386 BILEFIELD AND VINCENT : DE.TERMINATION OF CADMIUM (Vol. 86 Determination of Cadmium in Rocks by Neutron- activation Analysis BY L. I. BILEFIELD AND E. A. VINCENT (Department of Geology and Mineralogy, University Museum, Oxford) A neutron-activation method has been developed for determining 0.1 to 1 p.p.m. of cadmium in rocks. It depends on the separation of l15Cd (half-life 55 hours) with added cadmium carrier from a solution of the irradiated pow- dered rock by successive precipitations with trimethylphenylammonium iodide, 5,6-benzoquinoline and brucine. The precipitates are re-dissolved in a mixture of sulphuric and perchloric acids, and the cadmium is finally mounted for beta-counting as its brucinium salt. Standards are prepared by irradiating microgram amounts of cadmium acetate, The method has been applied to a suite of rocks from the Skaergaard Intrusion, Greenland, and also to the international standard rock samples G-1 and W-1.BECAUSE of the lack of sufficiently sensitive analytical methods, little is known of the occurrence of cadmium in igneous rocks other thian those rich in sulphide minerals. A colori- metric method1 has yielded most of the results hitherto available, but it is not claimed to be highly accurate. Cadmium has also been determined by ~pectrography~~~~~ and polaro- graphy5.6~7; the excitation of cadmium in flames has recently been reviewed.8 Neutron-activation analysis has been applied to the determination of cadmium in vinylite resins,g which were found to contain 100 to 400 p.p.m. of the element, and DeVoe and Meinkel0 have studied the radiochemical separation of cadmium from about twenty ele- ments.The results obtained when the method described here was applied to rocks and minerals of the Skaergaard Intrusion have already been discussedll from a geochemical standpoint. NUCLEAR CHEMISTRY From a consideration of the isotopic composition of natural cadmium and the half-lives of the expected (.n,y) products, it can be concluded that the only active nuclides present in significant amount in a sample of cadmium after irradiation for 1 week in a flux of thermal neutrons and then decay for 24 hours should be l15Cd (half-life 55 hours) and its metastable form 1lbCd (half-life 43 days). Since the cross-section for formation is eight times greater for 115Cd than for llbCd and since the shorter-lived nuclide attains 87 per cent.of its saturation yield in 1 week (against 9 per cent. for llhCd), it follows that the observed activity of a sample of cadmium treated as indicated above should be mainly due to l15Cd. This conclusion is substantiated by the behaviour of the cadmium activity separated from rock samples and standards, and, since a half-life of 2.3 days allows ample time for chemical treatment, the determination of traces of cadmium by neutron activation seems feasible. The decay scheme of 115Cd involves a radioactive daughter isotope, l151n (half-life 4.5 hours), so that the total activity of a freshly prepared sample of 115Cd increases during the first few hours. In general, for the scheme- A, AB A (active) -+ B (active) -+ C(stab1e) where AA and AB are the respective disintegration constants, if AA is less than AB maximum total activity occurs at a time after separation of pure A given by the equation12- In this instance, tmax.is 12.9 hours. After maximum activity has been reached, transient equilibrium is established between parent and daughter isotopes, and the total activity decays smoothly with the parent's half-life. In a method based on l15Cd, therefore, it is necessary to wait about 24 hours after separation before the sources are counted.June, 19611 IN ROCKS BY NEUTRON-ACTIVATION ANALYSIS 387 Consideration must be given to the possible occurrence of nuclear reactions yielding llTd in samples of rock, but not in standards, thereby invalidating comparison between the two; three such reactions are- (i) l161n (n,p) l16Cd (ii) ll*Sn (%,a) l15Cd (iii) 236U (n, fission) l15Cd 49 60 The types of reaction involved in (i) and (ii) rarely occur by the action of therrnal neutrons on elements of atomic number greater than about 40,13 and reaction (iii) is unlikely to be significant, since cadmium lies in the trough of the uranium fission yield curve.Experiments with indium, tin and uranium monitors showed that the maximum error that could be caused by the presence in a sample of rock of all three of these elements in concentrations similar to that of cadmium was about 0.5 per cent.; this was negligible for our purpose. Other potential sources of error are self-shielding and self-absorption. Self-shielding during irradiation is likely to be exceptionally severe for cadmium, owing to its high capture cross-section for thermal neutrons, and so the amount of a pure cadmium compound that can be used as an irradiation standard is expected to be correspondingly low.If self-shielding occurs, a large sample exhibits a lower specific activity than does a small one. Two separate irradiations of cadmium acetate dissolved in de-mineralised water, showed that only little self-shielding occurred in the range 0.01 to 1.0 pg of cadmium; the results were- Irradiation No. . . . . .. .. 1 2 Cadmium irradiated, p g .. . . 0.926 0-672 0.107 0.0115 0.908 0.474 0.095 0.0093 A A f \ I \ Activity per pg, counts per minute . . 3594 3466 3883 3899 3835 3973 4007 4103 If self-shielding occurs in the samples of rock, the lower specific activity of a large sample of a particular rock will lead to an under-estimate of its cadmium content when compared with the result from a smaller sample irradiated at the same time.The results in Table I do not show this effect. TABLE I REPLICATE RESULTS FOR SAMPLES OF ROCKS Sample No. Type of rock Weight of sample, Cadmium found, mg p.p.m. E.G.5321 . . .. . . Gabbro 92-6 0.32 208.6 0-34 E.G. 4507 . . .. . . Gabbro 105-5 0.125 216.1 0-131 w-1 . . .. . . Diabase 101.3 0.34 116.2 0.30 227.3 0-33 144.5” 0-34 * Separate irradiation. Self-absorption is severest when low-energy beta-particles are counted. The results below show that the apparent activity of a source of cadmium “brucinate” having constant surface area slowly decreases as its thickness (proportional to the amount of cadmium present) increases.To minimise this cause of error, it was desirable to restrict the thickness of these sources to the equivalent of 5mg of cadmium or less. Cadmium present, mg . . .. .. 15 12 9 6 3 Activity, counts per minute . . .. 235 242 247 273 267 ANALYTICAL CHEMISTRY In developing a method for isolating radiochemically pure cadmium from the complex mixture of isotopes produced in a rock by irradiation with neutrons, a survey was made of the organic precipitants available. The most popular gravimetric reagents for cadmium have been14 anthranilic acid, quinaldinic acid and oxine, but these were insufficiently specific for our purpose. 2-(o-Hydroxyphenyl) benzoxazole is said to be highly specific for cadmium,16 but this has been questionedlOsl6; further, the necessary careful adjustment of pH is tedious.388 BILEFIELD AND VINCENT : DETERMINATION OF CADMIUM [Vol.86 In the presence of potassium halide, brucine sulphate forms a white precipitate with cad- miuml7; this brucinium salt contains only 9.2 per cent. of cadmium and so is a convenient means of handling a small amount of the element. Cobalt, nickel, copper, iron, chromium and zinc form similar precipitates. 5,6-Benzoquinoline and trimethylphenylammonium iodide behave alike as precipitants for cadmium1* and are not subject to interference from nearly all the elements just mentioned. There is interference from copper, lead, zinc and bismuth, however, but these elements can be displaced by the addition of elemental iron.Hexamine allylo-iodide has been recommended for use with cadmium,19 but the results are not always satisfactory . 2o On the basis of this survey and experiments to ascertain the characteristics of the precipitates involved, 5,6-benzoquinoline, trimethylphenylammonium iodide and brucine were selected as precipitants, with the brucinium salt as the “plating-out” form. Aqua regia was found to be unsatisfactory for re-dissolving the precipitates, but a mixture of sul- phuric and perchloric acids (17 + 3) gave good results. I t was ascertained that the precipitated cadmium “brucinate” was of reproducible known composition. To assist in attaining radiochemical purity, scavenge steps with ferric hydroxide and silver iodide were included. Together with the iron solution were added manganese and magnesium carriers, which, when made alkaline, remove 56Mn and l14“In, respectively21 ; otherwise, indium closely follows cadmium.Holdback carriers were used to assist the puri- fication each time a cadmium compound was precipitated. PREPARATION OF STANDARDS AND SAMPLES FOR IRRADIATION- AnalaR cadmium acetate was chosen for preparing the irradiation standards in preference to the chloride, which gives rise to excessive extraneous activity, or the metal, which is not available in a pure condition and would suffer from self-shielding. About 0.2 ml of cadmium acetate solution, containing 0.01 to 1.Opg of cadmium, is introduced into a tared silica ampoule. After the ampoule has been re-weighed and its contents evaporated to dryness, it is sealed and sent for irradiation.As the concentration of the acetate solution is known, the exact amount of cadmium taken can be calculated. After return from irradiation, each ampoule is rinsed in a warm dilute solution of cadmium, dried and cut open. The portions are immersed in hot water for about 30 minutes, with thorough rinsing, and the solution is made up with washings to 50 ml. Samples of rock are ground to a fine powder in an agate mortar and sealed in lengths of silica or (more conveniently) polythene tubing marked with black Chinagraph pencil ; 100 to 300 mg of powder is a convenient amount. The tubes should be only two-thirds full, so that, after irradiation, the ends can be cut off and the powder removed by tapping without loss. Samples and standards were irradiated for 1 week in a flux of 10l2 neutrons per sq.cm per second. REAGENTS- Trimethyl~henylammonium iodide solution--Prepare an aqueous 3 per cent. solution of the reagent, and filter to remove dark insoluble matter. Trimethylphenylammoniu~ iodide wash solution-Dissolve 1 g each of the reagent and potassium iodide in 200 ml of water, and filter. Brucine sulphate so1utio.n-Dissolve 1 g of brucine in 100 ml of cold 20 per cent. sulphuric acid; alternatively, use an aqueous 1 per cent. solution of brucine sulphate. The yellow colour that develops when the solution is stored is not detrimental. Brucine wash solution-Mix 40 ml of the brucine sulphate solution, 15 ml of 20 per cent. potassium bromide solution and 100ml of water. 5,6-Benzoquinoline solution-Dissolve 2-5 g of the reagent in 100 ml of 0.2 N sulphuric acid.5,6-Benxoquinolinne wash solutiort-Mix 10 ml of the 5,6-benzoquinoline solution, 10 ml of 0-2 N potassium iodide and 150 ml of water, and add a few crystals of sodium sulphite. Holdback carrier solution A-Prepare to contain approximately 1 mg each of lead, copper, bismuth, mercury, arsenic and tin per ml, as nitrates or chlorides. Holdback carrier solution B-Prepare to contain approximately 1 mg each of cobalt, nickel, manganese, chromium, magnesium, zinc and sodium per ml, as nitrates or chlorides. METHODJune, 19611 IN ROCKS BY NEUTRON-ACTIVATION ANALYSIS 389 Holdback carrier soZution C-Prepare to contain 1 mg each of copper, magnesium, man- ganese, potassium and sodium per ml, as chlorides or sulphates, but not as nitrates.Perchloric - sulphuric acid mixtare-Add 3 ml of 60 per cent. perchloric acid to 17 ml of concentrated sulphuric acid, with stirring. Iron-scavenge solation-Prepare to contain 10 mg of ferric iron (as ammonium ferric sulphate), 5mg of manganese (as potassium permanganate) and 5 mg of magnesium (as sulphate) per ml. PROCEDURE FOR SAMPLES OF ROCK- Take norrnal precautions for handling moderate levels of radioactivity, and begin work as soon as the irradiation can is returned. Bring the powdered rock into solution in the presence of 15 mg of cadmium, usually by Rafter’s method22 (dissolve the cake resulting from the fusion with sodium peroxide in the minimum volume of water containing the cadmium carrier as sulphate). To the solution add 5ml of concentrated hydrochloric acid, and evaporate to incipient dryness; repeat this step, adding 1 ml of concentrated nitric acid if necessary, until only silica has not been dissolved.To the moist solid add 2 ml of holdback carrier solution A, 1 g of sodium sulphite, 10 ml of 20 per cent. sulphuric acid (7 N) and a few bright iron nails. Cover, and boil for about 20 minutes until the yellow or green of the solution is obscured by the dark colour of iron and the displaced metals. Decant the hot mixture into a centrifuge tube, spin, and decant the supernatant solution (or remove it by filtration). Wash the solids by centrifugation with hot water, and combine the solutions until 25 to 35 ml of mixture approximately 2 N in sulphuric acid are obtained.Add 1 g each of potassium iodide and sodium sulphite and 2 ml of holdback carrier solution B; then add, with vigorous stirring, excess of trimethylphenylammonium iodide solution. Set aside for a few minutes, spin in a centrifuge, reject the supernatant solution to active waste, and wash the precipitate once with trimethylphenylammonium iodide wash solution. At this stage, the activity is generally down to tracer level. Iodine is evolved, and then the solution slowly clears ; if necessary, add further small volumes of the acid mixture until a pale yellow or colourless solution results. Cool, dilute until no further heat is evolved, and cool again. Add 2ml of holdback carrier solution C, 8ml of 20 per cent. potassium bromide solution and 16ml of brucine sulphate solution.Stir vigorously to coagulate the precipitate, set aside for at least 10 minutes, spin in the centrifuge, and reject the supernatant solution. Wash the precipitate once with brucine wash solution, and boil it with small volumes of the acid mixture to obtain a pale solution. Cool, dilute a little, add 1 ml of iron-scavenge solution, and stand the tube in ice - water mixture. Cautiously add ammonia solution, sp.gr. 0.880, until an excess is present, boil, filter, and wash the residue with hot ammonia solution. Stir into the filtrate 0.1 g of potassium iodide, and add 1 ml of a silver nitrate solution containing 10 mg of silver per ml. Boil to coagulate the silver iodide, filter, and wash the precipitate with ammonia solution. Dissolve 1 g each of Rochelle salt and sodium sulphite in the filtrate, cool, and add concentrated sulphuric acid dropwise until the solution is yellow and acid.Add 2 g of potassium iodide and 2 ml of holdback carrier solu- tion B, and dilute to 30ml with water. Add 5ml of 5,6-benzoquinoline solution, with thorough stirring, and warm, if necessary, to initiate coagulation. Allow the precipitate to settle, add a surface layer of ethanol to repel the solid, and spin in the centrifuge. Discard the supernatant solution, and wash the precipitate once with 5,6-benzoquinoline wash solution ; stir well to dissolve any precipitated reagent, and add ethanol before centrifugation. Dissolve the residue in the minimum of acid mixture by boiling, and cool the clear solution. Dilute a little, and add 2 ml of holdback carrier solution C and some decolorising charcoal.Boil for 1 minute, allow to cool to about 50” C, add 5 ml of 20 per cent. potassium bromide solution, and filter. To the cooled filtrate, which should be nearly colourless, add 10ml of brucine sulphate solution, stir to initiate precipitation, and set aside for at least 20 minutes. Wash the white precipitate twice with brucine wash solution and then three times with a mixture of ethanol and diethyl ether (1 +- ti), breaking up the lumps of solid. Suspend the “brucinate” in a little pure diethyl ether, transfer to a tared aluminium counting tray, distribute the solid evenly, and dry under a heat lamp. Allow to cool, weigh, and count next day in a beta- counter. Boil the precipitate with 5 ml of perchloric - sulphuric acid mixture.Stir, spin in the centrifuge, and reject the supernatant solution.390 BILEFIELD AND VINCENT : DETERMINATION OF CADMIUM [Vol. 86 PROCEDURE FOR STANDARDS- Prepare 50-ml portions of standard solutions containing 0.01 to 1.Opg of cadmium as described under “Preparation of Standards and Samples for Irriadation.” Evaporate aliquots (containing 0.1 pg of cadmium) of the concentrated standards and the total volumes of the less concentrated ones to 5 ml. To each solution add 10 mg of cadmium carrier, as sulphate solution, and then 1 ml of iron-scavenge solution and excess of ammonia solution. Filter, wash the residue, add the washings to the filtrate, and stir in 0.1 g of potassium iodide. Add 1 ml of the silver nitrate solution, boil to ,coagulate the silver iodide, and filter or spin in the centrifuge. Dissolve in the filtrate 1 g each of Rochelle salt and sodium sulphite, cool, and acidify with concentrated sulphuric acid.Add 1 g of potassium iodide, 2 ml of holdback carrier solution B and excess of trimethylphenylammonium iodide solution, with vigorous stirring. Allow the precipitate to settle, spin in the centrifuge, discard the supernatant solution, and wash the precipitate once with trimethylphenylammonium iodide wash solution. Dis- solve it by boiling in the minimum amount of perchloric - sulphuric acid mixture, cool, dilute a little, and add 2 ml of holdback carrier solution C and 0.5 g of potassium chloride. Stir, filter, add 5 ml of 20 per cent. potassium bromide solution to the filtrate, thoroughly cool, and then add 10ml of brucine sulphate solution. Stir well, set aside, wash the precipitate, and mount it for counting as described for samples of rock.Wash the residue once with dilute ammonia solution. TESTS FOR RADIOCHEMICAL PURITY- The simplest way to check the radiochemical purity of the sources is to follow the decay of the activity from 24 hours after preparation for about a week. Since the sources prepared from an average rock give only 50 to 100 counts per minute at the outset, the activity usually merges into the background after this period. For the same reason, it is not usually convenient to obtain absorption curves, and hence the maximum beta-energy, with sources from rock samples. When sources prepared from standards, together with aluminium absorbers, are used, maximum beta-energy occurs at about 460 mg per sq.cm, which corresponds closely to the published value of 1.1 MeV for l15Cd. The observed half-life is usually 56 to 57 hours, slightly longer than the accepted value for 115Cd. The difference is ascribed to the presence of a small amount of l1wCd (half-life 43 days) ; since this is present in both samples and standards, which contain roughly the same weight of cadmium, it does not affect the validity of the comparison between the two. The small “tail” observed in the absorption measurements is attributed to the same cause (for 11hCd, maximum beta-energy is 1.6 MeV, equivalent to 740 mg per sq. cm of aluminium). RESULTS AND DISCUSSION OF THE METHOD The precision of the proposed method may be judged from the results in Table I1 for a series of rocks from the Skaergaard Intrusion, East Greenland.TABLE I1 CADMIUM CONTENTS FOUND IN SKAERGAARD ROCKS Sample No. Type of rock ’ * } Transgressive acid granophyre E.G. 5259 .. E.G. 4489 . . .. E.G. 4332 . . . . Basic hedenbergite granophyre E.G. 4328 . . . . Fayalite ferrogabbro, purple band E.G. 5196 . . . . Melanocratic ferrogabbro E.G. 5181 .. . . Average rock E.G. 5321 .. . . Plagioclase cumulite E.G. 5052 .. . . Middle gabbro (olivine-free) a } Lower olivine gabbro E.G. 5087 . . E.G. 5086 .. .. E.G. 4526 . . . . Gabbro picrite E.G. 4443 .. . . Border group E.G. 4507 . . . . Chilled marginal gabbro (initial magma) Cadmium content found, p.p.m. 0.57, 0-59, 0.54 (mean 0.57) 0.44, 0.44 (mean 0-44) 0-27, 0.28 (mean 0-28) 0.28, 0-27 (mean 0.28) 0.11, 0.12 (mean 0-12) 0.38, 0.43 (mean 0.40) 0.34, 0.32 (mean 0.33) 0.08, 0-09 (mean 0.09) 0-20, 0.19 (mean 0.20) 0.08, 0.09 (mean 0.09) 0.11 0.13, 0-11, 0.13, 0-13 (mean 0.13) { The accuracy of our results can only be a,ssessed for the standard rocks G-1 and W-1, The results in Table I11 show that agreement for which independent analyses are available.June, 19611 IN ROCKS BY NEUTRON-ACTIVATION ANALYSIS 391 with other methods is satisfactory for G-1, but not W-1, and, although we have endeavoured to eliminate potential sources of systematic error from our method, further independent determinations are clearly required.TABLE 111 CADMIUM FOUND IN STANDARD ROCKS BY VARIOUS METHODS Method Cadmium found in- Reference A \ No.* granite G-1, p.p,m.diabase W-1, p.p,m. Neutron activation (proposed procedure) - 0.054, 0.056, 0.066 0.30, 0.33, 0-34, 0.34 Anion-exchange enrichment and spectro- (mean 0-06) (mean 0-33) graphy . . .. .. .. .. 23 - 0.07, 0.09 (mean 0.08) Square-wave polarography . . . . 7 0.06, 0.08, 0.09, 0.10 0.24, 0.27, 0.33, 0.35 (mean 0.08) (mean 0.30) * See reference list below. The proposed procedure appears to be serviceable, but is probably capable of refinement. It may be possible with some samples to omit certain of the precipitation steps, so increasing the chemical yield (at present about 25 per cent.) and the sensitivity of the method. When carried out as described above, analysis of twelve samples of rock, with four standards, takes 4 days from the return of the irradiated material.I 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. We thank Mr. A. A. Smales, U.K. Atomic Energy Research Establishment, Harwell, for suggesting the determination of cadmium as a problem in radioactivation analysis and Professor L. R. Wager for making available the Skaergaard material and for his interest in the work. The necessary irradiations in the Harwell pile BEPO were arranged through the IsotoDe Division of the Atomic Energy Research Establishment. REFERENCES Sandell, E. B., and Goldich, S. S., J . Geol., 1943, 51, 99 and 167. Oftedal, I., Skr. Norske VidenskAkad., 1941, No. 8. Marks, G. W., and Jones, B. M., U.S. Bureau of Mines Report of Investigation No. 4363, Washing- Rusanov, A. K., and Alekseeva, V. M., Zavod. Lab., 1945, 11, 181. Miholic, S., J . Chem. Soc., 1950, 3402. Smythe, L. E., and Gatehouse, B. M., Anal. Chem., 1955, 27, 901. Carmichael, I., and McDonald, A., Geochim. Cosmochim, A d a , 1961, 22, 87. Gilbert, P. T., jun., Anal. Chem., 1959, 31, 110. Brooksbank, W. A., Leddicote, G. W., and Mahlman, H. A., J . Phys. Chem., 1953, 57, 815. DeVoe, J. R., and Meinke, W. W., Anal. Chem., 1959, 31, 1428. Vincent, E. A., and Bilefield, L. I., Geochim. Cosmochim. Acta, 1960, 19, 63. Cook, G. B., and Duncan, J. F., “Modern Radiochemical Practice,” The Clarendon Press, Oxford, -- , op. cit., p. 205. Weliher, F. J., “Organic Analytical Reagents,” D. Van Nostrand Co. Inc., New York, Volume Walter, J. L., and Frieser, H., Anal. Chem., 1952, 24, 1985. Smit, J . van R., D.Phi1. Thesis, University of Oxford, 1966. Nikitina, E. I., Zavod. Lab., 1938, 7, 409. Pass, A,, and Ward, A. M., Analyst, 1933, 58, 667. Evrard, V., Ann. Chim. Anal., 1929, 11, 322. Hurd, L. C., and Evans, R. W., Ind. Eng. Chem., Anal. E d . , 1933, 5, 16. Jacobi, E., Helu. Phys. Acta, 1949, 22, 66. Rafter, T. A., Analyst, 1950, 75, 485. Brooks, R. R., Ahrens, L. H., and Taylor, S. R., Geoclzim. Cosmochim. Acta, 1960, 18, 162. ton, 1948. 1952, p. 49. IV, 1948. Received February 15th, 1961