Determination of depleted uranium in environmental ...

Journal of Environmental Radioactivity 76 (2004) 295–310 www.elsevier.com/locate/jenvrad

Determination of depleted uranium in environmental samples by gamma-spectroscopic techniques D.J. Karangelos a, M.J. Anagnostakis a,, E.P. Hinis a, S.E. Simopoulos a, Z.S. Zunic b a

b

Nuclear Engineering Section, Mechanical Engineering Department, National Technical University of Athens, 15780 Athens, Greece Radiation Medicine Department, Institute of Nuclear Sciences, ‘‘Vincˇa’’, Belgrade, Yugoslavia

Received 31 March 2003; received in revised form 14 November 2003; accepted 25 November 2003

Abstract The military use of depleted uranium initiated the need for an efficient and reliable method to detect and quantify DU contamination in environmental samples. This paper presents such a method, based on the gamma spectroscopic determination of 238U and 235U. The main advantage of this method is that it allows for a direct determination of the U isotope ratio, while requiring little sample preparation and being significantly less labor intensive than methods requiring radiochemical treatment. Furthermore, the fact that the sample preparation is not destructive greatly simplifies control of the quality of measurements. Low energy photons are utilized, using Ge detectors efficient in the low energy region and applying appropriate corrections for self-absorption. Uranium-235 in particular is determined directly from its 185.72 keV photons, after analyzing the 235U–226Ra multiplet. The method presented is applied to soil samples originating from two different target sites, in Southern Yugoslavia and Montenegro. The analysis results are discussed in relation to the natural radioactivity content of the soil at the sampling sites. A mapping algorithm is applied to examine the spatial variability of the DU contamination. # 2004 Elsevier Ltd. All rights reserved. Keywords: Gamma spectroscopy; Depleted uranium; Soil; TENORM

Corresponding author. Tel.: +30-210-7722912; fax +30-210-7722914. E-mail address: [email protected] (M.J. Anagnostakis).

0265-931X/$ - see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvrad.2003.11.011

296

D.J. Karangelos et al. / J. Environ. Radioactivity 76 (2004) 295–310

1. Introduction The use of depleted uranium (DU) penetrators in military operations results in contamination of the target area, because of the DU aerosol formation during impact of the penetrator on a hard surface. DU aerosol is then dispersed into the environment at a distance that strongly depends on various parameters such as meteorological conditions and aerosol size distribution. Since the uranium isotopes existing in DU penetrators are naturally occurring radionuclides, DU is characterized as a Technologically Enhanced Naturally Occurring Radioactive Material (TENORM). The main problem in the determination of DU contamination in environmental samples is the interference caused by natural uranium, which is detected in almost all environmental samples as a trace element. The abundances of the U isotopes in DU, reported to 0.2% for 235U, 0.001% for 234U and the remaining (~99.8%) 238U (McClain et al., 2001), are significantly different from those in natural uranium, which are 0.72% for 235U, 0.0057% for 234U and 99.27% for 238U (Kathren, 1984).This characteristic difference forms the basis of several methods for the determination of DU contamination by determining the U isotope ratios, either by nuclear methods such as alpha or gamma spectrometry or by mass spectroscopic techniques such as ICP-MS. The present communication will report on a method for the determination of DU contamination by gamma spectrometry. Therefore, in the following, whenever detection of DU contamination is mentioned, it must be understood as detection of a statistically significant disruption of the natural U isotopic ratios. As 234U does not emit suitable photons for gamma spectrometry, the detection of DU contamination by gamma spectrometry is mainly based on the determination of the 235U/238U ratio. Furthermore, since 238U does not emit significant photons that can be used for its gamma spectroscopic determination, it is usually determined through its daughter products in equilibrium, namely 234Th and 234m Pa. Thorium 234, which, due to its short half-life (24 d) establishes secular equilibrium with 238U within a reasonable time period, emits low energy photons with low yield at the energies: 63.29 keV (3.8%), 92.35 keV (2.72%) and 92.78 keV (2.69%). The photons at 92.35 and 92.78 keV are very close, resulting in a multiplet photopeak, which is very difficult to resolve. For this reason these two peaks are rarely used, although there exist cases where the multiplet photopeak is treated approximately as a singlet with a yield of 5.41% (Roessler et al., 1979). Although 234mPa has a short half-life (6.7 h) and therefore is practically always in equilibrium with its parent, 234Th, it emits 1001.03 keV photons with very low yield (0.837%). These photons have been employed for the determination of DU in environmental samples, although no positive results were reported (Uyttenhove et al., 2002). Taking the above into account, the photons of 234Th at 63.29 keV were used for the determination of 238U in the present work. The activity of 235U in environmental samples is seldom reported, mainly due to the difficulties in its determination. Uranium 235 emits several photons, such as at


297

143.76 keV (10.96%), 163.33 keV (5.08%), 185.72 keV (57.2%) and 205.31 keV (5.01%). However, due to its long half-life (7.04108 years) and the very low photon emission rate of all these photons in environmental samples, the only photons that can be used for the determination of 235U are those at the energy of 185.72 keV. The problem with the analysis of these photons is that 226Ra, a daughter of 238U that emits 186.25 keV photons, co-exists with 235U in environmental samples, and the photons which these two nuclides emit result in a multiplet photopeak at ~186 keV. A common approach to this problem is to determine the activity of 226Ra from its daughters in equilibrium and subtract its contribution from the multiplet to get the 235U activity. This technique however requires that secular equilibrium between 226Ra and its daughter 222Rn be established, which can only be accomplished if the samples are hermetically sealed, after a waiting period of not less than 20 days. Furthermore, when using this approach, any uncertainties in the determination of 226Ra will propagate to the determination of 235U. However, if using a Ge detector with sufficient resolution at this energy it may be possible to resolve the two components of the multiplet at 186 keV. This is the technique that is actually developed and used in the present work. This communication reports: 1. A method for the gamma-spectroscopic determination of 238U and 235U in environmental samples, in order to determine any DU contamination. 2. An application of this method to soil samples from DU targeted areas in Southern Serbia and Montenegro. 3. Mapping of the DU contamination around the point of impact of a DU penetrator at a target site in Southern Serbia. 4. Gamma-spectroscopic determination of the 235U/238U ratio of a DU penetrator, collected at a target site.

2. Detection and determination of DU contamination in soil samples Since the only definite evidence of DU contamination in soil samples is the disruption of the natural isotopic abundance of 238U and 235U, an accurate determination of the activities of these nuclides should be performed. The 235U/238U atom ratio r can then be determined using the simple formula: r¼

Að235 UÞT1=2 ð238 UÞ Að238 UÞT1=2 ð235 UÞ

where A is the activity and T1/2 is the half-life of the respective nuclide. A soil sample may be considered as DU contaminated when the measured value of the 235 U/238U atom ratio differs to a statistically significant degree from its natural value of 0.7253%. Therefore, the ability to detect DU contamination depends on the uncertainty in the determination of the 235U/238U atom ratio, as well as the natural U content of the sample.

298


For the analysis of 238U and 235U in environmental samples, such as soil, two detectors are used in the Nuclear Engineering Section of the National Technical University of Athens (NES-NTUA): . A low energy Germanium (LEGe) detector, with a FWHM of 341 eV at 5.9 keV and 530 eV at 122 keV, calibrated in the energy region 0–200 keV. . An extended range Germanium (XtRa) detector, with a FWHM of 1003 eV at 122 keV and a relative efficiency of 107%, calibrated in the energy region 0– 2000 keV. According to the standard procedure for the gamma spectroscopic analysis of environmental samples applied in NES-NTUA, the samples fill completely 282 ml plastic cylindrical boxes. Since this analysis also includes the determination of 226 Ra and 224Ra through their gamma emitting daughters, the samples are then hermetically sealed and covered with a film of epoxy resin. It has been experimentally verified that the epoxy resin used is an efficient radon barrier, ensuring the establishment of secular equilibrium of the radium isotopes with their decay products inside the boxes; all samples are analyzed at least three weeks after the boxes are sealed. The 226Ra activity is derived from the weighted mean of the activities of two photopeaks of 214Pb (295.2, 352.0 keV) and three photopeaks of 214 Bi (609.3, 1120.3, 1764.5 keV). Further details on the calibrations performed, the c-spectroscopic analysis and the quality assurance procedures followed, may be found elsewhere (Simopoulos and Angelopoulos, 1987). The count-rate of the low energy photons used for this analysis, such as those of 234 Th at 63.29 keV, is highly affected by the intense self-absorption of the photons inside the sample, especially in large volume samples, depending on the sample geometry, the type of material, the packing density and the photon energy. It is also important to mention that the self-absorption of photons in the material is generally different than that of the efficiency calibration standard. A method has been developed in NES-NTUA (Anagnostakis and Simopoulos, 1995a), for the quantitative determination of radionuclides emitting low energy photons. This method takes into account the difference in the self-absorbing properties between the calibration source material and the sample material, introducing an efficiency correction factor. The efficiency correction factor is determined on the basis of the linear attenuation coefficient l(q,E) correlations determined experimentally for photon energies between 46 and 186 keV and soil densities between 0.5 and 2 g cm3 (Anagnostakis, 1998; Petropoulos et al., 2001). The technique used in this work for the determination of 235U is based on the analysis of the multiplet photopeak at ~186 keV using the LEGe detector calibrated in the low energy region 0–200 keV, which allows for sufficient number of channels per keV (~20 channels/keV). The analysis is performed using the SPUNAL (SPectrum UNix AnaLysis) software, a computer code developed inhouse for the analysis of complex gamma spectra with up to 10 component multiplets, using a modified Marquardt algorithm (Petropoulos et al., 2001). It should be noted that, in order to obtain a statistically adequate number of counts, analysis


299

Fig. 1. The 186 keV multiplet.

times of the order of 3105 s are required. Two examples of the analysis of this multiplet are given in Fig. 1. In conclusion, for every sample analyzed: . The 238U activity was determined indirectly from the 63.29 keV photons emitted by its decay product 234Th. . The 235U activity was determined directly from its 185.72 keV photons, after analyzing the multiplet at ~186 keV into two components. . The 226Ra activity was determined both indirectly from its decay products in equilibrium and directly from the 186.25 keV photons. It is already reported that the two methods agree, within the total uncertainty of the measurements (Petropoulos et al., 2001). Using the above analysis one can determine the radioactive content of the two uranium isotopes in the sample. However, in order to quantify any DU contamination in the sample, in the cases that the natural 235U/238U atom ratio is disturbed, one of the following assumptions, introduced in (Anagnostakis et al., 2001), has to be adopted: Assumption 1. The DU contaminant is unique and has a known isotopic abundance. A typical value for DU used in kinetic penetrators is 0.2% for 235U and ~99.8 for 238U. This value is in agreement with that determined in the present work by analyzing a penetrator collected in the field, as it will be explained later. In this case the fractions of 238U and 235U activities that should be attributed to the natural uranium of the soil and the DU contaminant may be calculated.

300


Assumption 2. Radioactive equilibrium between 238U and 226Ra exists in the soil prior to DU contamination. In this case, the 238U and 235U activities in the soil prior to DU contamination are determined from the 226Ra activity. Since 238U and 235 U activities in the soil prior to and after contamination are known, the activities of 238U and 235U due to the DU contaminant can be determined, as well as their isotopic abundance, assuming that all the DU present in the sample has the same isotopic composition. However, it is emphasized that the assumption of radioactive equilibrium between 238U and 226Ra in soil should be applied with caution, since this equilibrium is often disturbed (Anagnostakis et al., 2002). Both of these assumptions are examined in the present work, for samples collected at two sites with different natural radioactivity characteristics. 3. Uncertainty analysis and lower limit of detection For the analysis technique applied in the present work, the uncertainty in the activity of a radionuclide can be broken down into the following components: . Counting and peak analysis uncertainty, estimated by the spectrum analysis software. This includes counting statistical uncertainty as well as the uncertainty of peak fitting and deconvolution where applicable. . Uncertainty due to the calibration source, quoted in the source certificate by the manufacturer. . Uncertainty due to the calibration curve fitting, estimated during the fitting process. . Uncertainty due to the self-absorption correction. For the specific detector systems employed in the present work, the systematic uncertainty component values have been estimated as follows (1r level): . Calibration source uncertainty less than 1.5%. . Calibration curve fitting uncertainty less than 1%. . Self-absorption correction uncertainty less than 1%. The uncertainty components are combined following the root-sum-of-squares rule to give the combined standard uncertainty for each nuclide. The uncertainties calculated in this manner for 235U and 238U are then substituted in the uncertainty propagation formula to give the uncertainty rr for the 235U/238U atom ratio r: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 235 2 238 2 rð UÞ rð UÞ rr ¼ r þ Rð235 UÞ Rð238 UÞ This uncertainty can be used in a statistical test of the atom ratio against the natural value and to determine the critical level and lower limit of detection (Currie, 1968). It must be underlined that, as the counting and analysis uncertainty for the determination of the natural U radionuclides depends on the total U activity


301

present, the lower limit of detection for DU can only be calculated at a given level of natural U activity in the sample. In order to estimate the lower limit of detection for the technique applied in the present work, we used data from a large (~100) number of analyzed soil-type samples to estimate an empirical correlation between the radioactivity and the relative combined standard uncertainty for the determination of 235U and 238U. These correlations are presented in Fig. 2, and are given by the following equations, where activities are in Bq kg1: rð238 UÞ ¼ 0:18Rð238 UÞ0:29 Rð238 UÞ rð235 UÞ ¼ 0:126Rð235 UÞ0:40 Rð235 UÞ with RMS deviations of 23% and 34%, respectively. Using these estimates for the uncertainties we have determined the lower limit of detection for DU numerically, for soils with a natural activity concentration of 238U ranging from 10 to 1000 Bq kg1. For this calculation, we assumed a false positive probability of 5% and a false negative probability of 1%. The 238U activity due to DU corresponding to the lower limit of detection determined in this way can be seen in Fig. 3 both as an absolute value as well as relative to the natural 238U activity present in the soil. It can be seen that the lower limit of detection 238U activity due to DU ranges from more than 100% of the natural 238U activity at the lower end of the range examined down to approximately 20% for the highest natural activity concentrations considered.

Fig. 2. Empirical correlations between uncertainty and activity concentration for

235

U and 238U in soil.

302


Fig. 3. Uranium-238 activity concentration due to DU corresponding to the lower limit of detection against natural 238U activity concentration.

As an example, for soil containing 150 Bq kg1 of natural 238U, the lower limit of detection for DU corresponds to 62 Bq kg1 of 238U due to DU. This soil would contain 6.9 Bq kg1 of 235U, while the DU contaminant at the lower limit of detection would introduce an additional 0.8 Bg kg1. For this soil contaminated with DU at the lower limit of detection, the 235U/238U atom ratio would be equal to 0.57%, while the combined standard uncertainty in its determination is estimated to 0.038%. 4. Sampling The Nuclear Engineering Section of NTUA and Vincˇa Institute have been in cooperation since 1996, studying the natural radioactivity background in Yugoslavia. After the military operations in Kosovo in 1999, this cooperation was renewed, aiming to the determination of DU contamination in environmental samples and especially in soil and vegetation from DU target areas in the south of Serbia. An initial field mission was conducted in November 1999. A total of 20 samples—11 soil and nine vegetation samples—were collected from various target sites in Yugoslavia by scientists from Vincˇa Institute, and analyzed by NES-NTUA using the detectors and the gamma spectroscopic techniques described in the present work. Depleted uranium was detected in four soil samples, collected inside a DU crater in the area of Bratoselce, with the 238U content due to DU contamination exceeding in one case 2200 Bq kg1 (Anagnostakis et al., 2001). A DU penetrator was also recovered from the Bratoselce site. The site where contamination was detected was visited again in November 2001, aiming to study the spatial distribution of contamination around an impact


303

Fig. 4. Bratoselce site sampling grid.

point by applying small-scale grid sampling. A total of 24 surface soil samples (1 cm thick) were collected on a regular grid with diameter 10 m, centered on a visible penetrator crater. A sample thickness of 1 cm, which is customarily used in NES-NTUA sampling campaigns, was also adopted in this case. Although sampling of a thinner soil layer might result in more pronounced values of the 235 U/238U ratio, it would be more difficult to control and was therefore not preferred. Sample locations relative to the penetrator crater are presented in Fig. 4. A target site in Cape Arza, Montenegro was also visited in November 2001, where nine soil samples were collected, again around a visible DU crater. Although grid sampling was planned for this site as well, the morphology of the site did not allow for a grid diameter larger than 1 m, which is too small to allow any reliable conclusion to be drawn regarding the spatial distribution. It should be noted that the UNEP mission to Serbia and Montenegro investigated both these sites (UNEP, 2002). 5. Analysis results The results of the gamma-spectroscopic analyses of the samples collected in Bratoselce are presented in Table 1. In this table, as in all subsequent tables, uncertainty is given as combined uncertainty, at a 99% confidence level (3r). Depleted uranium contamination was detected in most of the samples. For comparison

304


Table 1 Analysis results—Bratoselce site 1 cm thick surface soil (Fig. 4) Sample code

Grid point

226 Ra (Bq kg1)

238

U (234Th) (Bq kg1)

235 U (Bq kg1)

235

U/238U (atom ratio, %)

238 U/226Ra (activity ratio)

MS161 MS162 MS163 MS164 MS165 MS166 MS167 MS168 MS169 MS170 MS171 MS172 MS173 MS174 MS175 MS176 MS177 MS178 MS179 MS180 MS181 MS182 MS183 MS184

East, 15 cm East, 50 cm East, 100 cm East, 200 cm East, 300 cm East, 500 cm West, 15 cm West, 50 cm West, 100 cm West 200 cm West, 300 cm West, 500 cm North, 15 cm South, 15 m South, 50 cm North, 100 cm a North, 100 cm South, 100 cm North, 200 cm South, 200 cm North, 300 cm South, 300 cm North, 500 cm South, 500 cm

118 9:8 120 10 130 11 120 10 104 9:1 120 10 119 9:8 106 8:9 104 8:6 87 7:1 90 15 140 12 130 10 117 9:8 130 11 102 8:4 119 9:7 121 9:9 98 8:3 130 27 102 8:4 130 10 76 6:2 130 10

430 42 170 24 140 21 140 14 150 18 100 13 150 15 130 20 120 18 160 17 170 16 170 22 3500 370 150 19 150 18 300 28 580 50 120 12 540 47 110 14 1200 100 130 15 180 20 120 13

9 1:4 6 1:1 5 1:1 5:9 0:9 5 1:1 4:9 0:76 5:9 0:79 51 4:9 0:9 5:2 0:79 5:1 0:79 6:2 0:85 50 6 61 6 1:1 7:0 0:86 10 1:8 4:3 0:76 10 1:3 4:9 0:94 16 3:1 5:6 0:91 5:0 0:74 4:4 0:69

0:34 0:061 0:6 0:13 0:6 0:15 0:6 0:12 0:5 0:13 0:8 0:15 0:6 0:1 0:6 0:15 0:6 0:15 0:51 0:096 0:47 0:085 0:6 0:1 0:22 0:036 0:7 0:14 0:6 0:14 0:37 0:057 0:28 0:055 0:6 0:11 0:29 0:044 0:7 0:17 0:22 0:046 0:7 0:14 0:43 0:078 0:60 0:12

3:6 0:47 1:4 0:23 1:1 0:18 1:2 0:15 1:5 0:21 0:8 0:13 1:3 0:17 1:3 0:21 1:2 0:2 1:8 0:24 1:9 0:35 1:2 0:19 27 3:6 1:3 0:2 1:2 0:17 2:9 0:36 4:9 0:58 1 0:13 5:5 0:67 0:8 0:2 11 1:4 1 0:15 2:4 0:33 0:90 0:13

a

Sampled at 5 cm depth (1 cm thick)—not included in mapping.

purposes, ranges for the activity concentration of 238U and the 238U/226Ra ratio of 16 Yugoslavian surface soil samples collected before the Kosovo conflict, as well as from a much larger data set (286) of Greek surface soil samples are presented in Table 2. It is evident from this table that the ratio 238U/226Ra for natural soil may reach values as high as 17, which should be attributed to the physicochemical characteristics of the soil, the surrounding rocks and any weathering process (Coles and Meadows, 1976; Anagnostakis et al., 2002). The UNEP mission to Serbia and Montenegro, which investigated several target sites, reported values in the range of 12.6–3.07105 Bq kg1 for 238U activity concentration, with the highest values observed in the case of soil collected in the immediate vicinity of a penetrator (UNEP, 2002). The amounts of DU contamination present in those samples where DU contamination was most evident, calculated according to the two assumptions presented earlier, are presented in Table 3. The results obtained by using Assumption 1 indicate that radioactive equilibrium at the sampling site was not significantly disturbed before contamination. This is in agreement with the results obtained from


305

Table 2 Reference ranges for 238U activity and the 238U/226Ra activity ratio of 1 cm thick surface soil

Sample size 238 U (Bq kg1) 226 Ra (Bq kg1) 238 U/226Ra (activity ratio)

Greek surface soil

Pre-war Yugoslavian surface soil

286 4–480 0.5–372 0.21–17

16 6.5–326 29–336 0.71–1.59

the samples where DU is not detected, where equilibrium or small deviations from equilibrium are observed. Application of Assumption 2 is possible only for a few of the samples, as the amount of 235U in some of them is less than what would be expected if 238U and 226 Ra were in equilibrium. However, the mean value of the DU contaminant 235 U/238U atom ratio, calculated for the samples where this was possible, is equal to 0.18% with a standard deviation of 0.037%. Furthermore, as illustrated in Fig. 5, where the DU contaminant 235U/238U atom ratio has been plotted against the total 238U activity for each sample, most of the variability in these results is due to the samples where contamination is low. In order to further confirm this result, a small amount of DU (~0.1 g) was removed from the surface of the penetrator that was recovered in Bratoselce, using a steel file. Concentrated HNO3 was added to the filings and evaporated repeatedly. The solution was finally filtered, converted to 4 M HCl and made up to a Table 3 DU content – Bratoselce site 1cm thick surface soil. Sample 235U/238U (atom code ratio, %)

MS161 MS165 MS167 MS170 MS171 MS172 MS173 MS176 MS177 MS179 MS181 MS183

0:34 0:061 0:5 0:13 0:6 0:1 0:51 0:096 0:47 0:085 0:6 0:1 0:22 0:036 0:37 0:057 0:28 0:055 0:29 0:044 0:22 0:046 0:43 0:078

Assumption 1

Assumption 2

DU isotope ratio equal to 0.2%

Equilibrium before contamination

238

U due to DU (Bq kg1)

238

238 U natural 238U/226Ra U due to 1 (Bq kg ) before DU (Bq kg1) contamination (activity ratio)

U/238U of contaminant (atom ratio, %)

320 72 50 41 30 32 60 33 80 32 60 40 3300 540 200 46 490 88 450 76 1100 170 110 35

110 83 100 45 120 35 90 37 90 36 120 46 200 650 100 54 100 100 100 89 – 80 40

0:19 0:078 0:2 0:37 0:2 0:4 0:3 0:19 0:2 0:19 – 0:20 0:036 0:18 0:077 0:17 0:065 0:2 0:05 0:17 0:049 0:2 0:12

0:9 0:71 1:0 0:44 1:0 0:31 1:1 0:44 1:0 0:42 0:8 0:33 1:0 5:1 0:9 0:53 0:8 0:85 1:0 0:91 – 1:0 0:54

310 43 50 20 30 18 70 18 80 22 30 25 3400 370 200 29 460 51 450 48 1100 100 110 21

235

306


Fig. 5. 235U/238U atom ratio attributed to DU against total Bratoselce site.

238

U activity of soil samples from the

standard volume of 282 ml for gamma-spectrometric analysis, which determined the 235U/238U atom ratio of the penetrator to be equal to (0:21 0:01)%. The values of the 235U/238U atom ratio of the DU contaminant determined in the present work, both directly from analysis of the penetrator and indirectly from the soil samples analysis, are in good agreement with the value of 0.2% reported in the literature. The results of the gamma-spectroscopic analyses of the samples collected in Cape Arza are presented in Table 4. These samples proved to be particularly interesting, as the 226Ra concentration was found in many cases higher than that of 238 U. Based on this information alone, one might conclude that DU is not present. However, examination of the 235U/238U ratio indicates that DU is indeed present; presumably, 226Ra/238U equilibrium was disturbed in favor of 226Ra prior to contamination. In this case it is clear that it is unreasonable to apply Assumption 2 for the quantitative determination of DU contamination. Table 5 shows the amount of 238 U due to DU present in each sample, assuming that the isotope ratio of the contaminant is equal to 0.2%. It is interesting to comment on the amount of natural 238 U present in these samples: it appears that the 226Ra–238U equilibrium was severely disturbed before contamination, with a minimum value of the 238U/226Ra activity ratio as low as 0.07. For comparison purposes note that the minimum value measured in Greek surface soil is equal to 0.21, while values as low as 0.11 have been reported for upland organic soil (Dowdall and O’Dea, 2002).


307

Table 4 Analysis results—Cape Arza site 1 cm thick surface soil Sample code

226 Ra (Bq kg1)

238 U (Bq kg1)

235 U (Bq kg1)

235 U/238U (atom ratio, %)

238

U/226Ra (activity ratio)

MS185 MS186 MS187 MS188 MS189 MS190 MS191 MS192 MS193

350 29 370 30 400 32 390 32 370 30 290 23 510 41 330 27 360 29

170 20 340 37 490 45 330 30 360 33 200 22 270 29 170 20 540 48

3:1 0:39 6 1:4 8 1:3 5 1:1 6:4 0:9 3:3 0:91 6 1:5 3:5 0:88 8 1:3

0:29 0:051 0:29 0:071 0:27 0:048 0:24 0:057 0:28 0:048 0:27 0:078 0:33 0:096 0:33 0:092 0:24 0:045

0:48 0:069 0:9 0:12 1:2 0:15 0:8 0:1 1 0:12 0:69 0:094 0:53 0:072 0:50 0:073 1:5 0:180

Information regarding other radionuclides detected in the samples collected at the two sampling sites is presented in Table 6. It is interesting to underline the large variability in the concentration of 137Cs, probably due to the Chernobyl fallout. This variability may be of significance when surveying for DU ‘‘hot spots’’ with hand-held instrumentation. Some authors (Bikit et al., 2001) have suggested a method to detect DU by examining the 226Ra–238U equilibrium, using disequilibrium in favor of 238U as evidence of DU contamination. In order to test this method, the 226Ra/238U ratio of all samples analyzed has been plotted against the 235U/238U ratio in Fig. 6, along with the relevant ranges for Greek surface soil and a limited set of Yugoslavian surface soil samples. It appears that the 226Ra/238U ratio cannot be used to distinguish between contaminated and non-contaminated samples, except for the cases where contamination is very high. Table 5 DU content—Cape Arza site 1 cm thick surface soil Sample code

MS185 MS186 MS187 MS188 MS189 MS190 MS191 MS192 MS193

235

U/238U (atom ratio, %)

0:29 0:040 0:29 0:063 0:27 0:038 0:24 0:050 0:28 0:037 0:27 0:073 0:33 0:089 0:33 0:085 0:24 0:037

Assumption 1 DU isotope ratio equal to 0.2% 238

U due to DU (Bq kg 1)

238

U natural (Bq kg1)

238

U/226Ra before contamination (activity ratio)

140 23 280 52 430 45 300 36 300 32 170 33 200 52 130 33 490 47

30 27 60 57 70 50 30 38 60 36 20 36 70 55 40 36 40 51

0:09 0:077 0:2 0:15 0:2 0:13 0:07 0:096 0:16 0:097 0:1 0:13 0:1 0:11 0:1 0:11 0:1 0:14

308


Table 6 Other radionuclides detected at the sampling sites Bratoselce—24 samples

232

Th K 137 Cs 40

Cape Arza—9 samples

Range

Mean

Standard deviation

Range

Mean

Standard deviation

44–68 799–1131 3–232

59 1000 100

5.2 90 62

52–89 430–567 19–121

80 500 50

11.0 40 38

All values in Bq kg1.

6. Mapping of the DU contamination in the immediate area of an impact point In order to examine the spatial distribution of DU contamination around the point of impact of a DU penetrator, the depleted uranium deposition pattern within the sampling area around the targeted point in Bratoselce was produced using an in-house built Unix-based complex data base/geographical information system (DBGIS) (Anagnostakis et al., 1995b). This DBGIS system contains: (1)

Fig. 6. 238U/226Ra activity ratio against ranges.

235

U/238U atom ratio of all samples analyzed, with reference


309

Fig. 7. Mapping of the 235U/238U atom ratio around a DU penetrator impact point.

detailed information regarding each soil sample collected, including the geographical coordinates of the sampling location and (2) advanced plotting routines. The 235U/238U ratio was selected as the most appropriate parameter for mapping, as it is a definite indicator of DU contamination, independent of any assumption. From the map produced (Fig. 7), it can be seen that the contamination is strongly inhomogeneous, as well as asymmetrically distributed around the point of impact. It should be noted that, although the sampling grid is centered on an impact point, the existence of other impact points or penetrators buried in the soil within the sampling area cannot be ruled out. 7. Conclusions The results presented in this work conclude that gamma spectroscopy can be applied for the determination of depleted uranium contamination in environmental samples. Detection is based on a direct determination of 235U, and thus it is reliable and not subject to interference from other physicochemical processes. Furthermore, very little sample preparation is required and the analysis is non-destructive. Analysis of the uncertainties involved in this type of measurement has shown that the lower limit of detection for DU contamination in a soil sample depends on the natural U content. Both the analyses of contaminated soil samples as well as the analysis of a DU penetrator recovered in the field indicate that the value of 0.2% for the 235U/238U atom ratio of DU penetrators found in the literature is absolutely reasonable.

310


Therefore, it can safely be used for the quantitative determination of the DU content of a sample, and should be preferred over an assumption about the status of the 226Ra/238U equilibrium in the soil before contamination. The amount of DU in soil samples collected at target sites in Serbia and Montenegro was found to range up to 3300 360 Bq kg1 of 238U. Contamination was found to be strongly inhomogeneous within the 5 m-radius grid around the point of impact used. This inhomogeneity should be taken into account when planning field sampling for the determination of DU. References Anagnostakis, M.J., 1998. Gamma-spectroscopic analysis of low radioactivity samples in the low energy region. Ph.D. thesis, NTUA, Athens (in Greek). Anagnostakis, M.J., Simopoulos, S.E., 1995a. An experimental numerical method for the efficiency calibration of Low Energy Germanium detectors. Environment International 22(1), S93–S99. Anagnostakis, M.J., Hinis, E.P., Simopoulos, S.E., Angelopoulos, M.G., 1995b. Natural radioactivity mapping of Greek surface soils. Environment International 22(1), S3–S8. Anagnostakis, M.J., Hinis, E.P., Karangelos, D.J., Petropoulos, N.P., Rouni, P.K., Simopoulos, S.E., Zunic, Z.S., 2001. Determination of depleted uranium in environmental samples by gamma-spectroscopic techniques. Archive of Oncology 94(4), 231–236. Anagnostakis, M.J., Hinis, E.P., Simopoulos, S.E., 2002. 238U and its daughter products in Greek surface soils. In: Proceedings of the 7th International Symposium Natural Radiation Environment (NRE-VII), in press. ˇ onkic´, L., Varga, E., C ´ urcˇic´, S., Mrda, D., 2001. Bikit, I.S., Slivka, J.M., Krmar, M., Veskovic´, J., C Determination of depleted uranium at the Novi Sad low-level laboratory. Archive of Oncology 94(4), 241–243. Coles, D.G., Meadows, J.W.T., Lindeken, C.L., 1976. The direct measurement of ppm levels of uranium in soils using high-resolution Ge (Li) gamma-ray spectroscopy. The Science of the Total Environment 5, 171–179. Currie, L.A., 1968. Limits for qualitative detection and quantitative determination—application to radiochemistry. Analytical Chemistry 40, 586–593. Dowdall, M., O’Dea, J., 2002. 226Ra/238U disequilibrium in an upland organic soil exhibiting elevated natural radioactivity. Journal of Environmental Radioactivity 59, 91–104. Kathren, R.L., 1984. Radioactivity in the Environment. Harwood Academic Publishers, Chur, Switzerland. McClain, D.E., Benson, K.A., Dalton, T.K., Ejnik, J., Emond, C.A., Hodge, S.J., Kalinich, J.F., Landauer, M.A., Miller, A.C., Pellmar, T.C., Stewart, M.D., Villa, V., Xu, J., 2001. Biological effects of embedded depleted uranium (DU): summary of Armed Forces Radiology Research Institute research. The Science of the Total Environment 274, 115–118. Petropoulos, N.P., Anagnostakis, M.J., Simopoulos, S.E., 2001. Photon attenuation, natural radioactivity content and radon exhalation rate of building materials. Journal of Environmental Radioactivity 61, 257–269. Roessler, C.E., Smith, Z.A., Bolch, W.E., Prince, R.J., 1979. Uranium and Radium-226 in Florida phosphate materials. Health Physics 37, 269–277. Simopoulos, S.E., Angelopoulos, M.G., 1987. Natural radioactivity releases from lignite power plants in Greece. Journal of Environmental Radioactivity 5, 379–389. United Nations Environmental Program, 2002. Depleted uranium in Serbia and Montenegro: postconflict environmental assessment in the Federal Republic of Yugoslavia. United Nations Environmental Program, Geneva. Uyttenhove, J., Lemmens, M., Zizi, M., 2002. Depleted uranium in Kosovo: results of a survey by gamma spectrometry on soil samples. Health Physics 83, 543–548.