Radiochemistry questions

Journal of Radioanalytical and Nuclear Chemistry, Vol. 264, No. 2 (2005) 437–443

0236–5731/USD 20.00 Akadémiai Kiadó, Budapest © 2005 Akadémiai Kiadó, Budapest Springer, Dordrecht

Determination of radium and uranium isotopes in natural waters by sorption on hydrous manganese dioxide followed by alpha-spectrometry

R. Bojanowski,1* Z. Radecki,2 K. Burns2 1 Institute of Oceanology, Polish Academy of Sciences, Sopot 81-712, Poland

2 International Atomic Energy Agency, Agency’s Laboratories Seibersdorf, 2444 Seibersdorf, Austria (Received April 6, 2004)

Water samples, spiked with 133Ba and 232U radiotracers, are scavenged for radium and uranium isotopes using hydrous manganese dioxide which is produced in-situ, by reacting manganese (+2) and permanganate ions at pH 8–9. The precipitate is solubilized with ascorbic and acetic acids and the resulting solution filtered through a glass fibre filter GF/F to remove particulate matter. The radium is co-precipitated with barium ions by the addition of a saturated Na2SO4 solution where a small amount of BaSO4 suspension is introduced to initiate crystallization. The micro precipitate containing the radium is collected on a 0.1 µm membrane filter and the filtrate saved for follow-up uranium analysis. The 226Ra on the filter is determined by alpha-spectrometry and its recovery is assessed by measuring the 133Ba on the same filter using gamma-spectrometry. The filtrate containing uranium is passed through a Dowex AG 1x4 ion-exchange resin in the SO42– form which retains uranium while other ions are eluted by dilute (0.25M) sulphuric acid. Uranium is eluted from the column by distilled water, electrodeposited on a silver disc and the uranium isotopes and their recovery are determined by alpha-spectrometry. The method was tested on a variety of natural and spiked water samples with known concentrations of 226Ra and 238U and was found to yield accurate results within ±10% RSD of the target values.

Introduction

Concentrations of uranium and radium in natural waters vary widely spanning the range from a fraction of a mBq to thousands of Bq per liter.1 Analytical methods must, therefore, be sensitive and flexible enough to cover such a wide range in activity concentration. Alpha-spectrometry in combination with radiochemistry offers many advantages in this respect.2 It permits the analyst to identify and quantify alpha-emitting radionuclides from the energy and intensity of the peaks in the acquired alpha-spectra. The lowest detectable activity is limited only by the blank value, which typically can be kept under the one mBq level. However, preparation of sources of spectrometric quality from bulky natural samples requires many chemical operations in order to eliminate interfering elements. This task is particularly difficult for radium, for which few simple and specific separation methods are available. Since each step in the chemical separation process can result in unavoidable losses of the analyte, yield tracers must be used to account for them. This introduces an additional complication, because no convenient tracers exist for radium isotopes.

In some methods, no tracers are used and radium recovery is assumed to be quantitative or is deduced from independent recovery studies.3–7 In other methods 133Ba is used as the yield tracer.8,9 In the case of 133Ba, however, the analyst should be aware that there is experimental evidence that barium does not follow exactly the behavior of radium ions in all chemical reactions.9,10

* E-mail: [email protected]

Naturally occurring radium isotopes (223Ra, 224Ra)11–16 as well as anthropogenic 225Ra have been widely used as tracers for 226Ra determination in many methods based on alpha-particle spectrometry.17–22 Application of 223Ra or 224Ra as tracers requires, however, prior knowledge of their levels in the samples to be analyzed. The relatively short half-lives of these radionuclides (T1/2 = 11.2 and 3.64 days, respectively) combined with the in-growth of their progeny and their associated peaks which overlap those of 226Ra increase the confidence intervals about the final results. Radium- 225 has an advantage over the aforementioned radionuclides in that it does not exist in the natural environment. Being a pure beta-emitter, it decays through two of its progeny to 217At, which emits alpha- particles that do not interfere with the evaluation of the peaks in the spectrum produced by alpha-emitting radium radionuclides.

Calculation of 225Ra recovery using ingrowth-decay equations yields accurate results only if this radionuclide is present in the final source in a radiochemically pure state.23 Any 229 Th and 225Ac must be absent. They are removed, either at the stage of 225Ra tracer preparation or in the course of analysis. Due to this requirement, all methods dependent on this tracer are more complex and labour intensive. The wait time between source preparation and counting is approximately two weeks at which time the 217At/225Ra ratio attains maximum (ca. 0.44). Thus, if fast return of analytical data is the major concern, then 225Ra is not the optimum tracer and the analyst may wish to use 133Ba.

R. BOJANOWSKI et al.: DETERMINATION OF RADIUM AND URANIUM ISOTOPES IN NATURAL WATERS

438

There are several problems associated with the use of 133Ba as a tracer. First, its chemical and thermo- dynamical properties differ from those of radium. As a result, they behave differently during adsorption on sorbents, co-precipitation and ion-exchange.9,10 Second, two independent measurements must be performed on each sample, using two different pieces of instrumentation that are calibrated with different standards. Care must be taken to carry out all measurements under reproducible conditions or else the results may be excessively scattered and even seriously biased.

Although 226Ra measurements have a long history dating back to the beginning of the last century, and notwithstanding the fact that hundreds of analytical methods have since been developed,24 the quality of radium data are a matter of concern. A survey of seven interlaboratory comparisons made in the early 1990’s showed, that six out of seven had failed to reach target values within the uncertainties claimed by the participants.25 In one of these intercomparisons, the between-laboratory RSD ranged from 17 to 80%, with a median value of 20%. Variance due to the presence of systematic errors was estimated to be as high as 17%. A series of interlaboratory comparisons on 226Ra in sea water carried out under the GEOSECS programme disclosed systematic discrepancies of up to 20% while the precision of individual measurements was maintained at less than 5%.26 Such discrepancies are not uncommon in the real measurements world and call for some corrective action.

In this work an attempt was undertaken to develop a method which would minimize an impact of uncontrolled variables on the accuracy of obtained data. Attention was focused on the application of 133Ba for tracing the flow of 226Ra through the analytical steps from pre-concentration on aqueous manganese dioxide (MnO2aq.) to preparation of Ba(Ra)SO4 source for alpha-particle spectrometry. Uranium was incorporated in this method simply because the scheme provides for such an option with little additional effort and without any modification of the main 226Ra flow chart, except the addition of the 232U tracer.

Experimental

Reagents and apparatus

All reagents should be of analytical reagent grade purity. The solutions and all dilutions are made with distilled or deionised water.

Reagents for radium determination: Acetic acid: 100%; sulphuric acid, conc. (96–98%); hydrochloric acid, approx. 6M: dilute conc. acid in a proportion 1+1 by volume; ammonium hydroxide 25%: concentrated

aqueous solution; potassium permanganate, 0.2M: dissolve 7.90 g of KMnO4 in water, filter the solution through a glass fibre filter GF/F, dilute the filtrate to 250 ml and store the solution in an amber or dark glass bottle in a refrigerator; manganous chloride, 0.3M: dissolve 14.84 g of MnCl2·4H2O in water and dilute to 250 ml; ascorbic acid, 0.5M: dissolve 4.40 g of the reagent in 25 ml of water and dilute to 50 ml, keep in a refrigerator. The solution is stable for approximately two weeks; EDTA, 10% (0.2686M): dissolve 50 g of the di- sodium EDTA.2H2O salt in 400 ml of water together with 25 ml of conc. ammonia and dilute to 500 ml; sodium sulphate: prepare a saturated solution and filter it through a glass fibre filter GF/F; barium chloride, 10 mg Ba2+.ml–1: dissolve 2.443 g of BaCl2·2H2O in water and dilute to 100 ml; barium chloride, 1 mg Ba2+.ml–1: dilute the above solution 1 : 9 with distilled water; sodium hydrogen sulphate, 70%: dissolve 120 g of anhydrous sodium sulphate in 200 ml of water and slowly add 50 ml of conc. sulphuric acid; BaSO4 suspension, 0.5 mg.ml–1: mix 0.63 ml of barium chloride solution (10 mg Ba2+.ml–1) with 5 ml of sodium hydrogen sulphate solution in a Pyrex beaker and evaporate to dryness. Heat the residue over flame to obtain a transparent melt of sodium pyrosulphate. Dissolve the melt in 12.5 ml of saturated sodium sulphate and 12.5 ml of water; copper sulphate, 0.2686M solution: dissolve 6.71 g of CuSO4.5H2O in water and dilute to 100 ml; iso-propanol, n-propanol or ethanol; pH indicator papers (pH range 4–10); glass fibre filters 47 mm diameter: Whatman GF/F or equivalent; membrane filters 25 mm diameter, pore size 0.22 and 0.1 µm; vacuum filtration unit for filter diameter 25 mm and 47 mm; electrodeposition apparatus; ultrasonic bath; gamma-spectrometer; alpha- spectrometer.

Additional reagents for uranium determination: Nitric acid, conc. (65%); hydrochloric acid, 9M; perchloric acid, conc. (60 or 70%); sulphuric acid 0.25M and 2M solutions; Na2SO4 10% solution; plating solution: dissolve 100 g of ammonium sulphate in 500 ml of water and dilute to 1 l. Using a pH-meter carefully adjust the pH to 2.0±0.2 with conc. H2SO4; electroplating apparatus; ion-exchange columns φ 10 mm×100 mm with resin AG1×4, 100–200 mesh; 232U standard tracer solution, approx. 0.2 Bq.ml–1.

Procedure The course of analysis follows the scheme shown in

Fig. 1. Place a known aliquot (0.5–2 l) of filtered and

acidified (pH≤2) water sample into a suitably sized glass beaker, or an Erlenmeyer flask, and add accurately measured aliquots of 133Ba and 232U tracer solutions.

R. BOJANOWSKI et al.: DETERMINATION OF RADIUM AND URANIUM ISOTOPES IN NATURAL WATERS

439

Fig. 1. Analytical scheme for the determination of 226Ra and uranium in water samples

Boil the solution for 10 minutes, then add 1 ml of permanganate solution followed by 0.5 ml manganous dichloride solution. Adjust pH to 8–9 with ammonium hydroxide and keep the sample on a hot plate, without boiling, until the precipitate of MnO2.xH2O coagulates (~10 minutes). Add an additional 0.5 ml aliquot each of permanganate and manganous dichloride solution and continue heating for another 10 minutes. Collect the precipitate on a glass fibre filter GF/F, wash the precipitate with distilled water and place the filter in a small beaker. Return the filtrate to the original vessel and repeat a second precipitation of MnO2.xH2O by adding 1 ml each of permanganate and manganous chloride solutions in two sequences, as described earlier.

Repeat the filtration of the second MnO2aq. precipitate and combine the filters containing the

MnO2aq. precipitates and discard the filtrate. Add 3 ml of ascorbic acid and 0.5 ml of acetic acid to the filters, dilute with water to 5–10 ml and warm gently, with occasional swirling, until the MnO2 completely dissolves (as evidenced when the filters turn colourless). Transfer the contents (GF/F filters as well) to a filtration unit and filter it through a 0.22 µm Millipore filter. Wash the original beaker and the residue on the Millipore filter with a few small portions of distilled water, collect and transfer the filtrate to a new beaker. Transfer Millipore filter containing the undissolved residue to the original beaker, add 0.5 ml of EDTA, 3 drops of concentrated ammonium hydroxide and 5 ml of water, cover with a watch glass and heat gently for 10 minutes, swirling occasionally. Then filter the contents of this beaker through a new 0.22 µm Millipore filter

R. BOJANOWSKI et al.: DETERMINATION OF RADIUM AND URANIUM ISOTOPES IN NATURAL WATERS

440

and collect the filtrate and the washes (use distilled water) and combine with the first filtrate. Check the residue on this filter for 133Ba content on a gamma- spectrometer using a 133Ba standard prepared in the same geometry for comparison and discard the filter if the activity is less than 1% of the initial tracer quantity, which is normally the case. Otherwise repeat the EDTA leaching once more.

Add to the combined filtrate 3 ml of propanol and introduce 50 µg of Ba2+ ions if the original water sample is estimated to contain less than 50 µg of Ba2+. Place the beaker in a ultrasonic bath, add 5 ml of saturated sodium sulphate solution and, simultaneously, 0.1 ml of BaSO4 suspension. After 3 to 5 minutes remove the beaker from the bath and filter the solution through a 0.1 µm membrane filter. The filter chimney should be rinsed with 20% propanol solution prior to use, to keep the walls in a wet state. Retain the filtrate for uranium analysis. All subsequent washings should be carried out with 20% propanol. Accurately measure the 133Ba activity on the filter and in the filtrate by gamma- spectrometry ensuring that the system has been calibrated for these geometries, and calculate its recovery in both media. Place the filter in an alpha- spectrometer and measure it for sufficient time to acquire a satisfactory number of counts in the 226Ra peak.

Transfer the filtrate to a column filled to a height of 10 cm with anion-exchange resin AG 1x4 (100–200 mesh) that has been previously converted to sulphate form by passing through it 50 ml of 2M H2SO4 solution followed by 10 ml of 0.25M H2SO4. Wash the resin with 50 ml of 0.25M H2SO4 followed by 25 ml of 9M HCl. The washings are discarded. Elute the uranium with 70 ml of distilled water.

Add to the effluent 5 drops of 10% Na2SO4 and evaporate to dryness. Moisten the residue with a few ml of conc. HNO3 and a few drops of HClO4 and evaporate to dryness. Dissolve the remaining salts in 10 ml of plating solution and transfer to a plating cell equipped with platinum wire as anode and a silver disc as cathode. Plate for 2 hours at a current density of approximately 2 A.cm–2. Wash the disk with distilled water. Place the disc in an alpha-spectrometer and acquire the uranium alpha-spectrum.

Results and discussion

Preconcentration

Water samples with activities above 0.1 Bq/l can be analyzed directly using volumes not exceeding 50 ml. At larger volumes, radium recovery falls below unacceptable level which makes precise uncertainty assessment difficult. As the majority of natural waters have low radium and uranium contents, typically in the range from a fraction of mBq to several tens of mBq per litre, some kind of preconcentration must be applied to obtain meaningful data. Reduction of volume by evaporation has limited application because the gain on concentration rarely exceeds one order of magnitude and there is a risk of uncontrolled loss of radium to a solid phase which may form as evaporation continues.

Most preconcentration methods depend on adsorption or co-precipitation on suitable carriers, ion- exchange and extraction. Radium is known to co- precipitate with barium and lead sulphates, barium chromate, calcium and aluminium phosphates. In all these cases, an additional step must be, however, included to separate the analyte from the carrier and other interfering ions, and prepare it in a form suitable for alpha-spectrometry. Sorption on manganese dioxide is more suitable because it leaves behind nearly all major cations normally present in natural waters and requires no other carriers for the sorption to be efficient. It is also capable to efficiently adsorb uranium, enabling thereby inclusion of this element in the analytical scheme.

A series of experiments were carried out to determine the effect of principal cations on sorption efficiency of radium and uranium. Two-liter water samples containing 5 g/l of sodium chloride and varying concentrations of calcium and magnesium ions were spiked with known amounts of 133Ba and 226Ra ions and processed as described in the procedure above. The results have shown that more than 90% of barium and 95% of radium are removed by the first MnO2 precipitate provided that concentrations of calcium or magnesium ions do not exceed 0.2 and 1.0 g/l, respectively. When two successive precipitations are made, 95% of barium and nearly 100% of radium are recovered from solutions containing up to 1 g/l of either calcium or magnesium. A third precipitation ensures nearly quantitative recovery of radium, even though the recovery of barium may turn out to be appreciably lower. It is noteworthy that concentration of sodium chloride up to 100 g/l have no effect on sorption efficiency of both elements, which is nearly quantitative (Fig. 2). Two successive sorptions ensure quantitative recovery of barium and radium from up to 200 g/l NaCl solutions.

R. BOJANOWSKI et al.: DETERMINATION OF RADIUM AND URANIUM ISOTOPES IN NATURAL WATERS

441

Fig. 2. Effect of major cations on 133Ba sorption on MnO2aq. precipitate

Fig. 3. Method performance for the determination of 226Ra in various types of waters

Application of 133Ba as yield determinant for radium is complicated by the fact, that it is not scavenged with the same efficiency as radium. In all experiments with spiked water samples of various compositions, the measured Ra/Ba ratio was invariably found to be greater than one in the first MnO2 aq. precipitate and less than one in all subsequent precipitates. Since this ratio varies from sample to sample, and since a correction factor varies somewhat from sample to sample, it is recommended to strive for maximum recovery, monitoring it with 133Ba tracer. As a rule, two precipitations from low salinity waters (<1 g TDS/l) and three precipitations from samples with high salinity will satisfy this demand in most cases, but an experimental check is strongly recommended to ensure quantitative recovery of the 133Ba (Fig. 3).

At the onset of analysis, the acidified samples should be boiled for a few minutes to remove carbon dioxide before the pH is adjusted to the required level (8–9). Carbonate ions, if present, will keep uranium in solution and, at the same time, will enhance co-precipitation of calcium which, if present in moderate quantities, will interfere with subsequent analytical operations. To minimize adsorption of CO2 from the air, excessive exposure of the samples that have been made alkaline, to ambient air should be avoided.

Dissolution of MnO2aq. and separation of insoluble matter Manganese dioxide may be conveniently brought

into solution by reduction of manganese(IV) to its divalent state with hydrogen peroxide, hydroxylamine

R. BOJANOWSKI et al.: DETERMINATION OF RADIUM AND URANIUM ISOTOPES IN NATURAL WATERS

442

hydroxide or ascorbic acid. In the described procedure the latter was used with good effect. A 50% excess of the reducing agent, gentle heating and mixing with a spatula shortens the time for complete reduction to a few minutes.

The insoluble matter is separated by vacuum filtration and washed with a small volume of water followed by a hot alkaline EDTA solution. The latter step ensures complete recovery of radium and barium that are occasionally retained on the filter. These amounts are normally well below 1% when low salinity waters are analyzed, but may run as high as 10% for barium and 15% for radium in waters with high calcium or magnesium contents. The amount of EDTA used for washing the insoluble matter does not interfere with subsequent analytical steps.

Preparation of Ba(Ra)SO4 sources for alpha-particle spectrometry

Micro precipitation was applied as a method of source preparation due to its simplicity, rapidity and the high recovery it assures. The quality of spectra obtained with such sources depends on the amount of barium present in the original water sample as well as on the way the precipitate is formed. Satisfactory spectral resolution can be obtained on sources containing up to 0.5 mg/cm2 of precipitate, using the described procedure.

The Ba(Ra)SO4 precipitate is difficult to transfer onto the filter quantitatively as it has a tendency to adhere to solid surfaces. The kind of material with which the suspension is in contact does not seem to play

a role, provided the surface is kept wet. Surface active agents (surfactants) may reduce this tendency. In this work, iso-propyl alcohol was utilized as the surfactant. Its presence in the wash solutions, in moderate amounts, also helps reduce the amounts of Ba and Ra lost to the filtrate as it reduces the barium sulphate solubility in the wash solution. That loss should be kept as small as possible to reduce uncertainties resulting from different Ra/Ba ratios in the precipitate and the filtrate.

In this procedure, the amount of 133Ba in the filtrate was found to vary from less than 0.1% up to 6%, but typically was less than 1%. Since the Ra/Ba ratio in the filtrate was observed to be always less than one (typically in the 0.4 to 0.8 range), such losses can be neglected in most cases. However, it is recommended to confirm this assumption by measurement.

Losses by sorption on the walls of the beaker and filtration chimney were found to range from less than one up to 15%, most frequently from 3 to 5%. The Ra/Ba ratio in this fraction is equal to that in the main precipitate collected on the filters. The amount of barium recovered on the filter usually ranged from 88 to 98% and the corresponding radium recovery was estimated to range from 92 to nearly 100%.

Spectral purity was checked on samples spiked with natural uranium and thorium salts. The amount of 234U and 230Th contained by the Ba(Ra)SO4 precipitate was found to be 0.0% and up to 1%, respectively. Since the concentrations of thorium ions in filtered natural waters are usually low compared to 226Ra levels, application of additional purification steps was considered unnecessary.

Fig. 4. Method performance for the determination of uranium in various types of waters

R. BOJANOWSKI et al.: DETERMINATION OF RADIUM AND URANIUM ISOTOPES IN NATURAL WATERS

443

An experiment was carried out to study the behavior of 229 Th in the described analytical procedure in order to evaluate its usefulness as a source of 225Ra and 225Ac radionuclides for the determination of 226Ra recovery. The experiment confirmed that thorium remains in the filtrate and does not contaminate the Ba(Ra)SO4 precipitate. However, the actinium was captured nearly quantitatively. Attempts to keep it in solution with citric acid and EDTA were so far only partly successful. It follows from the foregoing, that the application of 229 Th instead of pure 225Ra as a source of 225Ra is not yet feasible.

Determination of uranium This part of the procedure was taken from an earlier

work,13 with slight modifications. No difficulty has been encountered while performing this part of analysis (Fig. 4). The columns can be re-used after regeneration with 2M sulphuric acid, as described in the preparation paragraph. Typical recoveries range from 60 to 85%.

Conclusions

The presented method is an attractive alternative to the existing, lengthy and more laborious methods based on alpha-particle spectrometry. Monitoring of all analytical steps with 133Ba facilitates identification of the operations responsible for significant losses of the analyte and helps to overcome this problem. Information acquired in the course of analysis is used to improve subsequent analytical operations so as to attain high analyte recovery and reduce the uncertainty of the final results.

The performance of the method was tested on a variety of the IAEA’s proficiency test samples and was found to be satisfactory for both low and high salinity waters at 226Ra concentrations covering four orders of magnitude (Fig. 3).

References

1. M. A. R. IYENGAR, The natural distribution of radium, in: The Environmental Behaviour of Radium, Vol. 1, Techn. Rept. Ser. No. 310, International Atomic Energy Agency, Vienna, 1990, Chapter 2–1, p. 59.

2. C. J. BLAND, Methods for measuring radium isotopes: alpha particle spectrometry, in: The Environmental Behaviour of Radium, Vol. 1, Techn. Rept. Ser. No. 310, International Atomic Energy Agency, Vienna, 1990, Chapter 3–3, p. 173.

3. J. C. BAYÉS, E. GÓMEZ, F. GARCIAS, M. CASAS, V. CERDA, Appl. Radiation Isotopes, 47 (1996) 849.

4. C. J. BLAND, Intern. J. Appl. Radiation Isotopes, 30 (1979) 557. 5. J. EIKENBERG, A. TRICCA, G. VEZZU, S. BAJO, M. RUETHI,

H. SURBECK, J. Environ. Radioact., 57 (2001) 109. 6. H. SURBECK, Appl. Radiation Isotopes, 53 (2000) 97. 7. A. S. GOLDIN, Anal. Chem., 33 (1961) 406. 8. J. C. LOZANO, F. FERNÁNDEZ, J. M. G. GÓMEZ, J. Radioanal.

Nucl. Chem., 223 (1997) 133. 9. C. W. SILL, Nucl. Chem. Waste Managem., 7 (1987) 239.

10. D. LANGMUIR, A. C. RIESE, Geochim. Cosmochim. Acta, 49 (1985) 1593.

11. M. L. ACENA, M. T. CRESPO, J. Radioanal. Nucl. Chem., 126 (1988) 77.

12. J. S. ALVARADO, K. A. ORLANDINI, M. D. ERICKSON, J. Radioanal. Nucl. Chem., 194 (1995) 163.

13. R. BOJANOWSKI, R. FUKAI, S. BALLESTRA, H. ASARI, Determination of natural radioactive elements in marine environmental samples by ion-exchange and alpha-spectrometry, Proc. 4 th Symp. on the Determination of Radionuclides in Environmental and Biological Materials, London, UK, 1983, Paper No. 9, 15 p.

14. R. G. DITCHBURN, N. E. WHITEHEAD, J. Radioanal. Nucl. Chem., 189 (1995) 115.

15. M. KOIDE, K. W. BRULAND, Anal. Chim. Acta, 75 (1975) 1. 16. W. C. LAWRIE, J. A. DESMOND, D. SPENCE, S. ANDERSON,

C. EDMONDSON, Appl. Radiation Isotopes, 53 (2000) 133. 17. P. BLANCO, J. C. LOZANO, F. VERA TOMÉ, Appl. Radiation

Isotopes, 57 (2002) 785. 18. M. T. CRESPO, Appl. Radiation Isotopes, 53 (2000) 109. 19. G. J. HANCOCK, P. MARTIN, Appl. Radiation Isotopes, 42 (1991)

63. 20. M. T. CRESPO, A. S. JIMENEZ, J. Radioanal. Nucl. Chem., 221

(1997) 149. 21. M. J. RODRIGUEZ-ALVAREZ, F. DANCHEZ, J. Radioanal. Nucl.

Chem., 191 (1995) 3. 22. M. YAMAMOTO, K. KOMURA, K. UENO, Radiochim. Acta, 46

(1989) 137. 23. K. A. SMITH, B. O. BARTLETT, J. Radioanal. Chem., 4 (1970) 383. 24. B. B. BANDONG, E. B. GUTHRIE, S. A. KREEK, K. U. MARSH,

J. M. BAZAN, F. A. ZAKA, M. A. RUTH, I. A. DUPZYK, W. L. EDWARDS, M. L. HALL, A Bibligraphic Review on the Analytical Methods for the Determination of Radium in the Environment, Rept. No. PB 96-178496, U.S. Department of Commerce, National Technical Information Service, Washington, D.C., 76 p. + III App.

25. A. R. WILLIAMS, Error analysis: evaluation of interlaboratory comparisons against a precision model, in: The Environmental Behaviour of Radium, Vol. 1, Techn. Rept. Ser. No. 310, International Atomic Energy Agency, Vienna, 1990, Chapter 3–11, p. 301.

26. Y. CHUNG, H. CRAIG, T. L. KU, J. GODDARD, W. S. BROECKER, Earth Planet. Sci. Lett., 23 (1974) 116.