ANALYST, SEPTEMBER 1993, VOL. 118 Under-determination of Strontium-90 in Soils Containing Particles Irradiated Uranium Oxide Fuel Deborah H. Oughton and B. Salbu Isotope and Electron Microscopy Laboratories, Agricultural University of Norway, 1432 As, Norway Tom L. Brand and J. Philip Day Department of Chemistry, University of Manchester, Manchester, UK M 13 9PL Asker Aa rkrog Environmental Science and Technology Department, Risaf National Laboratory, DK-4000, Denmark 1101 of A much used method for the determination of 90Sr in soil depends on extraction of the soil with 6 mol I-’ HCI, followed by P-counting. For soils containing particles of irradiated uranium oxide, we postulate that this extraction could result in a variable underestimate owing to incomplete chemical recovery of strontium from the uranium oxide matrix.In experiments on two soils, collected from near Windscale (now Sellafield) in 1956 and near Chernobyl in 1990, about 25% of the total 90Sr present in the soil was recovered in 24 h by HCI extraction at room temperature, and the presence of high-radioactivity particles both before and after extraction was demonstrated by autoradiography. For a further 11 particle-containing Chernobyl soils, 90Sr determination, based on classical HCI extraction, yielded, on average, 54% (range 3345%) of the total 90Sr, as determined by oxidative alkaline fusion. While we accept that HCI extraction is well established as a reliable method for the determination of soil 90Sr derived from weapons fallout, we conclude that more rigorous analytical pre-treatment is essential in instances where the 90Sr may be associated with fuel particles.Keywords: Strontium-90 determination; soil extraction; uranium oxide fuel particle; Windscale; Chernob yl Historically, on at least two occasions, particles of uranium oxide from irradiated reactor fuel have been released in significant amounts into the environment.1 In the early 1950s, the operation of the two Windscale piles (at Sellafield, UK) resulted in the continuous release over several years of an estimated 20 kg of uranium, in the form of relatively large particles (up to 700 pm in length) of oxidized uranium fuel.2 More recently, the Chernobyl Reactor accident, in 1986, released a large amount of radioactive particles .3-7 These varied in size, shape and composition, but were mainly of two types.Firstly, fragmentation of the reactor core during the initial explosion released about 3.5% of the irradiated uranium dioxide fuel: the larger fuel particles (20-400 pm) were deposited within about 60 km of the reactor, while smaller particles were carried at least 1500 km. Secondly, during combustion of the reactor core, following the explo- sion, small particles of irradiated fuel and ‘condensation’ particles (Le., inactive or active materials on which the volatile fission products had condensed) were released. These par- ticles were carried considerable distances from Chernobyl, and constituted the major component of long-lived fission products deposited at distances over 30 km (that is, outside the so-called ‘30 km exclusion zone’). The radiochemical analysis of soils containing discrete radioactive particles can present special problems, apart from the difficulty of obtaining representative samples.In direct y- ray spectrometry (e.g., for 137Cs), inhomogeneity could clearly give rise to errors in calibration, whereas for determi- nations requiring chemical isolation of the nuclides, for example, in the determination of 90Sr or actinides, problems arise if the analyte nuclide is not completely extracted (or, more precisely, in instances in which a yield monitor is used if complete exchange between the analyte and yield monitor nuclides is not achieved). Such problems have been recog- nized for some time, and the subject has been reviewed recently.8 In particular, Sill and co-workers9JO showed that, for soils containing particles of refractory uranium or pluto- nium oxides, extraction methods involving HCl and/or HN03, which would be appropriate for soils containing only weapons fallout, achieve very low recoveries (typically <30%) of plutonium and other actinides.In such instances, fusion techniques, resulting in complete dissolution of the sample, were recommended.10 In relation to 90Sr determination in soils containing fuel particles, a crucial part of the analytical procedure is the initial chemical extraction of the 90Sr from within the fuel particle matrix. Traditionally, the determination of 90Sr in soil has depended upon an extraction procedure using HCl, typically 6 mol 1-1 HCl at room temperature.llJ2 This extraction step, which was developed in the 1950s and remains a recommen- ded procedure ,8913 achieves good recovery of 90Sr originating from weapons fallout, where 90Sr is deposited in condensed, sub-micrometre particles.However, general problems of incomplete recovery have been recognized, and some labora- tories now routinely use much more vigorous conditions to ensure complete extraction of strontium.8 As uranium oxide is relatively inert to cold HCl (although more reactive in oxidizing conditions) ,I4 it would be expected that complete recovery of 90Sr from fuel particles might not be attained by HC1 extraction alone. Hence, in instances where the fuel particle component was significant, the method could give rise to an underestimate of deposited 90Sr in soils.To test this hypothesis, 90Sr determinations have been carried out by HCl extraction of particle-containing soils from two sources: from near Sellafield (then Windscale), collected in 1956; and from within 15 km of Chernobyl, collected in 1990 and 1991. The efficiency of the HC1 extraction was then assessed by determination of the remaining WSr , following total dissolution of the residues. Experimental Two types of experiment have been carried out. In the first, intended to examine the behaviour of identifiable radioactive particles during extraction, samples of Chernobyl and Wind- scale soil were sequentially extracted with HCl(6 moll-l; two portions), ammonium acetate and finally (for total analysis) a mixture of HF and HN03. The presence of radioactive particles in the samples was demonstrated by autoradiography both before and after the HC1 extractions.In the second experiment, intended to investigate the extent to which 90Sr in particle-containing soils might be under-1102 ANALYST, SEPTEMBER 1993, VOL. 118 estimated in normal laboratory practice, 90Sr in 11 samples of Chernobyl soil was determined by using HC1 extraction, and the results were compared with the total obtained following oxidative alkaline fusion of the residues. Sample Origins and Preparation The Chernobyl soils were taken to a depth of 5 cm, from Bourykovka (May, 1990 and 1991) and from Novo Shepilichi (May, 1991), 15 and 5 km, respectively, from the Chernobyl reactor. Stones and roots were removed, and the soils were dried, crushed and homogenized by gentle grinding.Soils from these locations have previously been characterized both at the Isotope Laboratory, Norway, and the Rise National Laboratory, Denmark, and were found to contain relatively high activities of the refractory fission products (e.g., 144Ce), consistent with the presence of fuel particles.7.15 The Windscale soil was one of a number of grass and topsoil samples collected between 1955 and 1957 from a garden in the village of Seascale, approximately 3 km from the Windscale piles, as part of a contemporary investigation of the emissions of oxidized fuel particles. The piles operated from 1951 to 1957, when both were shut down because of a major fire in one of them.16 Particle releases were first discovered in mid-1955, and were caused by oxidation of incorrectly discharged (metallic uranium) fuel elements in the ducts leading from the air-cooled reactor core to the discharge stack, which housed an ineffective filter system .2717 One such particle (approxi- mately 300 pm in length) was isolated, at the time, from the Seascale garden18 and some of its characteristics have already been reported.7 The Seascale soil samples were stored dry and intact, together with their surface vegetation, until 1991, when they were delivered to the University of Manchester, and thereafter to the Isotope Laboratory in Norway.A number of additional experiments to characterize these materials have now been carried out, and will be reported separately. The Windscale soil sample described in the present paper had been collected in February, 1956, to a depth of approximately 1 cm.In 1991, the vegetation and obvious stones were removed, and the soil (15.2 g) was further dried, crushed, and mixed gently, yielding seven 2 g portions for further investigation. These were first screened by y-ray spectrometry. The only fission product readily detectable was *37Cs, averaging 0.12 k 0.05 Bq g-' in six of the sub-samples. The seventh portion contained (in total) 22.8 Bq of 137Cs, and was assumed to contain one or more fuel particles. This high- activity portion, and a low-activity sample for control purposes, was further examined by autoradiography and then subjected to chemical extraction with HCl. Sequential Extraction Experiments Two 2 g samples of one particular Chernobyl soil (Boury- kovka, 1990) were compared with two 2 g samples of the Windscale soil (i.e., the active sample and the control, as described earlier).Sample preparation, autoradiography and chemical extractions were carried out in Norway, at the Isotope Laboratory. Extracts and residues were then divided, and radiochemical measurements were carried out in dupli- cate, both at the Isotope Laboratory and at the University of Manchester (actinides were also determined at Manchester, and will be reported separately). Measurement differences between the two laboratories for parallel samples did not exceed 10%. As the soils were known to contain small radioactive particles, the loss of even a very small fraction of the sample could potentially affect the results disproportion- ately. Therefore, at all stages in the experiment the 137Cs content of the samples, sub-samples, residues and empty containers was monitored.No losses were detected. For autoradiography, the soils (the initial 2 g portions and the corresponding dried residues after the solution extrac- tions) were examined for the presence of Ply-emitting hot particles using standard X-ray film (Kodak X-OMAT AR-5). The samples were thinly spread on paper, covered with a thin polyethylene membrane, and the X-ray film was placed in close contact on top for 2 weeks exposure. The results are shown in Fig. 1. After the first autoradiographic exposures, the soil samples were treated sequentially with two portions of 6 mol 1-l HCl (20 mi each for 12 h with gentle shaking at 20 "C) and one of 1 mol 1-l ammonium acetate (20 ml for 2 h, intended as a precautionary measure to ensure that none of the extracted 137Cs would be re-adsorbed by soil components). Strontium chloride (20 mg of Sr) was added as a yield monitorlcarrier with each treatment, and on each occasion the residue was washed with water (10 ml), and the washings were added to the appropriate extractant solution.The 137Cs was determined in the initial starting materials, the extracting solutions at each stage, and in the residues. To determine total WSr, the extracted residues were totally dissolved in a 1 + 1 mixture of 60% HF and 16 moll-' HN03, the solutions were evaporated to dryness, and the residual solids were dissolved in 4 moll-' HN03. Comparative Analysis of Chernobyl Soils Strontium-90 was determined in 11 samples of Chernobyl soils (five from Bourkovka and six from Novo Shepilichi, collected in May, 1991) using a standard HCl extraction technique, and the residual 9OSr was then recovered by oxidative alkaline fusion.19,20 Duplicate 1 g portions of each soil were treated with cold 6 mol 1-1 HCI for 24 h, with occasional stirring.The residual solids from the HCl extractions were fused with a mixture of NaOH (41 g), NaKC03 (14 g) and KN03 (0.7 8). The fused mixture was extracted with HN03, and 90Sr was determined in both sets of extracts. Fig. 1 Autoradiographs of Windscale and Chernobyl soils: (a) Windscale soil before chemical extraction; ( b ) Windscale soil after extraction; (c) Chernobyl soil before chemical extraction; and ( d ) Chernobyl soil after extraction.The light patch on the top edge of (a), (c) and ( d ) is due to exposure from exterior lightANALYST, SEPTEMBER 1993, VOL. 118 1103 Radiochemical Measurements At the Isotope Laboratory, WSr was determined by an established laboratory procedure involving extraction of equilibrated 90Y .21 Solutions were evaporated to dryness, the residues were ashed at 400 "C, and the ashes were dissolved in 1 moll-' HCl, with addition of natural yttrium and strontium carriers. Yttrium was isolated by extraction into toluene containing 5% of bis(2-ethylhexyl) hydrogen phosphate (HDEHP), back-extracted into 3 mol 1-1 HN03 and finally determined by Cerenkov counting (Quantilus 1220 low-level liquid scintillation spectrometer; LKB-Wallac, Turku, Fin- land).Chemical yields were determined by complexometric titration. The 137Cs in solids was determined by means of an HPGe (high-purity germanium) detector, and in the extrac- tant solutions using a sodium iodide well (MiniAxi) detector (Canberra Packard Benelux, Tilburg, The Netherlands), cross-calibrated against the HPGe system. At the University of Manchester, strontium (i.e., carrier plus 90Sr) was isolated from 4 mol 1-1 HN03 by chromato- graphy on a strontium-specific solid phase (Sr-SPEC; EI- Chrom Industries, Chicago, IL, USA) and subsequently eluted with water.22 The chemical yield was determined by analysis for total strontium (inductively coupled plasma atomic emission spectrometry; Perkin- Elmer Model 6500, Norwalk, CT, USA), and 90Sr (with equilibrated 9OY) was determined after 3 weeks by liquid scintillation spectrometry [Canberra (Pangbourne, Dorset, UK) Model 2250CAl , using an equilibrated (90Sr-9OY) low-level standard for calibration (National Physical Laboratory, Ref.R715). Caesium-137 was determined by y-ray spectrometry. At the RisG National Laboratory, 90Sr was determined in all extracts using the classical technique of precipitation with fuming HN03, and P-counting (after 3 weeks).19,20 Yields were monitored with 85Sr. Results and Discussion Au toradiography The autoradiographs for the high-activity Windscale soil, before and after HCl extraction, are shown in Fig. l ( a ) and ( 6 ) , respectively. The low activity (control) Windscale soil produced no visible effect on the X-ray film. The analogous autoradiographs for one of the Chernobyl soil samples are shown in Fig.l(c) and (d) (both Chernobyl samples yielded similar results). At the pre-extraction stage, the presence of a considerable number of particles is clearly seen in the Chernobyl soil, while one particle is visible in the active Windscale soil. The resistance of these particles to HC1 extraction is demonstrated by the post-extraction autoradiographs [Fig. l(b) and (d), Windscale and Chernobyl, respectively], although the Windscale particle appears to have broken into two pieces during extraction. Using the particle positions identified by the autoradiograph [Fig. l(b)], the bulk of the residue was physically separated from the two particles. It was found that most of the 137Cs activity was still associated with the particles, not with the bulk soil.Hence, the possibility that 137Cs had been extracted from the particle and resorbed by the soil fraction can be excluded. Sequential HCl Extraction Results for the Windscale soil are summarized in Table 1. The total 137Cs activity of the soil plus particle was 22.8 k 0.5 Bq, whereas that of the control soil was 0.24 f 0.02 Bq (both 2 g). It seems reasonable to infer that, in the more active sample, at least 99% of the 137Cs (and presumably 90Sr) activity was associated with the one particle observed by autoradiography . For this particle, totals of 23 and 25% of the 137Cs and 90Sr, respectively (i. e., within experimental error, the same relative amounts), were extracted. Additionally, the 9OSr : 137Cs ratio (mean 1.7; range 1.5-2.0) does not vary markedly as between the various extracts or between extract and residue.These results suggest strongly that the extraction process results essentially from dissolution of the particle matrix itself. It might also be noted that the 90Sr : 137Cs ratio observed is higher than that assumed (0.9 : 1 .O) for the irradiated fuel that was released.2 The reason for the discrepancy is not clear, but could reflect preferential removal of Cs nuclides (e.g., by volatilization) at some stage in the life of the uranium fuel, perhaps during its oxidation and the formation of particles, which occurred in a hot-air stream over a long period. Possible, but in our view less likely, would be preferential extraction of caesium relative to strontium from the fuel particle after deposition, by weathering processes in the soil.For the analogous sequential extraction of the Chernobyl soil (Table 2), the relative amounts of 137Cs and 90Sr extracted by HC1 were 66 and 2670, respectively. The result for 9OSr, taken together with the demonstration by autoradiography of the presence of extraction-resistant particles, clearly demon- strates the ineffectiveness of HCl extraction for the determina- tion of total 90Sr in this soil. However, the greater extractabil- ity of 137Cs relative to 90Sr, which is also reflected by the Table 1 90Sr and 137Cs activities in sequential extracts and residue from a soil sample collected from near Windscale in February, 1956 Amount Amount WSr : 137Cs Component 137Cs/Bq* (%) WSr/Bq* (%) ratio 1.6rn011-~HC1(12h) 3 . 3 f 0 . 1 14.5 5.1k0.1 14.5 1.5 3. 1 moll-' NH40Ac (1 h) 0.3 f 0.1 1.3 0.5+0.1 1.4 1.7 Total extracted 5 . 2 f 0 . 3 22.8 8.8k0.3 25.0 1.7 Residue (measured) 17.6 f 0.2 77.2 26.4 k 2 75.0 1.5 Total in 2 g sample 22.8 100.0 35.2 100.0 1.54 2. 6m0ll-~HCl(12h) 1.6 f 0.1 7.0 3.2k0.1 9.1 2.0 * Activity f standard deviation (counting statistics). Table 2 WSr and 137Cs activities in sequential extracts and residue from a soil sample collected from near Chernobyl in May, 1990 Amount Component 137Cs/Bq* (%) WSr/Bq* 1. 6moll-' HCl(12h) 84.7 f 0.1 57.8 6.6 Ifr 0.2 2. 6 mol 1-' HCI (12 h) 10.9 Ifr 0.1 7.4 0.7k0.1 3. 1 rnol 1-1 NH40Ac (1 h) 1.0 f 0.3 0.7 0 . 3 f 0 . 1 Total extracted 96.6 f 0.5 65.9 7.6 f 0.4 Residue (measured) 50.0 f 1 34.1 21.4 k 1.5 Total in 2 g sample 146.6 100.0 29.0 Amount 22.7 2.4 1.0 26.2 73.8 100.0 (%> 90Sr : 137Cs ratio 0.078 0.064 0.30 0.079 0.43 0.198 * Activity 1- standard deviation; mean of two 2 g samples.1104 ANALYST, SEPTEMBER 1993, VOL.118 differing 90Sr : 137Cs ratios in the residue and the HCl-extract- able component (0.43 and 0.079, respectively), also suggests that a significant amount of the total 137Cs activity is bound to a matrix other than fuel particles. In relation to the analytical determination of 90Sr in particle- containing soils, the hypothesis outlined in the introduction to this paper has been substantiated by these experiments. In both instances, the HCl extraction method recovered about 20-26% of the total 90Sr present in the soil samples.For the Windscale soil, the incomplete extraction was shown to be due to the presence of a single fuel particle, resistant to extraction by the technique used, and a similar explanation is likely for the low 90Sr recoveries from the Chernobyl soil. Comparative Analysis of Chernobyl Soils In general, it might be expected that the fraction of the total 90Sr extracted from a soil sample would be affected by the fraction of the isotope associated with radioactive particles, the properties of the particles themselves (composition, size/ surface area, degree of post-depositional weathering, etc. ) and by factors related to sample preparation (e.g., drying tempera- ture or amount of sample grinding) and chemical extraction (e.g., temperature, agitation or length of extraction pro- cesses).In the sequential extraction experiments reported above, we deliberately reduced the mechanical factors thought likely to affect particle dissolution (e.g., we avoided fine grinding of the soil before extraction). Hence, while demonstrating that large, intact particles are resistant to chemical attack by HCl, our sequential extraction experi- ments may not be fully representative of routine laboratory soil analysis. In order to evaluate whether HCl extraction underestimates 90Sr in normal practice, we have also carried out determina- tions on 11 particle-containing soil samples from the Cher- nobyl area by the procedure (based on HCl extraction) routinely used at the Ris@ National Laboratory, and which follows a long established and widely used protocol19320 (although we recognize that soil pre-treatment methods for 9% determination vary considerably between laboratories , as indeed do the extraction methods themselves).The results for these soil analyses are presented in Table 3. The fractional W r recovery by HCl extraction (relative to that obtained by alkaline fusion and complete dissolution) ranged from 33 to 85%, with a mean of 54%. It is clear that, for soils containing uranium oxide fuel particles, extraction by cold 6 mol I-' HCI is not a satisfactory technique for the determination of total 90Sr. A significant deposit of condensation-type particles, in addition to fuel particles, in this soil is suggested by the relatively low values for the 90Sr : 137Cs ratios (range 0.27- 0.62), which are below the average ratio (0.71) for the irradiated fuel at the time of the accident.Extraction by HCl and the Alternatives Under conditions in which WSr has been deposited in sub- micrometre particles, and probably in an acid-soluble matrix, as is the case with weapons fallout, HCl extraction has been demonstrably and understandably effective.8.12 However, extraction of trace nuclides from larger particles of a refractory oxide presents a very different problem, as was identified some years ago by Sill and co-workers,9JO in relation to actinide determination. At that time, plutonium in soil from weapons fallout was determined accurately and reliably by an extraction procedure involving HCl or a mixture of HC1 and HN03. Sill9 demonstrated that, for soils contami- nated by 'an accidental release of plutonium' (site not stated), the standard extraction failed to recover all the plutonium, and additionally, if the soil had been strongly heated before extraction, recoveries as low as 530% were obtained.He concluded that treatment with these acids alone 'is grossly inadequate for dissolution of refractory compounds of pluto- nium'. Sill et aZ.10 concluded that only complete dissolution of the sample initially could ensure reliable analysis, and they evolved a method of sample pre-treatment based on a combination of KF and pyrosulfate fusions to meet this problem. Oxides of the lower oxidation states of uranium are only easily soluble in acids under oxidizing conditions.14 In our view, the HCl extraction method is ineffective for soils containing fuel particles, in part because of the lack of oxidizing capability (hence, extraction is enhanced by the use of H202 and/or HN0315).However, in instances in which fuel particles could be involved, even more rigorous oxidizing conditions might be needed. Additionally, thorough grinding of a sample to reduce the particle size (and increase the relative surface area) of the chemically resistant components would clearly assist recovery in instances where an acid- extraction method is to be used (in this context, we are aware of recent work23 in which a Windscale soil, similar to our own, afforded good recovery of 9OSr when extracted with 6 mol 1-1 HCl; however, it is probable that the soil sample was ground to a much finer powder than in our extraction experiments, which could explain the difference in the results obtained).It has been suggested .to us24 that the HCl extraction technique for 90Sr determined in soil is no longer in common usage, and that soil extraction with hot aqua regia (HCl- HN03, 3 + l), which is more likely to achieve complete recovery, is now more commonplace. Traditionally, recom- mended alternatives to acid extraction, for intractable mat- rices, have been alkaline and/or fluoride fusion under oxidizing conditions (as outlined earlier in this paper) .6,7 Table 3 WSr determination (Bq g-') by HCl extraction (and subsequent oxidative alkaline fusion of the residue) for samples of 11 soils containing radioactive particles, collected from near Chernobyl in May, 1991 Yh/Bq g-I in soil Amount in HCl extract Isotope ratio Sample Location* HCl extract Residue ("/.I (WSr : 137Cs) B2-11 B2-11 C2-7 C2-8 C2-9 E2-9 E2-10 E2-11 M-9 M-10 F2-11 BK 1 BK BK BK BK NS NS NS NS NS NS 11.3 f 0.2 6.2 k 0.1 27 f 0.4 23 k 0.3 42 k 0.6 58 k 0.8 49 f 0.7 59 f 0.7 134 f 2.0 132 f 2.0 155 k 2.0 9.9 * 0.1 5.3 rfr 0.1 12.3 k 0.2 18.7 t 0.2 7.4 2 0.1 42 rfr 0.6 66 k 0.8 59 k 0.8 122 k 2.0 270 f 4.0 230 k 3.0 53 54 69 55 85 58 43 50 52 33 40 0.62 0.39 0.30 0.29 0.37 0.43 0.42 0.39 0.27 0.43 0.41 * BK = Bourykovka; NS = Novo Shepilichi (15 and 5 km, respectively, from the Chernobyl reactor site).ANALYST, SEPTEMBER 1993, VOL.118 1105 These methods are all discussed in the recent RADREM (Radioactivity Research and Environment Monitoring Com- mittee) report,g and it does indeed seem likely that many UK laboratories would now not use HC1 extraction routinely for soil analysis.However, we also note that the use of HCl extraction for the determination of WSr in soils, including those which could contain fuel particles, is specifically recommended in the recent IAEA (International Atomic Energy Agency) ‘Guide- book’ for environmental radionuclide determination,13 inten- ded particularly to establish ‘reliable analytical methods to obtain data capable of inter-comparison in the case of radioactive releases’. In the recommended procedure, soil is first dried, weighed, and crushed to pass through a 4 mm mesh sieve. The dry soil (500 g recommended) is then treated with 6 mol 1-1 HCl and allowed to stand, with occasional stirring, for at least 8 h.After filtration, the procedure is repeated and the extracts are combined. This procedure is almost identical with that originally described in 1957 by Bryant et aZ.11 and used (scaled down) by ourselves in the present work. It now seems clear, both from the earlier work and from our own experiments, that HCl extraction is an unreliable method for the extraction of nuclides from soils, at least in instances where fuel particles are likely to be involved. Therefore, we suggest that the IAEA recommendation13 should be viewed with considerable caution. In our view, in instances where oxide fuel particles are likely to be present, only methods based on complete dissolution of the sample should initially be assumed to be reliable, and although methods based on mixed-acid extraction (HC1-HNO3-HF) could be effective, they should be evaluated for the particular circumstances of the analysis. Re-evaluation of Previous Results In view of our conclusions above, it may be necessary to examine the reported experimental methods carefully when evaluating determinations of soil 90Sr reported in connection with the Chernobyl accident, certainly for locations close to the site.Similarly, the determinations of 90Sr in soil near Sellafield in the period following the Windscale pile fire (October, 1957) are likely to have been affected by the presence of uranium oxide particles, as it was established at the time that most of the soil 90Sr identified within 10 km of Sellafield originated from particle emissions that had taken place before the fire.25 It seems to us possible that in both instances the total release of 90Sr will have been under- estimated as a result of the analytical problems we have discussed.We thank Dr. F. Leslie, formerly resident in Seascale, for supplying materials used in this investigation, H. N. Lien of the Isotope Laboratory, Agricultural University of Norway, for carrying out analyses for 90Sr, and Drs. F. R. Livens (University of Manchester) and B. T. Wilkins (UK National Radiological Protection Board), and also the two (anony- mous) referees of this paper, for helpful comments and suggestions. 2 3 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Chamberlain, A. C., Sci. Total. Environ., 1987, 63, 139. Loshchilov, N. A., Kashparov, V.A., Yudin, Ye. B., Protsak, V. P., Zhurba, M. A., and Parshakov, A. E., in The Radiobiological Impact of Hot Beta- Particles from the Cher- nobyl Fallout: Risk Assessment, IAEA, Vienna, 1992, Part I, Bogatov, S. A., and Borovoy, A. A., in The Radiobiological Impact of Hot Beta-Particles from the Chernobyl Fallout: Risk Assessment, IAEA, Vienna, 1992, Part 11, pp. 1-16. Salbu, B., in Proceedings of an International Workshop (Thueren, FRG), eds. von Philipsborn, H., and Steinhausler, F., Bergbau- und Industriemuseums, Thueren, 1988, vol. 16, pp. 83-84. Sandells, F. J., Segal, M. G., and Victorova, N., J. Environ. Radioact., 1993, 18, 5. Salbu, B., Krekling, T., Oughton, D. H., Ostby, G., Kash- parov, V. A., and Day, J. P., Proceedings of the International Symposium on Radioecology. Chemical Speciation-Hot Par- ticles, CEC/IUR Joint Workshop on Hot Particles, Prague, TUR-European Branch, Znojmo, Czechoslovakia, 1992, pp.108-110. Sampling and Measurements of Radionuclides in the Environ- ment, UK Department of the Environment, RADREM Report, HM Stationery Office, London, 1989. Sill, C. W., Health Phys., 1975, 29, 619. Sill, C. W., Hindman, F. D., and Anderson, J. I., Anal. Chem., 1979, 51, 1307. Bryant, F. J., Chamberlain, A. C., Morgan, A., and Spicer, G. S., Radiostrontium Fallout in Environmental Materials in Britain, Report AERE HP/R.2056 (unclassified), AERE, Harwell, UK, 1957. Wilken, R. D., and Diehl, R., Radiochim. Acta, 1987,41, 157. Measurement of Radionuclides in Food and the Environment: a Guidebook, Technical Report series, No. 295, IAEA, Vienna, 1989. Katz, J. J., and Rabinovitch, E., The Chemistry of Uranium, Dover Publications, New York, 1951, p. 322. Oughton, D. H., Salbu, B., Riise, G., Lien, H., 0stby, G., and Noren, A., Analyst, 1992, 117,481. Arnold, L., The Windscale Fire, 2957, Anatomy of a Nuclear Accident, Macmillan Press, London, 1992. Howells, H., Ross, A. E., and Gausden, R., The Release of Oxide from Irradiated Uranium in the Windscale Area Since October 1955, UKAEA Report No. IGO/TM/W036, 1957 (available from the Public Records Office, UK). Leslie, F. R., personal communication, 1990. Harley, J. H., Health and Safery Laboratory Procedures Manual, HASL-300, US Energy Research and Development Administration, New York, 1972. Aarkrog, A., Environmental Studies on Radioecological Sensi- tivity and Variability With Special Emphasis on the Fallout Nuclides WSr and 137Cs, Risd-R-437, Rise National Laboratory, Roskilde, Denmark, 1979. Bjornstad, H. E., Lien, H. N., Yu-Fu, Y., and Salbu, B., J. Radioanal. Nucl. Chem., 1992, 156, 165. Horwitz, E. P., Diety, M. L., and Fisher, D. E., Anal. Chem., 1991, 63, 522. McMahon, A. W., Toole, J., Jones, S. R., and Gray, J., unpublished work. Wilkins, B. T., personal communication, 1992. Bryant, F. J., Spicer, G. S., Chamberlain, A. C., Morgan, A., and Templeton, W. L., Radiostrontium in Soil, Grass, Vege- tables and Milk from Seven Farms in the Windscale Area. Report No. HP/R-2636, AERE, Harwell, UK, 1958. pp. 34-39. References 1 Eisenbud, M., Environmental Radioactivity Academic Press, New York, 3rd edn., 1987, pp. 244-391. Paper 210.51396 Received September 25, I992 Accepted February 11, 1993