivity might be taken as that weight of element which on irradiation a t a flux of 10l2 neutrons cm.-2 sec.-l to saturation or for one month (whichever is the shorter) followed by decay for 2 hours gives 100 dis. rnin.-l of residual induced activity. The Table sets out the absolute weights of various elements which one could expect to measure the corresponding concentrations depend on the sample weight-in general 1 g. would be a fair figure possibly rather more .for solids and rather less for liquids.The coefficient of variation of single analyses for a given element would be about 10% if the sample and the background could each be measured for an hour. sec.-l assumed in calculating the results given above is appropriate for irradiations carried out near the centre of a graphite-moderated natural-uranium reactor of the B.E.P.O. type. Higher fluxes (available over a smaller effective volume) are available in water-moderated reactors of similar or higher power rating e.g. figures of 7 x 1013 and 5 x 1014 have been quoted for the Canadian heavy-water reactor at Chalk River l7 and for the American light- water-moderated Materials Testing Reactor at 1daho.l' Heavy-water reactors now under construction at Harwell l8 should in time provide irradiation facilities which will extend the sensitivities quoted in the Table by a factor of nearly 100.In certain cases elements listed with poor sensitivities in the Table may be analysed more readily by other methods of activation or detection. The most important alternative to pile irradiation is the use of a cyclotron beam of protons or deuterons. The disadvantages compared with pile irradiation are the limited cross-sectional area of the beam and the limited degree of penetration of charged particles into solid samples. The radioactivation analyses of carbon in steel by von Ardenne and Bernhard and others,25 and in organic compounds by Sue 2r exemplify a cyclotron-induced reaction which could not have been brought about in the pile namely 12C(d,n)13N (9.9-min. 1-2 Mev positron emitter). Bade and others 27 used the 22 Mev betatron at the G.Roussy Institute Villejuif France for the radioactivation analysis of oxygen down to a 6 mg. level in organic compounds and down to 0.1% in aluminium metal by using the reaction 160(y,n)150(t$ 2.1 min. 1.7 Mev Pf). These workers indicated l7 " Catalogue of Nuclear Reactors " C.R.R.-590 Atomic Energy of Canada Ltd. 18 " First Annual Report of the United Kingdom Atomic Energy Authority 1954- The flux of 10l2 neutrons Chalk River Ontario 1955. 5 5 " H.M.S.O. London 1955. 94 QUARTERLY REVIEWS Estimated sensitivities of radioactivation analyses by irradiation in a natural uranium graphite-moderated reactor (for notes see pp. 96 97). a s Element Actinium * . Aluminium . Antimony. . Argon. . . Arsenic . . Astatine * . Barium . . Beryllium.. Bismuth . . Boron. . . Bromine . . Cadmium. . Czesium . . Calcium . . Carbon . . Cerium . Chlorine . . Chromium. . Cobalt. . . Copper. . . Dysprosium . Erbium . . Europium. . Fluorine . . Francium * . Gadolinium . Gallium . . Germanium . Gold . . . Hafnium . . Helium . . Holmium . . Hydrogen. . Indium . . Iodine . . . Iridium . . Iron . . . Krypton . . Activated form Parent zz7Ac '*A1 ."Sb with lZ4Sb L'A 7 6As Parent zloAt L 3 g B ~ l0Be L2B 30mBr + 80Br with szBr l15Cd with l17Cd and l1 5mCd 134cs 45Ca 14c 21OBi 41Ce with 143Ce 51Cr 60Co 64Cu 38c1 5Dy 171Er 152mE~ 20F Parent 3Fr 159Gd 2Ga 75Ge with 77Ge lg8Au lslHf 6He 3H 166H~ 3H l161n with 1281 with lg21r 59Fe 85Kr 11 4rn-ll 4mIn 1941r Half-life 22 y. 1% a 2-3 min. 2.8 d. 60 d. 1.8 hr. 26.8 hr. 3.3 hr. EC 0*20y0 a 35 min.2.7 x 106 y. 5.0 d. 0.03 sec. 4.6 hr. 35.9 hr. 54 hr. 2.9 hr. 43 d. 2.3 y. 152 d. 5580 y. 32 d. 33 hr. 37.3 min. 27 d. EC soft p 3% Y 58% p- + ps 5.2 y. 12.8 hr. EC 2-3 hr. 7.5' hr. 9.3 hr. p EC 12 see. 21 min. 18.0 hr. 14.1 hr. 82 min. 12 hr. 2.7 d. 45 d. (unknown) 12.4 y. 27.2 hr. 12.4 y. 54 min. 50 d. 25 min. 19 hr. 75 d. 45 d. 4.4 hr. soft /3 soft p Estimate sensitivit (g.1 5 x 10-13 Poor 1 x 10-10 5 x 10-10 5 x 10-11 1 x 10-16 1 x 10-9 Poor 5 x 10-8 Poor 1 x 10-10 1 x 10-9 5 x 10-10 Poor i x 10-7 1 x 10-9 5 x 10-9 1 x 10-7 5 x 10-10 1 x 10-10 1 x 10-12 1 x 10-10 1 x 10-12 Poor 1 x 10-18 5 x 10-10 1 x 10-10 5 x 10-12 5 x 10-10 5 x 10-6 5 x 10-12 Poor 1 x 10-11 5 x 10-9 i x 10-9 1 x 10-7 1 x 10-11 5 x 10-11 NCteil * K 0.8 * K 0.11 K < 0.4 K 10-4 To give 100 (p-+p+) min.-l K 0.05 * n,p on 3He K 1 x 10-5 JENKINS AND SMALES RADIOACTIVATION ANALYSIS 95 Element Lanthanum .Lead . . . Lithium . . 7 . . 7 9 . . Lutecium . . Magnesium . Manganese . Mercury . . Molybdenum . Neodymium . Neon . . . Nickel. . . Niobium . . Nitrogen . . Osmium . . Oxygen . . Palladium. . Phosphorus . Platinum . . Polonium * . Potassium. . Praseodymium Promethium * Protactinium * Radium" . . Radon* . . Rhenium . . Rhodium . . Rubidium. . Ruthenium . Samarium. . Scandium. . Selenium . . Silicon . . . Silver . . . Sodium . . Strontium. . Sulphur . . Tantalum . . Technetium * . 9 Tellurium . . Activated form ________ 140La zOgPb 8Li n,u 3H '8F 177Lu with 176Lu 7Mg zo3Hg 147Nd with 149Nd 3Ne s5Ni 9 4 m E b 5 6 ~ n 9 9 ~ 0 16N with lglOs or 18F l09Pd 32P 1 9 7 P t with lg9Pt Parent 210Po 42K Parent 7Pm Parent z31Pa Parent 226Ra Parent 2 2 2Rn ls8Re with lssRe 10 4mRh-10 4Rh 86Rb with 88Rb lo5Ru with lo3Ru 3Sm 46Sc with 8% 31Si llOmAg-110Ag 24Na 89Sr 3 5 s or 32P 18ZTa Parent g * T ~ 99mTc 'Te 1 9 3 0 s 1 9 0 2Pr 8lmSe-8lSe with lz9Te Tabk Half-life 40 hr.3.3 hr. 0.8 sec. 12.4 y. 112 min. fl+ 6.8 d. 3.7 hr. 9-6 min. 2.6 hr. 48 d. 67 hr. 11.6 d. 1.8 hr. 40 sec. 2.6 hr. 6.6 min. IT 7.3 sec. 31 hr. 16 d. soft /3 29 sec. 112 min. /3+ 13.4 hr. 14.3 d. 18 hr. 31 min. 138 d. cc 12.4 hr. 19.3 hr. soft fl l3 (0.1%) 2.6 y. 1620 y. u 3.8 d. u 17 hr. 91 hr. 4.3 min. 19 d. 18 min. 4.5 hr. 40 d. 46 hr. 85 d. 57 min. 17 min. 2.6 hr. 270 d. 15 hr. 54 d. 87 d. soft fl 14.3 d. 111 d. 6 h. IT y 9.3 hr. 70 min. 3.4 x 105 y. 61 > 107 y. Estimated sensitivity (g.1 5 x 10-11 5 x 10-6 Poor 5 x 10-10 Possibly < 5 x 10-12 5 x 10-5 1 x 10-9 1 x 10-11 1 x 10-8 5 x 10-10 Poor 1 x 10-8 5 x 10-2 Poor 1 x 10-9 Poor 1 x 10-10 5 x 10-10 5 x 10-7 1 x 10-9 1 x 10-10 5 x 10-11 1 x 10-11 5 x 10-18 1 x 10-11 1 x 10-9 5 x 10-14 5 x 10-13 5 x 10-3 1 x 10-9 5 x 10-10 1 x 10-11 5 x 10-1' 5 x 10-8 1 x 10-8 1 x 10-10 5 x 10-8 1 x 10-10 Poor Possibly 5 x 10-9 1 x 10-7 5 x 1 0 - 7 sensitive 5 x 10-9 :ontinued on next page.Notes K 0.4 d K 0.02 To give 100 fl min.-l K 2 x 10-8 K 2 x 10-8 d * * * * * 96 QUARTERLY REVIEWS Terbium . . Thallium . . Thorium . . Thulium . . Tin. . . . Titanium . . Tungsten . . Uranium . . Vanadium. . Xenon. . . Ytterbium . Yttrium . . Zinc . . . Zirconium. . Activated form 16oTb 233Th or 233Pa 17OTrn 121Sn with 12%n 61T i 187W or 2 3 9 N ~ or 140Ba 52V 133Xe with I35Xe 175Yb SOY 69Zn g7Zr with g5Zr 2 0 4 ~ 1 239U Half-life 73 d.2.7 y. 22 min. 27.4 d. 129 d. 27 hr. 40 min. 6 min. 24 hr. 23.5 min. 2.3 ci. 12-S d. 3.7 min. 5.3 d. 9.2 hr. 4.2 d. 61 hr. 52 min. 17 hr. 65 d. Estimated sensitivity (6.) 5 x 10-11 1 x 10-8 1 x 10-10 5 x 10-11 5 x 10-9 5 x 10-9 1 x 10-2 6 x 10-1’ 1 x 10-8 1 x 10-10 Poor 5 x 10-9 5 x 10-9 5 x 10-11 1 x 10-10 5 x 10-9 1 x 10-7 I i Notes I Fission of 235U K 9 Notes t o Table ( a ) The sensitivities given in the Table have been rounded to the nearest 5 times an integral power of 10. The irradiations are assumed to have been carried out a t a thermal neutron flux of 10l2 sec.-l for one month (or to saturation if this period is less) and to have been followed by a two-hours’ delay during which radiochemical purification is carried out with quantitative yield.In rare cases where the fast neutron component of the pile flux is used e.g. to induce the S(n,p) reaction the sensitivities given apply to the neutron distribution a t the centre of the B.E.P.O. reactor a t Harwell. The radiochemically pure sources are assumed to be counted under an end-window Geiger counter a t an efficiency of 10% for p-particles or positrons of maximum energy > 0.2 Mev and of 1% for ?-particles of lower energy (tabulated as “ soft p ” ) . These efficiencies are deemed to include self-absorption effects in the sources. In the rare cases where the only suitable activated form decays only through isomeric-transition (IT) or electron- capture (EC) processes it is assumed that the source is y-counted under a scintilla- tion counter a t 10% efficiency ; the sensitivity limit has then been set a t a weight sufficient t o give lo3 y-photons min.-l (rather than lo2 particles min.-l as in /I-counting) to allow for the higher background normally characteristic of the y-counter.It is impossible in one compact table to indicate all the possibilities of pile activation the data presented here relate particularly to the possibilities of conducting highly sensitive analyses within two hours of irradiation. I n some cases the analyses could be carried out (with some loss of sensitivity) by using longer-lived isotopes e.g. analysts not working near a pile might not wish to use the 52-min. 69Zn for zinc analysis a t 5 x g. but could still use a 14-hr. 6gmZn or a 250-day 65Z1~ a t a level of perhaps lo-’ g or more.These possibilities can only adequately be appreciated by consulting the tables of nuclear data. Elements marked * occur in Nature as radioactive isotopes of reasonable specific activity or else the only known forms are artificially prepared isotopes in which case the isotope of greatest half-life has been selected. Many of these elements are x-emitters because of the much lower background the sensitivity of cr-counting is about 100 times that of P-counting but a proportionately longer time must be spent on the counting periods to attain reasonable precision. The Table is based upon the General Electric Company’s Chart of the Nuclides,l9 and upon the “ Table of Isotopes ” by Seaborg and others.20 “ Chart of the Nuclides ” General Electric Company Schenectady New York 4th Edn.1952. 2o Hollander Perlman and Seaborg Rev. Mod. Phys. 1953 25 469. JENKINS AND SATALES RADIOACTIVATION ANALYSTS 97 ( b ) The abundance of 3He in norinal helium has been assumed to be 1.3 x %. The 3He(n,p) reaction could detect 5 x 10-l2 g. of enriched 3He. ( c ) The 2H(n,y)3H reaction would be moderately sensitive as a means of analysis of enriched deuterium with a sensitivity of 1 x 10-5 g. ( d ) The possibility of radioactivation analysis for lithium or oxygen by means of lSF arises from the sequence *Li(n,~x)~H lSO( 3H,n)18F. The sensitivity given by Osmond and Smales 21 for oxygen applies to the determination of oxygen in finely divided (< 50 p ) metallic beryllium intimately mixed with an excess of lithium fluoride. In principle the reactions could be applied also to the determina- tion of lithium.The flux of fast neutrons in pile irradiations may be increased by placing the sample inside a hollow uranium cylinder under these conditions the yield in this reaction in the B.E.P.O. pile increases 22 by a factor of seven-the sensitivity of the sulphur determination would then be about 1 x lo-* g. This fast neutron reaction is included in the Table to illustrate the possibilities of using peutron-induced reactions other than the ( n y ) reaction. A possible advantage of a method for sulphur based on 32P is the ease of measuring the 1.7 Mev b-radiation rather than the 0.17 Mev b-radiation from 35S. An obvious disadvantage is the parallel production of 32P by ( n y ) reaction on any phosphorus in the sample which necegsitates a separate analysis.The interference from phosphorus might be greatly reduced by the use of a cadmium screen to absorb the thermal neutrons. (f ) Herr 2 3 and Alperovitch and Miller 24 have obtained preliminary evidence for the presence of a long-lived 9 8 T ~ in mineral samples by activation to S9mTc. It is not yet possible to state the sensitivity of the method in terms of the mass of @Tc. ( 9 ) When sensitivity calculated according to the convention adopted is par- ticularly low the value of the function K = lOOOu,/M~ has been tabulated as well as the half-life. Certain elements show a high K value and the poor sensitivity is entirely due to the very short or very long half-life. the possibility of extending the method to carbon and nitrogen activated to 20-min. 11C and 10-niin. 13N respectively.The use of the pile or of the cyclotron or betatron discussed so far in this Review has been to produce an activated isotope which can be isolated chemically and measured e.g. by using a Geiger-Muller counter after the irradiation has ceased. This intention is soinetimes frustrated by the over- long or overshort half-life of the product. An alternative approach is to measure each activating collision actually during irradiation. This has been adopted by Gaudin and Pannell 28 in measuring beryllium down to 1-2 p.p.m. in minerals by counting the prompt neutrons emitted during the reaction 9Be(y,n)24He. In the specific case of 9Be the threshold for the photoneutron reaction is below 2 MeV and a portable y-source (60d-l24Sb max. y-energy 2-04 MeV) is adequate for the irradiation.The prompt neutrons are counted by means of a boron trifluoride pulse ionisation chamber. The principle of the instantaneous detection of activating collisions has been extended to pile-irradiations. A natural limitation is the requirement that the measuring device must be very sensitive to the required nuclear event 21 Osmond and Smales Analyt. China. Acta 1954 10 117. 22 Whitehouse anti Putman “ Radioactive Isotopes ” Oxford Univ. Press Oxford 23 Herr 2. Naturforsch. 1954 9a 907. 2 4 Alperovitch and Miller Nature 1955 176 299. 25 von Ardenne and Bernhard 2. Physik 1944 122 740. 26 Sue Conapt. rend. 1953 237 1696. 27 Basile Hur6 LBvhque and Schuhl ibid. 1954 239 422. 28 Gaudin and Pannell AizaZyt. Chem. 1951 23 1261. ( e ) The 32S(n,p)32P reaction requires a neutron energy of a t least 1 Mev.1953 p. 124. G 98 QUARTERLY REVIEWS and yet remain insensitive to the pile flux of slow and fast neutrons and y-photons. In principle a pulse ionisation chamber might be used to deter- mine traces of say boron lithium or uranium,2Sa all of which eject heavily ionising particles (a-particles tritons or fission fragments) when irradiated with neutrons. The sample would be contained within the counting chamber. Actual analyses for each of these elements have in fact been made by using a nuclear emulsion as detector. The sample is evaporated from a microdrop (commonly 0.001 nil.) on to a metallic support which is irradiated in close contact with a sensitive emulsion. The emulsion is then developed and fixed and the photographic record is examined under a microscope.It is possible to distinguish tracks due to a-partJicles tritons or fission fragments from the shorter tracks due for example to protons. The reactions involved are (i) 10B(n,a)7Li (used by Faraggi by Mayr and by Loveridge and Smales to determine boron down to 2 x lop9 g.).29 (ii) 6Li(n,~)~H (used by Picciotto and Van Styvendael to determine lithium down to (iii) 235U(n fission) (used by Curie and Faraggi t,o study the localisation of uranium on the surface of polished mineral specimen^).^^ Of the 82 elements which are not naturally radioactive with high specific activity all but hydrogen seem susceptible to radioactivation analysis in favourable circumstances. Beryllium boron carbon and oxygen although unsuited to neutron activation may be activated by other means. Aluminium fluorine neon niobium rhodium titanium and vanadium while susceptible to activation by thermal neutrons give relatively short- lived isotopes which would decay prohibitively during moderately prolonged radiochemical separations.(c) Is the Element determined identified with certainty ?-The specificity of radioactivation analysis is high. It relies on three and sometimes four characteristics viz. the specific radiochemical purification the half-life of the product the maximum P-energy as determined from aluminium absorption measurements and in favourable cases the y-spectrum. (d) Can several Elements be determined simultaneously ?-It is obviously possible to irradiate an unknown sample for say 1 month and then to take it into solution add carriers separate the groups and isolate specific coni- pounds.This technique would not normally be considered as valuable a means of qualitative analysis as the simultaneous recording of a wide range of elements by emission or mass spectrometry. In certain cases it is feasible to carry out analyses for a limited range of elements after a single irradiation of a sample and a mixed standard for instance Smales and others during their analyses of marine sediments and of rock samples separated 2.5-hr. 65Ni 12.9-hr. 64Cu and 5.2-y. 6oCo in that g.).30 28a Stewart and Bentley Xcience 1954 120 50. 2B Faraggi Kohn and Doumerc Cornpt. rend. 1952 235 714 ; Mayr Nucleonics 30 Picciotto and Van Styvendael Compt. rend. 1951 232 855. 31 Curie and Faraggi ibid. p. 959. 32 Smales Geneva Conference on Peaceful Uses of Atomic Energy August 1955 1954 12 No.5 58 ; Loveridge and Smales unpublished work at Harwell. Paper 770. JENKINS AND SMALES RADIOACTIVATION ANALYSIS 99 The simultaneous analysis for several trace elements by irradiation followed by direct estimation of individual disintegration rates without radiochemical separations is sometimes feasible provided that the activity clue to the main constituents is not excessive. The individual trace elements may be recognised by resolving the gross decay curve or with much greater certainty by a combination of decay studies and y-spectroscopy. For clxample Smales 32 irradiated a sodium-potassium alloy for a long time and tjhen allowed two weeks for decay ; subsequent y-spectroscopy without chemical separation revealed 134Cs (y-energies 0.59 0-80 Mev) lloAg (0.88 1.36 1.48 Mev) and 87Rb (1.09 Mev).In these cases identification was by ;)-energy only as the products were long-lived. The quantitative analyses revealed Ag 100 p.p.m. Rb 5 p.p.m. and Cs 0.1 p.p.m. the last figure being ;tctually obtained in a separate analysis which included a radiochemical purification. In a similar study Morrison anti Cosgrove 33 irradiated silicon for 3 days then measured 69Zn 7 6 A ~ lX7W 59Fe 24Nn 42K and 182Ta by ( lirect y mintillation spectrometry. (e) How reliable axe the Quantitative Results ?-The precision of results (an index of the reproducibility of repeated individual analyses) must be (listinguished from the absolute accuracy or closeness of the mean value to the truth. Radioactivation like the other general methods of trace analysis should give a coefficient of variation of no worse than 10% for single determinations < L t an adequate level of trace element i.e.at the levels for the various chlements quoted in the Table. By careful work at somewhat higher levels the coefficient of variation can be lowered to 1%. The possibilities of very precise radioactivation analysis have been demonstrated by Seyfang’s deter- mination 34 of the isotopic content of depleted and enriched uranium this analysis which involves measuring the fission product 14*Ba produced on pile-irradiation of the 235U constituent is similar in principle to a trace- element analysis and the method was originally used to determine traces of natural uranium. In Seyfang’s most recent work 7 portions of natural uranium as the oxide U30, were simultaneously irradiated then treated chemically and the final barium sulphate sources were P-counted each source recording about 50,000 counts in 4 minutes.The series of 7 c:orrected counts showed a coefficient of variation of only 0.5%. The statistical errors in the determination of the counting rates account for 0.4% i.e. for most of the observed variation. It may be estimated that tihe analysis of an unknown sample by simultaneous irradiation of one portion of sample and one portion of standard should be subject to a caoefficient of variation of 0.7 yo in these particular conditions. The precision attainable by a given trace-element method normally tlecreases as the amount of the required constituent present becomes smaller. ‘Che precision a t any given level depends not only on the intrinsic sensitivity of the method but also on a factor present in greater or smaller degree in all trace-element analytical methods namely the “ background ”.The Morrison and Cosgrove AnaZylt. Chem. 1955 27 810. 3 4 Seyfang Analyst 1955 80 74 ; cf. ref. 11. 100 QUARTERLY REVIEWS “ background ” may take the form of contamination collected froni reagents or the atmosphere as in absorptiometry and certain other methods or of the residual current in polarography or of the electronic noise level in recording spectrometry or of the extraneous radioactivity due to cosmic radiation etc. in radioactivation analysis Strictly i t is not the absolute level of the background but rather the fluctuation between one measurement and another which is the limiting factor. One of the most restrictive types of background is that set by coil- tamination from impurities in reagents.This often remains a problem even after reagents of the highest quality have been subjected to further elaborate purification. The radioactivation niethod avoids errors introduced by impure reagents and it remains to take adequate precautions against surface contamination by e.g. atmospheric dust during the physical preparation of the sample. This method like any other is still subject to a background which however is normally of a low level it is the intrinsic background of the shielded counting assembly normally of the order of 10 counts per minute for an end-window Geiger-Muller tube of 1 inch diameter ; its relative insignificance for many trace analyses is illustrated b y the facts that even the iniiiiite weights of various elements listed in the Table would normally give a count rate (due to sample plus background) of 20 counts per minute and that the coefficient of variation of the background (for an hour’s counting) would be as little as 4%.Next must be discussed the absolute accuracy of the method i.e. the degree to which the experiineiitally determined level of the trace constituent obtained as the mean of many analyses with high reproducibility approaches the truth. This is sometimes expressed as the bias (positive or negative) of the method. In many methods of analysis including radioactivation the sample is compared directly or indirectly with a standard. If then there is adequate precision the absolute accuracy of the method depends on the availability of a suitable standard i e .one whose composition is accurately known to an accuracy exceeding the maximum accuracy of the trace- element determination. This requirement is often met by weighing out elements or compounds of known composition or by dispensing solutions of such reference substances. Further the behaviour of the standard in colour development excitation activation etc. niust exactly parallel that of tlhe trace constituent in the sample. For example a sample of finely ground rock might be excited for say 40 seconds by a graphite arc for emission spectro- metry volatilisation of a trace constituent froin the rock matrix might then not be matched in a parallel excitation of say a solid dilution of the oxide of the element in graphite or silica. Radioactivation is independent within reasonable limits of the nature of the solid matrix of the sample provided that the weight and concentration of the sample and the standard are sufficiently low to avoid the self-shielding errors discussed on p.85. This arises because radioactivation is a nuclear process and does not involve considerations of the volatility of atomic species their adhesion to other species or the excitation behaviour of their JENKINS AND SMALES RADIOACTIVATION ANALYSIS 101 external electrons. Direct radioactivation followed by measurement of dis- integration rate all on the intact solid sample and accompanied by a parallel analysis of a standard should approach absolute accuracy. The standard eould be a pure element or compound or a liquid or solid dilution. In this hypothetical case it has been supposed that the disintegration rate of the activated product could be measured directly on the solid sample (and on the standard) without loss due to self-absorption or interference due to foreign activities.This ideal has been attained in cert'ain insta,nces such as the direct y-spectrometer nieasurements (pp. 99 103). More generally the sample must be dissolved after irradiation and a carrier added followed by radiochemical purification of the element concerned. The radioactivation method will retain high accuracy-if the activated species passes completely into true solution without loss by volatility adsorption or persist'ent tlraces of insoluble residues and if the dissolved species undergoes complete isotopic exchange with the added carrier. Summarising t>he radioactivation method should give a low bias in those I'avourable cases where it can be applied without radiochemical separation.In its more general applications its accuracy is as good as that of the mass- spectrometer isotope-dilution method and greater than that of direct einissioii or mass spectrometry. It is very valuable as a method for the standardisat)ion of samples e.,q. rock samples containing trace elements which can then be used as standards for emission or mass spectrometry (see below). Practical applications of radioactivation analysis to the determination of trace elements There are four major fields \$here analysis of trace elements is required namely geochemistry biology physics of the solid state and nuclear physics. Geochemistry.-Geochemistry requires reliable analysis of the distri- imtion of the elements often only in minute amounts throughout the earth's erust the oceans and meteorites.The analytica81 results provide essential data for the theorist a'nd are occasionally of immediate practical interest as in geochemical prospecting i.e. detection of trace elements in neighbouring soils vegetation and waters. Brown and Goldberg 35 used radioactivation in analyses of iron meteorites for gold gallium rhenium and palladium ; and Reed and Turkevitch 36 used i t for uranium. By its use Morris and Krewer 37 determined gallium in blende and Long 38 and later workers detlermiiied tant alum in minerals Smales 39 has reported the determination of nickel copper cobalt palladium gold and rubidium in samples of granite and diabase which have had world-wide circulation as standards in rock 3 5 Brown and GoltUx~g 8ciw7c0 1940 109 347; Goltlherg arid Brown Analyt.36 Reed ~ n t l 'I'urkevitc.h Nuturc 1955 176 7'34. 37 Morris aiitl Brewer G'eockim. C'ositiocJ~it)L. Acftr 1954 6 134. 38 Long Analyst 1951 76 644. 3D Smales Geochi,ri. C'osmochini. Acta 1956 8 300. Clteirz. 1950 22 3013. 102 QUARTERLY REVIEWS analysis. The same author l1 determined traces of uranium in monazite zircon and dunite and Jenkins 7 determined thorium in similar materials. Determination of traces of uranium and thorium in igneous rocks e.g. granite is important in attempts to estimate the age of the earth's crust (the radioactivation analyses would of course require to be supplemented by ?n alternative and more sensitive method of determination of lead- preferably by the mass-spectrometer isotope-dilution technique).In a different method of dating applicable to potassium minerals Moljk Drever and Curraii 40 determined 40A by radioactivation. Herr 23 and (independ- ently) Alperovitch and Miller 24 recently published preliminary evidence for the occurrence of a long-lived 9 * T ~ in Nature based on neutron activation and detection as 99mTc. Smales and Wiseman 41 have discussed the origin of the nickel found in deep-sea sediments which Pettersson and Rotschi 41a had ascribed to the deposition of meteoritic dust. Radioactivation analysis for nickel copper and cobalt in representative samples of globigerina ooze red clay and oceanic rocks from the At'lantic the Pacific and the Indian Ocean showed that the ratios nickel cobalt nickel copper and copper cobalt were not those accepted for meteorites (13-1 92 and 0.14 respectively) and were very close to the normal ratios for igneous rocks (3.5,1.1 and 3.0 respectively).Recent Harwell determinations by radioactivation of trace elements in sea-water include arsenic,42 rubidium cmium 43 and strontium.44 Biology.-Biological work often necessitates determining major con- stituents in minute samples e.g. sodium and potassium in single nerve fibres.45 Trace analysis is of great importance in studies of the metabolism of potentially toxic elements and of elements which appear to be essential to an organism. In certain cases determination of non-radioactive traces may be sounder than the alternative of following the distribution and excretion of an added radioactive tracer for sensitive systems may conceivably be modified by the radiation.This objection does not apply to activation of the samples after the required metabolic process has taken place. Harrison and Raymond 46 point out that administration of radioactive isotopes while giving valuable information on the relative retention and distribution of elements cannot give information on the absolute excretion rates of these elements. Radioactivation has been applied to problems of animal meta- bolism by Tobias and Dunn 47 who studied the distribution of gold through- out the tissues of a mouse 30 days after administration of 1 X g. 40 Moljk Drever' and Curran Nucleonics 1955 13 No. 3 44. 41 Smales and Wiseman Nature 1955 175 464. 414 Pettersson and Rotschi Geochim. Cosinochiin. Acta 1962 2 81.4 2 Smales and Pate ,4iaalyst 1052 77 188. 43 Smales and Salmon ibid. 1955 80 37. 4 4 Hummel and Smales Analyst in the press. 4 5 Keynes and Lewis Nature 1950 165 809. 46 Harrison and Raymond J. Nuclear Energy 1955 1 290. 47 Tobias and Dunn U.S. Atomic Energy Commission Unclassified Report No A.E .C.D. -2099B. JENKINS AND SMALES RADIOACTIVATION ANALYSIS 103 of inactive gold and by Harrison and Raymond 46 who studied the fzcal and urinary excretion of strontium and barium from a human subject on a normal diet. Smales and Pate illustrated the potentialities of radioactivation in the study of arsenic metabolism by carrying out analyses on the individual organs of a normal mouse.1 Few applications of radioactivation analysis appear to have been made to studies of the metabolism of trace elements during plant growth.The sensitivity of this method is adequate (at a flux of 10l2 neutrons (3111.12 sec.-l) to measure submicrogram amounts of P Ca K Cs Fe Mn Cu Zn Mo Co or C1 it is interesting that the method is not as sensitive as micro- biological assay for molybdenum. 48 Boron would require a specialised technique (see p. 98) and nitrogen and magnesium could be determined a t present only a t relatively high levels. I n certain cases the radioactivation method which determines the mass of the trace element may usefully supplement microbiological methods which determine the avaiZabiZity of the element. The determination of czsium and rubidium in seaweed by Smales and Salmon 43 showed that these elements are enriched with respect to sodium compared to the corresponding values for sea-water.Physics of the Solid State.-This includes problems as to the effect of trace impurities on electrical optical and mechanical properties of solids. For example the electrical properties of semiconductors such as germanium and silicon are profoundly changed by the presence of 1 part in loB of copper or nickel. The scope of the various methods of trace analysis in the semi- conductor field has recently been reviewed by one of us.49 Radioactivation has been used for the analysis of arsenic l6 down to 0-005 p.p.m. and for nickel 32 down to 0.1 p.p.m. in germanium by Smales and his co-workers ; for rare earths antimony molybdenum copper and zinc in germanium oxide and metal a t 0.1 p.p.m. by J a k o ~ l e v ~ ~ and for copper in germanium to 0.001 p.p.m.by S ~ e k e l y . ~ ~ Arsenic and copper have been determined by Smales 32 and by James and Richards 52 in silicon down to 0.0001 p.p.m. In some cases it is now possible to measure a range of elements simultaneously as discussed on p. 98 by y-spectrometry on the activated sample. This approach has been followed by Smales 32 * and by Morrison and C0sgrove.3~ Trace analyses by means of radioactivation have given important results in the study of phosphors and luminescent solids and of y-dosimeter glasses ".g. Delberg Glendenin and Yuster 53 measured thallium down to g. in potassium iodide crystals Grillot 5 4 measured g. of copper and 48 Nicholas Analyst 1952 77 629. 49 Smales J. Electronics 1955 1 327. 50 Jakovlev Geneva Conference on Peaceful Uses of Atomic Energy August 1055 61 Szekely Analyt.Chem. 1954 26 1500. 5 2 James and Richards Nature 1955 175 769. 5 3 Delbecq Glendenin and Yuster Analyt. Chem. 1953 25 350. 5 4 Grillot Compt. repzd. 1952 234 1775 ; cf. Bancie-Grillot and Grillot ibid. * Cf. also ref. 49. Paper 632. 1953 237 171. 1 (a4 QUARTERLY REVIEWS chlorine in zinc sulphide powders and Peirson 5 5 has identified and measured manganese as an impurity in metal phosphates of possible application to dosimeter preparation. In the metallurgical field traces of various impurities have been deter- mined by radioactivation in high-purity iron,56 32 aluminium 57 and mag- nesium 58 metal. Nuclear Physics.-The application of nuclear physics to the atomic- energy project has set high standards of purity for such basic materials as uranium thorium graphite light and heavy water beryllium zirconium and plutonium.For example traces of strong neutron-absorbers such as cadmium boron lithium or the rare earths cannot be tolerated in the uranium rods the aluminium cans or the graphite moderator of the B.E.P.O. reactor. Among other methods radioactivation analysis has been used at Harwell for the determination of the individual rare earths 59 and of vana- dium 60 in graphite of hafnium in zirconium,61 of oxygen in beryllium,21 and of magnesium chromium rubidium czsium silver antimony strontium and cobalt in a sodium-potassium alloy of possible interest as a reactor coolant .62 Practical applications of radioactivation analysis on intact samples The first distinctive feature of radioactivation analysis is its extreme sensitivity already discussed.The second important feature has only been touched on above that is the analysis of intact samples. This concept of non-destructive analysis has implications which extend even to industrial processing. There are at least four reasons for interest in an analytical method which might be applied non-destructively. (a) If it is very rapid and can be made automatic it could be applied to automatic process-control e.g. to sorting of mineral mixtures. Or it may be used in process control by providing a human controller with rapid information. This information need not always be of the highest precision provided it is prompt. ( b ) Certain samples are non-consumable by reason of their historical scientific or other interest or they may be highly toxic. ( c ) In certain cases any chemical treatment of t,he sample may be suspect owing to potential loss of the required constituent by e.g.volatilisation during acid digestion. (d) It is sometimes possible to derive inforination about the spatial or surface distribution of the required constituent in an intact sample. 55 Peirson unpublished work a t Harwell. 513 Albert Caron and Chaudron Gonzpt. rend. 1953 236 1030. 67 Idem ibid. 1951 233 1108. 58 Atchison and Beamer Annlyt. C'heiiz. 1052 24 1812 59 Cornish U.K. Atomic Energy Authority Report A.E.R.E. C/R 1224 1953. 130 Sinales anti Mapper '1J.K. Atomic Energy Authority Report A.E.R.E. C/R 131 Smales and Fullwood mipuhli~heti wolk. 6 2 Smales Geneva Conference on Peaceful Uses of Atomic Energy August 1955 607 1950. Paper 766. Cf. ref. 32. JENKINS AND SMALES RA DIOACTIVATION ANALYSIS 105 Radioactivation analysis offers advantages in each of these four fields.In most cases the induced activity is measured by its y-radiation which is normally able to penetrate the intact solid without prohibitive loss by self- absorption. The increasing use of y-scintillation spectrometers l3 facilitates the specific recognition of characteristic y-energies. It should be emphasised that this type of application accepts each sample very much on its own merits and the user must be aware of potential interference by extraneous induced activities-that is it is less universally applicable than the normal sequence of radioactivation analysis including radiochemical purification. Nevertheless in specific instances it is invaluable. The potentialities are illustrated by the following examples (i) Rapid routine analysis (automatic or otherwise) will normally measure a major constituent rather than a trace element and great interest will attach to short-lived induced activities (many of the data in the Table are irrelevant in this respect and the full compilations should be consulted).Gaudin and his co-workers 63 investigated the activit'ies induced in 150 specimens of 51 different mineral species by irradiation for 2-5 seconds in a pile at 1012-1013 neutrons sec.-l. The y-activities were measured 30 seconds after irradiation. Under these conditions even oxygen and fluorine give significant activities. The aut)hors concluded that it should be possible to separate felspars from iron minerals copper minerals from pyrites or galena from limestone but they stress that each ore-body would present a specific problem as the activities induced in a given mineral species varied considerably from one sample to the next owing to the varying content of impurities.Industrial application of this type of work would require a portable (or readily accessible) neutron source-Gaudin concludes that a flux of about 100 times that offered by curie-level radium-beryllium sources is desirable. It seems possible that this gap can be closed both by the use of more sensitive y-count!ers (a sodium iodide scintillation countm would give about 100 times the sensitivity of tJhe Geiger-Muller tube used by Gaudin) and by the development of neutron sources of higher flux. For example a large antimony-beryllium source might be prepared by pile irradiation of a massive compact or large sources of the Be(ct,n) type might become available by the use of cheaper ct-emitters than radium.Further particle accelerators might be designed specifically to offer moderate neutxon fluxes (by bombardment of an appropriate target) at a reasonable capital cost. (ii) Radioactivation analysis is being applied currently in the authors' laboratories 64 to the rapid determination of plutonium in metallic and other inorganic compounds by pile-irradiation for 5 seconds followed by quantita- tive measurement of a specific y-emitting fission product e.g. 1311 (0.36 MeV) or 140Ba-La (1.60 Mev) by means of a scintillation spectrometer. In certaiii cases the quantitative measurement can be carried out with moderate precision by use of an inexpensive ionisation chamber (radiation nioiiitor) without discrimination between individual y-emitting fission products.Chemical methods are of course available for the analysis of snch materials 63 Gaudin Senftle aiid Freyberger Eiig. Min. J . 1952 153 No. 11 95 174. 6 4 Atkins Phillips a i d Jenkins unpublished n ork. 106 QUARTERLY REVIEWS but radioactivation is quicker and avoids excessive handling of a toxic material. Again several workers have determined the arsenic content of human hair in medicolegal analyses. Griffon and Barbaud 65 measured the in- duced 76As activity without chemical separation at various points along intact hairs. This approach might be useful in demonstrating relatively high levels of arsenic content if backed up by y-spectrometry and decay studies to provide a specific identification.It would have the virtue of preserving the specimen intact as a legal exhibit. (iii) Examples of the application of radioactivation to the analysis of silicon semiconductors by direct y-spectrometry immediately after irradia- tion have been given on p. 99. This technique avoids possible loss e.g. of arsenic during chemical dissolution. (iv) Autoradiography of uranium and thorium inclusions in polished mineral specimens has been carried out by placing a sensitive emulsion against the surface and recording the tracks of the ionising ct-particles. Curie and Faraggi 31 extended this method by irradiating the specimen and emulsion in mutual contact in the Chatillon pile. The densely ionising recoil fragments from the fission of uranium atoms left heavy tracks and served to distinguish uranium from thorium segregations in polished granite samples.Further applications of activation analysis to autoradiography have either involved direct irradiation of the sample and the emulsion as above or have brought the sample and the emulsion into contact after irradiation. The method has been used in studying the segregation of uranium and of lithium in mineral~,~l 30 of arsenic,66 boron,67 and carbon 68 in steels of lithium 69 and possibly boron and iodine 70 in biological samples and of sulphur phosphorus and chlorine- and bromine-containing organic compounds on paper chromatograms.71 The autoradiography of an activated sample need not be confined to the use of contact emulsions or films. The y-emitting centres on the active surface could be made to form an image on a distant film by the use of a y-sensitive pin-hole camera it appears that a resolution of a t least inch can be achieved.72 Alternatively the surface may be scanned by a lead- shielded collimator and the activity a t any given spot recorded.Conclusions Radioactivation analysis is extremely sensitive for a large number of the elements and in this respect may be compared with emission and mass 6 5 Griffon and Barbaud Compt. rend. 1951 232 1455. 66 Kohn " Radioisotope Conference 1954 " Butterworths London 1954 Vol. 11 13' Faraggi Kohn and Doumerc Compt. rend. 1952 235 714. 68 Curie J . Phys. Radium 1952 13 497. 69 Ficq Compt. rend. 1951 233 1684. 7o Mayr Nucleonics 1954 12 No. 5 58. T 1 Winteringham Harrison and Bridges ibid. 1952 10 No. 3 52 ; Schmeiser and Jerchel A?zqezo.Chenz. 1953 65 366 490. 7 2 Mortimer Anger and Tobias U.S. Atomic Energy Commission Unclassified Report No. U.C.R.L.-2584 1954. p. 68. JENKINS AND SMALES RADIOACTIVATION ANALYSIS 107 spectrometry. The ubiquitous traces of contaminants in most chemical reagents do not normally cause difficulties in radioactivation this is perhaps its most distinctive feature. The ease with which both the sample and the standard may be activated at the same time and under closely similar conditions makes the method valuable in the calibration of reference samples later to be used as standards for other methods such as emission spectroscopy. Activation is best carried out in the high flux of thermal neutrons within a nuclear reactor. At least 50 reactors l7 of various types were known to be operating in 1955 including several comparatively inexpensive low-power reactors which are suitable for use as neutron sources by universities and industrial research organisations. The chief significance of radioactivation analysis will probably be its contribution to the measurement of trace elements a secondary interest may be that of rapid industrial control analyses on intact samples.