oxidation states of uranium in depleted uranium particles from kuwait

www.elsevier.com/locate/jenvrad

Journal of Environmental Radioactivity 78 (2005) 125–135

Oxidation states of uranium indepleted uranium particles from Kuwait

B. Salbua,), K. Janssensb, O.C. Linda, K. Proostb,L. Gijselsb, P.R. Danesic

aDepartment of Plant and Environmental Sciences, Isotope Laboratory,

Agricultural University of Norway, P.O. Box 5028, 1432 Aas, NorwaybDepartment of Chemistry, University of Antwerp, Universiteitsplein 1, Antwerp, Belgium

cThe International Atomic Energy Agency’s Seibersdorf Laboratories,

A-2444 Seibersdorf, Vienna, Austria

Received 21 November 2003; received in revised form 31 March 2004; accepted 7 April 2004

Abstract

The oxidation states of uranium in depleted uranium (DU) particles were determined by

synchrotron radiation based m-XANES, applied to individual particles isolated fromselected samples collected at different sites in Kuwait. Based on scanning electronmicroscopy with X-ray microanalysis prior to m-XANES, DU particles ranging from sub-

microns to several hundred micrometers were observed. The median particle size dependedon sources and sampling sites; small-sized particles (median 13 mm) were identified in swipestaken from the inside of DU penetrators holes in tanks and in sandy soil collected belowDU penetrators, while larger particles (median 44 mm) were associated with fire in a DU

ammunition storage facility. Furthermore, the 236U/235U ratios obtained from acceleratormass spectrometry demonstrated that uranium in the DU particles originated fromreprocessed fuel (about 10�2 in DU from the ammunition facility, about 10�3 for DU in

swipes).Compared to well-defined standards, all investigated DU particles were oxidized.

Uranium particles collected from swipes were characterized as UO2, U3O8 or a mixture of

these oxidized forms, similar to that observed in DU affected areas in Kosovo. Uraniumparticles formed during fire in the DU ammunition facility were, however, present asoxidation state C5 and C6, with XANES spectra similar to solid uranyl standards.Environmental or health impact assessments for areas affected by DU munitions should

therefore take into account the presence of respiratory UO2, U3O8 and even UO3 particles,

) Corresponding author. Tel.: C47-649-48351; fax: C47-649-48359.

E-mail address: [email protected] (B. Salbu).

0265-931X/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jenvrad.2004.04.001

http://www.elsevier.com/locate/jenvrad

mailto:[email protected]

126 B. Salbu et al. / J. Environ. Radioactivity 78 (2005) 125–135

their corresponding weathering rates and the subsequent mobilisation of U from oxidizedDU particles.

� 2004 Elsevier Ltd. All rights reserved.

Keywords: Depleted uranium particles; Oxidation states; XANES; SEM–XRMA; Kuwait

1. Introduction

During the Kuwait and Balkans conflicts large amounts of depleted uranium(DU) ammunition were expended. When hitting hard targets such as tanks,penetrators made from metallic DU alloyed with small amounts of titanium generatea cloud of DU dust (aerosol) that ignites and creates a fire. In addition, unspent DUpenetrators deposited in the field will corrode over time and fragments and particlesare formed. Such material is generally localised close to the source, but can betransported in the environment due to resuspension and agricultural practices.Furthermore, due to weathering of fragments and particles and the subsequentremobilisation of uranium ecosystem transfer of uranium from soil/sand to the foodchain may take place.

To evaluate potential health hazards of depleted uranium munitions severalreview articles have been published (e.g., Fetter and von Hippel, 1999; Priest, 2001;The Royal Society, 2001) based on previous or recent investigations that focusedmainly on the radiological aspect of uranium, being depleted in 235U. When health orenvironmental impacts are assessed, information is needed not only on theradioactivity of particles, but also on particle size distribution and particlecharacteristics such as morphological structure and oxidation state of U influencingparticle weathering rates and subsequent mobilization of U from DU particles, whenin contact with body fluids or in natural aquatic systems. Previous studies ofradioactive particles released from different nuclear sources and under differentrelease conditions, such as reactor accidents involving explosions or fires,demonstrate that the particle characteristics are source and release scenariodependent (Salbu et al., 1994, 1998, 2000; Salbu, 2000). Inert U particles with lowweathering rates were released from the Chernobyl UO2 fuelled reactor duringexplosion (at high temperature and pressure conditions), while oxidized U3O8

particles with high weathering rates were released during the subsequent fire(Kashparov et al., 1999; Salbu et al., 2001). As recommended by The Royal Society(2001), research should therefore include information on the particle size distributionand properties of the DU particles in the aerosol.

Following the 1999 Balkan conflict, IAEA joined a UN mission to Kosovo. Basedon samples collected, the particle size ranged from submicrons to about 30 mm withaverage size of 2 mm, i.e. the particles were included in the respiratory fraction(Danesi et al., 2003a, b; Salbu et al., 2003). The size distribution of DU particles,crystallographic structure as well as the oxidation state of U were determined onselected samples using scanning electron microscopy with X-ray microanalysis

127B. Salbu et al. / J. Environ. Radioactivity 78 (2005) 125–135

(SEM–XRMA) and synchrotron radiation microscopic techniques (Salbu et al.,2003). These techniques proved valuable for analysis of DU particles, showing thatmost of the small-sized DU particles were oxidized.

Following the Gulf War investigations have been performed, focusing on theradiological aspect of DU particles in man. Moreover in 2002, IAEA organiseda field mission to collect samples at selected sites in Kuwait, where DU ammunitionswere used (IAEA, 2003). Similar to the DU samples from Kosovo, the samples fromdifferent sites in Kuwait have been subjected to a variety of analytical techniques,including gamma and alpha spectrometry, inductively coupled plasma massspectrometry (ICP-MS) and accelerator mass spectrometry (AMS) to determinethe concentration and isotopic ratios of uranium isotopes.

In the present work, DU particles from few selected samples collected in Kuwaithave been characterized with respect to composition, particle size, morphology andoxidation state of U. Particles with low 235U/238U ratios are assumed to be DUsource-specific, while the particle size distribution, oxidation states of U andconcentration of Ti will depend on the release scenario.

2. Materials and methods

Samples of sand collected at four different sites in Kuwait (Table 1) were dried atthe IAEA Seibersdorf Laboratory and subjected to gamma spectrometry (CanberraHPGe detector 1.85 keV resolution, 30% efficiency) at Agricultural University ofNorway prior to identification and isolation of individual particles.

The sampling sites are described by IAEA (2003). Due to a fire in a DU munitionsstorage at Al Doha in 1991, about 660 rounds (charges) of DU munitions weredestroyed. Following clean up operations, about 300 penetrators, corresponding toabout 1500 kg DU, were unaccounted. Access to the affected area is still restricted.The Manageesh oil field, southwest of Kuwait City, was occupied by a large numberof Iraqi troops and subjected to repeated air raids involving DU munitions. Accessto the area is still restricted due to the presence of a large number of unexploded land

Table 1

Investigated samples

Site Site description Characteristics Amount Number of

particles

Particle labels

in text and

figures

Al Doha Fire in a DU

ammunition

storage

Large particles

of DU

w2 g 43 84–93

Manageesh Below a DU

penetrator

Small-sized DU w2 g 5 94, 95

Um Al Kwaty Swipe from inside

damaged tank

Corroded material

(steel) containing DU

Few mg 11 80–81

Um Al Kwaty Swipe from inside

damaged tank

Corroded material

(steel) containing DU

Few mg 20 82


mines and cluster bombes, as well as intact DU penetrators. The affected site at UmAl Kwaty is situated close to the Ali Salem air force base, housing thousands ofdestroyed military vehicles including tanks contaminated from DU munitions. Theswipes were taken from within penetrator holes in damaged tanks (IAEA, 2003).

Individual DU particles were characterized with respect to structure by scanningelectron microscopy (SEM), to elemental composition by X-ray microanalysis(XRMA) and to oxidation states of U by using synchrotron radiation (SR) basedmicro-X-ray absorption near edge spectroscopy (m-XANES) performed at theHASYLAB synchrotron facility, Hamburg in Germany. Preliminary structureanalysis using micro-diffraction (m-XRD) has also been performed at ESRF,Grenoble, France. The SR based microscopic techniques such as the m-XANEStechnique have previously been developed at ESRF, for particles containing uranium(Salbu et al., 2000).

2.1. Scanning electron microscopy with XRMA

Depleted uranium particles isolated from the matrix were mounted on carbontapes and subjected to scanning electron microscopy (SEM) using a JEOL JSM 840instrument interfaced with X-ray micro-analyser (ISIS 300, Oxford Instruments).Swipe sample particles were transferred to carbon tape by touching the swipe cloths.In secondary electron imaging (SEI) mode, information on the surface structures ofparticles was obtained. Using backscattered electron imaging (BEI) mode, brightareas reflected the distribution of high atomic number elements. Using X-raymapping, the distribution of individual elements such as U within specific particleswas attained, while XRMA provided semi-quantitative information on individualelements at the electron beam spot, such as the concentration of U and Ti at specificparticle sites. By superimposing X-ray mapping with SEI or BEI images, individualparticles containing U were identified prior to m-XANES analysis.

2.2. SR-based micro-XANES

Particles characterized by SEM with XRMA were subjected to m-XANES analysisusing the X-ray microscopic facility at beamline L, HASYLAB, Hamburg (Fig. 1a).m-XRF was utilized to determine the concentration of elements within individualparticles using the m-XANES set-up (HPGe detector). Using m-XANES, informationon the oxidation state of U in DU particles was obtained by tuning themonochromatic, focused (20!20 mm beam by using a polycapillary lens) X-raybeam over the U LIII absorption edge (17.163 keV), while keeping the beam positionon individual DU particles. The flux at the beam spot on the sample was about 109

photons per second at 17.1 keV. The incident and transmitted beam intensities (I0, I )were measured by ionisation chambers, while the U LIII fluorescence intensity wasrecorded by means of a HPGe detector having an area of 30 mm2 mounted at 45( tothe incident beam and 30 mm from the sample. The HPGe detector was wellcollimated (2 mm diameter pinhole). The m-XANES spectra were collected at 1 eVincrements over a 300 eV energy range (extending from about 80 eV below and


220 eV above the U LIII). Based on well-defined U oxidation state standards (UO2,Institute of Energy Technology, Kjeller; U3O8, Institute of Energy Technology,Kjeller; UO2Ac2!2 H2O p:a:, Riedel-De Haen AG, Seelze-Hannover; UO2ðNO3Þ2!6 H2O p:a:, Merck, Darmstadt) the m-XANES spectra were recorded and theinflection point energies for the oxidation states of U established. The m-XANESprofile shapes and energy shifts (inflection point energies) of U in DU particles werecompared with those of the standards.

2.3. SR-based micro-XRD

The m-XRD data were collected at the ESRF ID18F end station (Fig. 1b). Duringthe experiment, the monochromator was tuned to 28 keV and a compound refractivelens focused the monochromatic beam. The samples were mounted in slide framesand moved through the beam by means of a motorized XYZ sample stage.

Fig. 1. (a) Experimental design performing XANES at the X-ray microscopy facility, beamline L, Hasylab

synchrotron, Hamburg, Germany. (b) Experimental design performing XRD at the X-ray microscopy

facility, beam line 18F, ESRF, Grenoble, France.


Perpendicular to the direction of the primary beam, an energy-dispersive X-raydetector allowed the collection of fluorescent signals from the sample. Furthermore,a Photonics Science X-ray diffraction camera coupled to a CCD camera was used torecord the diffraction signals. Measurements included point measurements(5!10 mm, single XRF spectrum and m-XRD image collected during typically 10–100 s) and line scans (horizontal or vertical lateral movement of the particle throughthe primary beam with increment of typically 2–5 mm and the collection of a XRF-spectrum and/or m-XRD image at all positions along the line).

2.4. ICP-MS and AMS measurements of isotope ratios

Individual DU particles and certified reference ore materials (EC nuclearreference material 113 and 114, respectively) were dissolved in aqua regia. Uraniumconcentrations and 235U/238U isotope ratios were determined by ICP-MS (Perkin–Elmer ELAN 6000) after radiochemical separations based on selective sorption onanion exchange resins (Dowex AG 1!8) to separate U and Am from Pu andextraction chromatography (TRU-Spec) to separate U from Am (Clacher, 1995).The 235U/238U and 236U/235U isotope ratios were also determined by AMS using the14UD tandem accelerator at the Australian National University, Canberra, ac-cording to the procedure described by Marsden et al. (2001).

3. Results and discussion

Based on gamma spectrometry, the concentration of the uranium isotopes in thesand samples varied due to varying amount of DU particles in the samples. TheXRMA demonstrated also the presence of Ti as well as traces of other metals withinmost of the DU particles (Fig. 2a and b), while for other DU particles (Fig. 2c and d)the detection limit for Ti was reached. Based on ongoing ICP-MS and AMS work,the 235U/238U ratios in the DU particles were about constant, 0.002, while the236U/235U isotope ratios varied according to DU particles analysed; 10�2 for DUparticles associated with fire in the DU ammunition storage facility, and about 10�3

in DU particles from swipes taken from the inside of DU penetrators holes in tanks(Table 2). Thus, the DU munitions originated from reprocessed uranium fuel.

Based on scanning electron microscopy (SEI mode), a variety of particles wasidentified on the tapes. Using BEI mode, however, distinct individual particlescontaining high atomic number elements could be distinguished (Fig. 2a and b). Byintroducing X-ray mapping, the distribution of U coincided with the intense BEIparticle, reflecting U as the dominant high atomic number element. Based on SEM,sizes of isolated DU particles ranged from 2 to 64 mm (median 13 mm, n¼ 36) insamples taken from penetrator holes in tanks hit by DU ammunition or close topenetrators (Fig. 2a and b). DU particles with a wider size distribution ranging from0.2 to 1500 mm (median 44 mm, n¼ 43) were attributed to the fire in the DUammunition storage facility (Fig. 2c and d). The large DU particles appeared witha strong yellow colour typical for uranyl compounds and a crystalline structure,


quite different from all DU particles observed in Kosovo or at other sites in Kuwait,and different from U particles released from other nuclear event such as theChernobyl accident (Salbu et al., 2001).

Using hard energy synchrotron X-ray radiation tuned over the U LIII absorptionedge (17.163 keV), the m-XANES spectra of the defined standards (UO2, U3O8,UO2Ac2, UO2(NO3)2) were attained. The spectra showed a shift in the inflectionpoint energy according to the increase in the oxidation state of U, as illustrated for

Fig. 2. Scanning electron microscopy of a depleted uranium particle. (a) Particle originating from the

impact of DU munitions collected by swipe (#80): backscattered electron imaging (BEI) mode, bright

areas reflect high atomic number elements and (b) elemental analysis by XRMA. Bar 20 mm. (c) Large

particle originating from fire in DU ammunition storage facility (#85), collected from sand in Kuwait:

backscattered electron imaging (BEI) mode, bright areas reflect high atomic number elements and (d)

elemental analysis by XRMA. Bar 50 mm.

Table 2

Uranium isotope ratios in DU particles and natural U (certified reference uranium mine ores) determined

by ICP-MS and AMS

Sample description Isotope ratioG SD

n 235U/238Ua 236U/235Ub 236U/238Uc

DU munition

storage fire

4 0:00198G 0:00006 0:0097G 0:0005d ð2:0G 0:1Þ!10�5

Swipe from

inside damaged tank

4 0:0021G 0:0001 0:00097G 0:0004 ð2:1G 0:9Þ!10�6

Uranium mine ores 2 0:0071G 0:0007 ð3:5G 0:5Þ!10�7 ð2:5G 0:5Þ!10�9

a ICP-MS.b AMS.c Calculated by combining the ICP-MS and AMS data.d Mean of one AMS and three ICP-MS measurements.


UO2, U3O8 and the solid uranyl standards (Fig. 3a). Compared to the m-XANESspectra obtained from the standards, all DU particles investigated were oxidized (noU metal present). As shown in Fig. 3a, the m-XANES spectra for the small-sized DUparticles (No. 80, 82, 95) originating from the corrosion of DU penetrators that didnot hit the target or from corrosion of holes in the metal shield of tanks hit by DUmunitions, reveal profiles and inflection point energies corresponding to that of theUO2 standard (i.e. oxidation state C4:0G 0:5), that of the U3O8 standard (i.e.oxidation state C5:5G 0:5) or a mixture of these oxidation states. No higheroxidation state for Uwas observed in these DU particles. These findings correspondedwell with results obtained for DU particles from Kosovo, where about 50 % of theparticles were UO2, and the remaining particles were U3O8 or a mixture of UO2/U3O8

(Salbu et al., 2003). Transformation from UO2 to U3O8 has also been observed forparticles released from the Chernobyl reactor as oxidized U3O8 particles were releasedfrom the UO2 fuel during the reactor fire (Salbu et al., 2001).

For DU particles associated with the fire in the DU ammunition storage facility,however, (Fig. 3c), the m-XANES spectra indicated the presence of oxidation stateC5 or C6. The spectra showed a shift in the inflection point energy and in the shapeof the curve corresponding to an increase in the oxidation state of U, to a statesimilar to the solid uranyl standards. Thus, these particles clearly reflect theconditions under which they formed, e.g., extreme high temperature under theinfluence of oxygen.

Comparing the m-XRD pattern of the samples with standards, it appears that U ina few particles from the impact of DU munitions was present as UO2 (Fig. 4a). ForDU particles released during the fire, the XRD pattern is more difficult to interpretas several spectra indicated shifted peaks towards higher d-spacing, indicating an

Fig. 3. Oxidation states of U in DU particles. (a) m-XANES spectra of U in small-sized DU particles

originating from impact of DU munitions compared to UO2 and U3O8 standards. (b) m-XANES spectra of

U in large DU particles originating from fire in DU ammunition storage facility compared to different

uranyl standards.


Fig. 4. X-ray diffraction of DU particles. (a) XRD pattern of U in small-sized DU particles originating

from impact of DU munitions compared to an UO2 standard. (b) XRD pattern of U in a large DU particle

originating from fire in DU ammunition storage facility compared to U2O5 standard (literature data). (c)

XRD pattern of U in large DU particles originating from fire in DU ammunition storage facility

compared to UO3 standard (literature data).


expanded lattice. In a few particles, U showed a pattern similar to U2O5 (Fig. 4b),while most of the investigated particles originated from the fire could be attributed tosolid uranyl compounds (e.g., UO2(CO3)), although the fit to UO3 literature XRDdata suffers from uncertainties (Fig. 4c). Thus, the m-XRD analysis confirmed toa large extent results obtained from m-XANES.

As the particle weathering rate increases with the oxidation state for U and shouldbe higher for UO3 and U3O8 particles than for UO2 (Kashparov et al., 1999), thesubsequent mobilization of U from highly oxidized U particles and the trans-membrane transfer should be enhanced compared with natural uranium (NU)particles, when in contact with body fluids or natural water systems.

4. Conclusions

Based on m-XANES techniques and confirmed to a large extent by m-XRD, thepresent results demonstrate that DU particles in Kuwait are present in oxidationstates ranging from C4 to C6. Small-sized DU particles collected from swipes orfrom below penetrators were similar to those observed in Kosovo; present as UO2,U3O8 and as a mixture of UO2 and U3O8. In areas contaminated with DU particlesoriginating from the fire in the DU ammunition storage facility, however, theparticle characteristics differed from all previous observations of U-particles.Relatively large (0.2–1500 mm, median 44 mm) yellow particles were crystalline andm-XANES spectra corresponded to oxidation state C5 or C6, similar to U2O5 or tosolid uranyl standards.

As particle weathering rates, remobilisation and bioavailability of U will increasewith oxidation states of U, oxidized DU particles such as U3O8, U2O5 or especiallyUO3 particles will behave different from natural uranium. Thus, environmental orhealth impact assessments for areas affected by DU munitions should take intoaccount the presence of respiratory UO2, U3O8 and even UO3 particles, theircorresponding weathering rates and the subsequent mobilisation of U from oxidizedDU particles. As the radioactivity of DU is lower than for natural uranium impactassessments should focus on DU as a heavy metal contamination problem.

The present work also demonstrates that advanced solid-state speciationtechniques, such as SR-based X-ray m-XANES and m-XRD providing informationon crystallographic structure and oxidation states of radionuclides associated withradioactive particles, should be considered most useful within radioecology.

Acknowledgements

The authors are indebted to professor D.H. Oughton and scientist L. Skipperudfrom the Isotope Laboratory, Agricultural University of Norway and L.K. Fifield,National University Canberra for measurements of U isotope ratios. The project isfunded by EU (Contract No. FIGE–CT–2000–00108 ADVANCE), and theNorwegian Research Council (Contract no. 141479/720).


References

Clacher, A.P., 1995. Development and application of analytical methods for environmental radioactivity.

PhD. Thesis, University of Manchester, UK.

Danesi, P.R., Bleise, A., Burkart, W., Cabianca, T., Campbell, M.J., Makarewicz, M., Moreno, J., 2003a.

Isotopic composition and origin of uranium and plutonium in selected soil samples collected in

Kosovo. Journal of Environmental Radioactivity 64, 121–131.

Danesi, P.R., Markowicz, A., Chinea-Cano, E., Burkart, W., Salbu, B., Donohue, D., Ruedenauer, F.,

Hedberg, M., Vogt, S., Zaharadnik, P., Ciurapinski, A., 2003b. Depleted uranium particles in selected

Kosovo samples. Journal of Environmental Radioactivity 64, 143–154.

Fetter, S., von Hippel, F.N., 1999. The hazard posed by depleted uranium munitions. Science and Global

Security 8, 125–161.

IAEA, 2003. Radiological conditions in areas of Kuwait with residues of depleted uranium. IAEA,

Vienna, (STI/PUB/1164.ISBN 92-0-106603-1).

Kashparov, V.A., Oughton, D.H., Protsak, V.P., Zvarisch, S.I., Protsak, V.P., Levchuk, S.E., 1999.

Kinetics of fuel particle weathering and 90Sr mobility in the Chernobyl 30 km exclusion zone. Health

Physics 76, 251–259.

Marsden, O.J., Livens, F.R., Day, J.P., Fifield, L.K., Goodall, P.S., 2001. Determination of U-236 in

sediment samples by accelerator mass spectrometry. Analyst 126, 633–636.

Priest, N.D., 2001. Toxicity of depleted uranium. The Lancet 357, 244–246.

Salbu, B., 2000. Actinides associated with particles. In: Kudo, A. (Ed.), Plutonium in the Environment.

Elsevier Science Ltd, Tokyo, pp. 121–138.

Salbu, B., Krekling, T., Oughton, D.H., :stby, G., Khasparov, V.A., Day, J.P., 1994. The significance of

hot particles in accidental releases from nuclear installations. Analyst 119, 125–130.

Salbu, B., Krekling, T., Oughton, D.H., 1998. Characterisation of radioactive particles in the

environment. Analyst 123, 843–849.

Salbu, B., Janssens, K., Krekling, T., Simionovici, A., Drakopoulos, M., Raven, C., Snigireva, I., Snigirev,

A., Lind, O.C., Oughton, D.H., Adams, F., Kashparov, V.A., 2000. m-XAS-tomography and

m-XANES for characterisation of fuel particles. ESRF Highlights 1999. European Synchrotron

Research Facility, Grenoble.

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., 2001. High energy X-ray

microscopy for characterisation of fuel particles. Nuclear Instrumentation and Methods A 467 (21),

1249–1252.

Salbu, B., Janssens, K., Lind, O.C., Proost, K., Danesi, P.R., 2003. Oxidation states of uranium in DU

particles from Kosovo. Journal of Environmental Radioactivity 64, 163–167.

The Royal Society report–UK, 2001. The health hazards of depleted uranium munitions, part 1, Policy

Document 7/701.

oxidation states of uranium in depleted uranium particles from kuwait

Documents