Characterization of uranium and plutonium containing particles originating from the nuclear weapons accident in Thule, Greenland, 1968

Transcription

1 Journal of Environmental Radioactivity 81 (2005) Characterization of uranium and plutonium containing particles originating from the nuclear weapons accident in Thule, Greenland, 1968 O.C. Lind a, *, B. Salbu a, K. Janssens b, K. Proost b, H. Dahlgaard c a Isotope Laboratory, Department of Plant and Environmental Sciences, Norwegian 10 University of Life Sciences, P.O. Box 5003, N-1432 A s, Norway b Department of Chemistry, University of Antwerp, Universiteitsplein 1, Antwerp, Belgium c Riso National Laboratories, P.O. Box 49, DK-4000 Roskilde, Denmark Received 12 July 2004; received in revised form 12 October 2004; accepted 19 October 2004 Abstract To improve long-term radioecological impact assessment for the contaminated ecosystem of Bylot Sound, Greenland, U and Pu containing particles have been characterized with respect to particle size, elemental distribution, morphology and oxidation states. Based on scanning electron microscopy with XRMA, particles ranging from about 20 to 40 mm were isolated. XRMA and m-xrf mapping demonstrated that U and Pu were homogeneously distributed throughout the particles, indicating that U and Pu have been fused. Furthermore, m-xanes showed that U and Pu in the particles were present as mixed oxides. U was found to be in oxidation state IV whereas Pu apparently is a mixture of Pu(III) and Pu(IV). As previous assessments are based on PuO 2 only, revisions should be made, taking Pu(III) into account. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: Synchrotron radiation; Microscopic techniques; Oxidation states; XANES * Corresponding author. Tel.: C ; fax: C address: ole-christian.lind@umb.no (O.C. Lind) X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi: /j.jenvrad

2 22 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Introduction In the afternoon of the 21st of January 1968, an American B-52 aircraft carrying four thermonuclear weapons crashed on the sea ice in Bylot Sound (76 30#N) located on the NW Coast of Greenland, 11 km west of the Thule Air Base (Fig. 1). Due to explosive fire, fissile material, i.e. uranium and plutonium, from four weapons was dispersed over ice in a distance of a few kilometres. Most of the debris, including contaminated ice and snow, were removed in the clean-up programme conducted in the months following the accident. However, some of the fissile material was not recovered and has been located within the sediments of Bylot Sound (Aarkrog, 1971; Langham, 1970). Although several reports have estimated that about 90% of the plutonium was subsequently removed in the ensuing clean-up operation (Aarkrog et al., 1984; Langham, 1970), the residual contamination has recently been re-estimated to about 10 TBq 239 Pu, i.e. w3.8 kg (Eriksson, 2002). The Pu contamination is primarily found within a 6 km radius from the point of impact and Pu is to a certain extent present as particles, having a wide size distribution and a median diameter of 2 mm (Holm et al., 1987; Langham, 1970). During the years, it has been assumed that all radioactive particles contained Pu only and that Pu in the particles was present as insoluble plutonium oxides (PuO 2 ) due to the oxidation of the nuclear material (Aarkrog et al., 1987; Holm et al., 1987; Langham, 1970). However, the presence of U in the Thule particles has recently been reported (Eriksson, 2002). Observed activity concentrations of Pu and concentration ratios for Pu isotopes at Thule support the assumption that the residual fissile material exist as particles of relatively low solubility (Dahlgaard et al., 2001). A clear signal from the accident in the water column has only been identified (30 mbq m ÿ3, 3 6 Fig. 1. Bylot sound. Station V indicates both the point of impact and sampling location (McMahon et al., 2000).

3 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) times higher than the background level) in near-bottom water samples collected at the point of impact (Dahlgaard et al., 2001). The elevated concentration of Pu has been attributed to resuspension of contaminated sediments. According to Peurrung et al. (2001), weapons grade plutonium consists primarily of metallic 239 Pu. However, fissile material may also be prepared as mixed oxides ((U,Pu)O x ) employing sintering or melting processes (Keller, 1971). To the knowledge of the authors, it has not been established in the open literature that the nuclear weapons involved in the Thule accident nor in the similar Palomares accident, were made from metallic plutonium. Thus, the original form of the dispersed Pu is still somewhat unclear. Based on 240 Pu/ 239 Pu isotope ratio measurements Dahlgaard et al. (2001) found that the Thule debris has at least two different Pu sources. About two-thirds of the debris originated from a source having a 240 Pu/ 239 Pu atom ratio of 0.056, whereas about one-third of the debris featured a lower 240 Pu/ 239 Pu atom ratio of Dahlgaard et al. (2001) were not able to determine whether the sources originated from different weapons or from different parts of an individual weapon. In order to reduce the uncertainties of environmental impact assessments of particle contaminated ecosystems detailed information is required on particle characteristics such as particle size, morphological structure and oxidation states of matrix elements such as U and Pu influencing particle weathering rates and subsequent mobilization of radionuclides from particles present in sediment water systems (Salbu et al., 2004). Previous studies of radioactive particles released from different nuclear sources and under different release conditions such as reactor accidents involving explosions or fires have demonstrated that the particle characteristics are source and release scenario dependent (Salbu, 2000). Inert U particles with low weathering rates were released from the Chernobyl UO 2 fuelled reactor during explosion (at high temperature and pressure conditions without air), while oxidized U 3 O 8 particles with at least a factor of 10 higher weathering rates were released during the subsequent fire (Kashparov et al., 1999; Salbu et al., 2001). The solubility of released Pu particles may also vary, as demonstrated by Cooper et al. (1994), reporting solubilities of Pu ranging from 1 to 96% in 0.16 M HCl over 40 days, from Pu particles produced by U.K. safety trials at Maralinga, Australia. The solubility of Pu from the Thule particles is reported to be low as less than 3% of the weapon derived Pu could be attributed to the mobile fractions and the major fraction was associated with residues following leaching with concentrated HNO 3, HClO 4 and HF (Holm et al., 1987). In the present work, radioactive particles have been isolated from contaminated sediment samples collected at Bylot Sound. The isolated single particles have been characterized using scanning electron microscopy with X-ray microanalysis (SEM XRMA) and synchrotron radiation (SR) microscopic techniques (Salbu et al., 2003). The SR based microscopic techniques were especially developed for characterization of particles containing uranium (Salbu et al., 2001). SR based micro-x-ray fluorescence (m-srxrf) gives information on the elemental matrix composition of the particles, while SR based micro-x-ray absorption near edge spectroscopy (m-xanes) provides information on oxidation states of U and Pu.

4 24 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Materials and methods 2.1. Sampling and pre-treatment Sampling was performed onboard the Greenland fisheries research vessel Adolf Jensen during the August 1997 expedition in Bylot Sound (Dahlgaard et al., 2001). Thule sediments were sampled just below the point of impact using a Gemini corer. Aliquots of the w5 kg sediment sample were air dried prior to isolation of individual particles. Identification and isolation of radioactive particles in the environment is a tedious time consuming task. A set of three particles was isolated by numerous sample splittings, digital autoradiography using phosphor imaging (Molecular Dynamics) and subsequential gamma measurements (Low Energy Germanium detector, Canberra) of 241 Am activity concentrations in the sub-samples (Lind et al., 2002). Elevated activity concentrations of 241 Am found in aliquots indicated the presence of Pu. Single sediment grains or individual radioactive particles isolated from the matrix were mounted on carbon tapes and subjected to scanning electron microscopy (SEM; JEOL JSM 840) with X-ray microanalysis (XRMA) (Salbu et al., 2003). In Secondary Electron Imaging (SEI) mode information on the surface structures of particles was obtained. Using Backscattered Electron Imaging (BEI) mode, bright areas reflected the distribution of high atomic number elements. Using X-ray mapping, the distribution of individual elements such as U and Pu within specific particles was attained, while XRMA provided semi-quantitative information on individual elements at the electron beam spot, such as the concentration of U and Pu at specific particle sites. By superimposing X-ray-mapping with SEI or BEI images, individual particles containing U and Pu were identified prior to m-xanes analysis SR based micro-xanes Particles characterized by SEM with XRMA were subjected to m-xanes analysis using the X-ray microscopic facility at beamline L, HASYLAB, Hamburg (Salbu et al., in press). m-xrf was utilized to determine the elemental composition within individual particles using the m-xanes set-up (HPGe detector). Using m-xanes, information on the oxidation state of U and Pu particles was obtained by tuning the monochromatic, focused (20! 20 mm beam by using a polycapillary lens) X-ray beam over the U L III ( kev) and Pu L III ( kev) absorption edges, while keeping the beam position on individual particles. The flux at the beam spot on the sample was about 10 9 photons/s at 17.1 kev. The incident and transmitted beam intensities (I 0, I ) were measured by ionisation chambers, while the U L III and Pu L III fluorescence intensity was recorded by means of a Si(Li) well collimated (2 mm diameter pinhole) detector having an area of 30 mm 2 mounted at 45 to the incident beam and 30 mm from the sample. The m-xanes spectra were collected at 1 ev increments over a 300 ev energy range (extending from about 80 ev below and

5 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) ev above for U L III and 60 ev below and 130 ev above for Pu L III ). Based on well defined U oxidation state standards (UO 2 and U 3 O 8, Institute of Energy Technology, Kjeller; UO 2 Ac 2! 2H 2 O p.a., Riedel-De Hae n AG, Seelze-Hannover; UO 2 (NO 3 ) 2! 6H 2 O p.a., Merck, Darmstadt) the m-xanes spectra were recorded and the inflection point energies for the oxidation states of U established. The m-xanes profile shapes and energy shifts (inflection point energies) of U in particles were compared with those of the standards. For Pu no suitable reference compounds or solutions were available during the experiments. Thus, the recorded m-xanes Pu profiles were calibrated against Y and Zr reference foils and positioned on the same energy scale as profiles previously obtained (Denecke, pers. comm.). Thus, the oxidation state of Pu was determined by the position of the white line and a qualitative comparison to published Pu(III) Pu(VI) XANES profiles. XANES profiles were collected from both the cores and the outer spheres of the particles to ensure that inhomogeneities would be taken into account g-spectrometry and ICP-MS measurements Isolated single particles were counted on top of an LEGe detector (Canberra) using a purpose-built geometry set-up with particles mounted 6 mm from the detector surface. After having characterized the samples by non-destructive techniques one particle was sacrificed in order to identify the source of the particle. The particle was partly dissolved (28%) by microwave boiling (72 Bar, w280 C) in suprapur HNO 3 (65%). The final solution was filtered by Whatman GF/C filters. The 240 Pu/ 239 Pu atom ratio was determined by ICP-MS (Perkin Elmer ELAN 6100) after radiochemical separations based on selective sorption on anion exchange resins (Dowex AG 1! 8) to separate U and Am from Pu (Clacher, 1995). 3. Results and discussion 3.1. g-spectrometry and ICP-MS Based on g-measurements of individual particles (n Z 3) the 241 Am activities ranged from 1.0 to 2.7 Bq/particle. Based on ICP-MS measurements the 240 Pu/ 239 Pu atom ratio of particle #132a was G and the activity concentration of 239 C 240 Pu was 20 G 3.6 Bq. Thus, the plutonium in the analysed particle belongs to the high 240 Pu/ 239 Pu atom ratio group described by Dahlgaard et al. (2001). By combining the g-measurement data with the ICP-MS data, a 241 Am/ 239 C 240 Pu activity ratio of 0.14 G 0.02 could be calculated for particle #132a. The corresponding ratios reported for individual particles (n Z 5) at Thule is G (Eriksson, 2002), while Dahlgaard et al. (2001) found a mean ratio of 0.13 G 0.08 for contaminated sediments. Thus, the atom ratio and activity ratio

6 26 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) data obtained in the present work are in agreement with previously published results verifying that the particles originated from the Thule accident SEM with XRMA Based on digital autoradiography and scanning electron microscopy (SEI-mode) a variety of particles were observed. Using BEI mode, however, distinct individual particles containing high atomic number elements could be distinguished (Figs. 2 and 3). By introducing X-ray mapping, the surface distribution of U and Pu coincided with the intense BEI particle, reflecting that U and Pu were the dominant high atomic number elements. Based on SEM, the particle size distribution of isolated particles ranged between about 20 mm and about 40 mm. All of the particles investigated contained an apparently homogeneous mixture of uranium and plutonium on the particle surfaces. In addition, electron-induced X-ray spectra of the particles showed relatively strong Si, Fe and Al signals (Figs. 2b and 3b), probably reflecting that the particles were partly coated with sediment material. Particle #132a had a fluffy amorphous appearance, while particles #133 and #134 (not shown) looked more like agglomerated crystalline grains with a relatively high degree of porosity. Fig. 2. Scanning electron microscopy of particle #132a in SEI-mode (a), elemental spot analysis by XRMA (b) and X-ray mapping of uranium (c) and plutonium (d) superimposed on a BEI mode image (Bar 20 mm).

7 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Fig. 3. Scanning electron microscopy of particle #133 in BEI mode (a), elemental spot analysis by XRMA (b), and X-ray mapping of uranium (c) and plutonium (d) superimposed on the BEI mode image (Bar 20 mm) SR based micro-xrf Synchrotron radiation based m-xrf mapping was performed to identify the subsurface elemental distributions. Based on the pixels obtained during mapping, the correlation between U and Pu was almost linear (Fig. 4). The calculated linear Pearson correlation coefficient for the plot was r Z , i.e. an extremely high degree of linear correlation. However, the relatively large beam size limited the resolution of the system. The obtained correlation indicates that U and Pu have been fused, either as a consequence of the explosion and subsequent fire or during the production of the fissile material SR based micro-xanes UL III and Pu L III fluorescent XANES profiles (uncorrected for self-absorption) of three particles are shown in Figs. 5 and 6, respectively. Apparently all U and Pu in the particles investigated were oxidized (no metal present) and the particles could be characterized as mixed U and Pu oxide particles. No differences in oxidation states

8 28 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Fig. 4. U intensities plotted against Pu intensities in individual pixels as determined by m-srxrf mapping of particle #134. The calculated linear Pearson correlation coefficient for the plot is r Z were observed between cores and the outer spheres of the investigated particles. However, the low resolution reduced the ability to make definitive conclusions about the spatial distribution of the oxidation states. The U L III -profiles strongly resembled that of UO 2, indicating that most of the U in the particles were present as U(IV). As the solid-state speciation of U in the original material is unclear, the mechanisms leading to UO 2 particles are difficult to outline. If the original material was made from UO 2 no further oxidation took place during the explosive fire. However, at high temperature and pressure conditions in the presence of air a mixture of UO 2 and U 3 O 8 should be expected, as observed for UO 2 particles released during the Chernobyl fire (Salbu et al., 2001). If the original material was made from metallic U, oxidation to UO 2 could easily occur during the explosive fire, as observed for impacted depleted uranium munitions (Salbu et al., 2003; Salbu et al., 2004). For Pu, the position of the white line and the qualitative comparison to Pu L III XANES profiles for Pu(III) and Pu(IV) suggested that all Pu in the particles investigated were present as a mixture of Pu(III) and Pu(IV). The broad white lines were considered to be indicative of the presence of a mixture between Pu(III) and Pu(IV). The energy difference between the centroid of the white line and that of the first maximum supports this interpretation, suggesting a mean Pu-O bond distance intermediate between those found for Pu(III) and Pu(IV). The broadening of the white line might also be caused by self-absorption and/or the presence of Pu(IV) in a highly defective lattice, but would not influence the above-mentioned difference in

9 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Fig. 5. Fluorescent U L III m-xanes profiles obtained from Thule particles in comparison to UO 2 and U 3 O 8 reference compounds. energy. It should be emphasized that the present m-xanes results are the first results published for the oxidation state of Pu in environmental particles. Thus, assumptions made by earlier workers that Pu in released particles was only oxidized to PuO 2 during the accident (explosion, fire) should be revised. This is of importance as Pu solubility and particle weathering rate predictions depend on the oxidation state of Pu. If the original material was made of PuO 2 no further oxidation should have taken place during the explosive fire as PuO 2 is not likely to transform into superstoichiometric oxides (Keller, 1971). At temperatures above 1400 C PuO 2 starts losing oxygen and forms substoichiometric oxides, e.g. Pu 2 O 3, which is more readily soluble than PuO 2 (Keller, 1971; Sill, 1975). If the original material was made from metallic Pu, partial oxidation to Pu(III) and Pu(IV) seems plausible. Thus, in either

10 30 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Fig. 6. Fluorescent Pu L III m-xanes profiles obtained from Thule particles in comparison to literature Pu(III) and Pu(IV) reference profiles. case the presence of Pu(III) cannot be excluded. However, previously reported data from Bylot Sound, including field observations and sequential extraction results indicate that the solubility of the particles is low and that Pu is not readily available for uptake in bottom feeding organisms (Holm et al., 1987). 4. Conclusions Three individual U and Pu containing particles originating from the nuclear weapons accident at Thule have been isolated from Bylot Sound sediments and characterized with respect to elemental and isotopic composition, morphological structure and oxidation states of U and Pu. The origin of the particles was verified by

11 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) the 240 Pu/ 239 Pu atom ratio (0.055 G 0.007) and 241 Am/ 239 C 240 Pu activity ratio (0.14 G 0.02). At the lateral resolution employed, the particles appeared to contain a homogeneous mixture of U and Pu oxides. Based on the m-xanes results, both U and Pu were oxidized in all investigated particles. U was found to be present as UO 2 while Pu is suspected to be a mixture of Pu(III) and Pu(IV). Thus, the assumption made by earlier workers that all Pu was oxidized to PuO 2 should be revised. The knowledge of the oxidation state of Pu is essential for the prediction of weathering rates and the subsequent mobilization of radionuclides associated with the particles. Acknowledgements We gratefully acknowledge the support provided by the European Commission (contract no. IHP-Contract HPRI-CT ), IAEA (CRP, Characterization of radioactive particles ) and the Norwegian Research Council (project no / 720). The authors are indebted to Lindis Skipperud, AUN for ICP-MS assistance, to Gerald Falkenberg, HASYLAB BL for beamline assistance and Melissa A. Denecke, Forschungszentrum Karlsruhe for providing Pu XANES profiles. References Aarkrog, A., Health Physics 20, Aarkrog, A., Boelskifte, S., Dahlgaard, H., Duniec, S., Holm, E., Smith, J.N., Studies of transuranics in an Arctic marine-environment. Journal of Radioanalytical and Nuclear Chemistry 115, Aarkrog, A., Dahlgaard, H., Nilsson, K., Holm, E., Further-studies of plutonium and americium at Thule, Greenland. Health Physics 46, Clacher, A., Development and application of analytical methods for environmental radioactivity. PhD thesis. University of Manchester. Cooper, M.B., Burns, P.A., Tracy, B.L., Wilks, M.J., Williams, G.A., Characterization of plutonium contamination at the former nuclear-weapons testing range at Maralinga in south-australia. Journal of Radioanalytical and Nuclear Chemistry 177, Dahlgaard, H., Eriksson, M., Ilus, E., Ryan, T., McMahon, C.A., Nielsen, S.P., Plutonium in the marine environment at Thule, NW-Greenland after a nuclear weapons accident. In: Kudo, A. (Ed.), Plutonium in the Environment. Elsevier, Oxford, pp Eriksson, M., On weapons plutonium in the Arctic environment (Thule, Greenland). PhD thesis. Risø National Laboratory, pp Holm, E., Gastaud, J., Oregioni, B., Chemical partitioning of plutonium and americium in sediments from the Thule region (Greenland). In: Radionuclides: A Tool for Oceanography. Elsevier Applied Science Publishers, U.K., pp Kashparov, V.A., Oughton, D.H., Protsak, V.P., Zvarisch, S.I., Levchuk, S.E., Kinetics of fuel particle weathering and 90 Sr mobility in the Chernobyl 30 km exclusion zone. Health Physics 76, Keller, C., The Chemistry of the Transuranium Elements, vol. 3. Verlag Chemie GmbH, Weinheim, pp Langham, W.H., Project Crested Ice. USAF Nuclear Safety, pp Lind, O.C., Salbu, B., Janssens, K., Simionovici, A., Radioactive particle characterisation by means of synchrotron radiation-based X-ray micro beam techniques. In: Strand, P. (Ed.), Conference on

12 32 O.C. Lind et al. / J. Environ. Radioactivity 81 (2005) Environmental Radioactivity in the Arctic & Antarctic. Norwegian Radiation Protection Authority, Østerås, pp McMahon, C.A., Vintro, L.L., Mitchell, P.I., Dahlgaard, H., Oxidation-state distribution of plutonium in surface and subsurface waters at Thule, northwest Greenland. Applied Radiation and Isotopes 52, Peurrung, A., Arthur, R., Elovich, R., Geelhood, B., Kouzes, R., Pratt, S., Scheele, R., Sell, R., The 871 kev gamma ray from O-17 and the identification of plutonium oxide. Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment 474, Salbu, B., Source-related characteristics of radioactive particles: a review. Radiation Protection Dosimetry 92, Salbu, B., Janssens, K., Lind, O.C., Proost, K., Danesi, P.R., Oxidation states of uranium in DU particles from Kosovo. Journal of Environmental Radioactivity 64, Salbu, B., Janssens, K., Lind, O.C., Proost, K. Gijsels, L., Danesi, P.R., Oxidation states of uranium in DU particles from Kuwait. Journal of Environmental Radioactivity 78, Salbu, B., Krekling, T., Lind, O.C., Oughton, D.H., Drakopoulos, M., Simionovici, A., Snigireva, I., Snigirev, A., Weitkamp, T., Adams, F., Janssens, K., Kashparov, V.A., High energy X-ray microscopy for characterisation of fuel particles. Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment 467, Salbu, B., Lind, O.C., Skipperud, L., Radionuclide speciation and the relevance in environmental impact assessments. Journal of Environmental Radioactivity 74, Sill, C.W., Some problems in measuring plutonium in the environment. Health Physics 29,