Environ. Sci. Technol. 2003, 37, 4990-4993
Improvement in the Determination of 238U, 228-234Th, 226-228Ra, 210Pb, and 7Be by γ Spectrometry on Evaporated Fresh Water Samples C . C A Z A L A , †,* J . L . R E Y S S , † J. L. DECOSSAS,‡ AND A. ROYER‡ Laboratoire des Sciences du Climat et de l’Environnement, domaine du CNRS, av de la Terrasse 91198 Gif sur Yvette, France, and CEMRAD, Universite´ de Limoges, 83 rue d’Isle 87000 Limoges, France
For the U-Th series radionuclides investigation in natural freshwater, a simple, fast, and not laboratory intensive method which consists of evaporating the water samples to dryness in the presence of carriers is presented. The small volume of the residue (1-2 cm3) leads to a good efficiency for γ counting and limits the self-absorption effect for the low energy γ rays (less than 200 keV). The best efficiency is obtained with a well-type Ge detector. To determine the evaporation yields a river with a common uranium content, the Seine river (France), was selected. By using internal spikes and more conventional techniques of investigation, we demonstrate that the evaporation is quantitative for U, Th, Ra, Pb, and Be. The residue of a 3 L, standard superficial freshwater, evaporated sample was analyzed in a high efficiency, low background Ge detector, which leads to a sufficient precision for most environmental studies. The method has been applied to rain, river, and lake waters to study the impact of disused uranium mine water inputs on the 238U, 228-234Th, 226-228Ra, 210Pb, and 7Be river and lake contents in the U mining area of Limoges (France).
Introduction Many studies have been conducted on the behavior of natural radionuclides in the environment because of their use as tracers and their involvement in the recent development management of the environment (1-7). For radionuclide measurements, the major advantage of γ spectrometry compared with R and β spectrometry consists of the simultaneous determination of several nuclides in a single counting stage (1-8). γ analyses have often been used on suspended particles removed from water by filtration (9) or centrifugation (10), on particles from traps (9), on sediments (9, 4), and on soils (11-14). In these kinds of solid samples, the activities of radionuclides are generally well above the detection limit, and a direct nondestructive investigation is then possible. On the opposite, the activities of radionuclides in common natural freshwater are as low as a few mBq‚L-1, and difficulties arise in accurately assessing the activities. Preconcentration step before γ analysis is required. Coprecipitation with Fe(OH)3 is quantitative for Th (1, 15). * Corresponding author phone: (33) 1 69 82 43 68; fax: (33) 1 69 82 35 68; e-mail:
[email protected]. † Laboratoire des Sciences du Climat et de l’Environnement. ‡ Universite ´ de Limoges. 4990
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 21, 2003
FIGURE 1. Localization of the sampling points. A is upstream the input of mine water (lake 1) while B (brook) and C (lake 2) are downstream. Coprecipitation with Ba as BaSO4 or plastic columns containing Mn-treated acrylic fiber have been widely used to concentrate Ra (1, 16-19). Uranium is generally radiochemically purified and analyzed by R spectrometry (20). To measure 210Pb and/or 7Be in rain, samples can be evaporated and analyzed by γ spectrometry in a Marinelli beaker (10, 21, 22). To study the “Speciation of 226Ra, 238U and 228Ra in an Upland Organic Soil Overlying a Uraniferous Granite”, Dowdall and O’Dea (13) used sequential extractions. Then the liquid extracts were reduced, when necessary, to 1 L by evaporation and placed in a Marinelli beaker for γ analysis. However direct measurements of natural water in a Marinelli beaker are limited by the low detection efficiency due to the geometry and the self-absorption of low energy γ rays (210Pb and 234Th) (23). To remedy these disadvantages, we propose (i) to reduce the volume of the analyzed sample and (ii) to measure the residue in a well-type Ge detector. Evaporation is conducted to dryness in the presence of carriers. The remaining solid is then analyzed by γ spectrometry. The purpose of the present work is first to demonstrate that the method is quantitative for the following seven radionuclidess238U, 234Th, 226Ra, 210Pb, 228Th, 228Ra, and 7Besand then to present applications aimed at measuring these radionuclides in freshwater.
Methods The validation step consists of determining the evaporation yield by (i) the use of gamma emitters spikes (229Th-225Ra) for Th and Ra, (ii) the determination by atomic absorption spectrometry of stable Pb and Be carriers added before evaporation for Be and Pb, and (iii) the measurement of 238U and 226/228Ra activities on aliquot samples by more conventional and quantitative methods (R spectrometry for U and γ spectrometry on a BaSO4 precipitate for Ra). Investigated Sites. The Seine river (France) was selected for the validation experiment because its U activity of about 10 mBq‚L-1 ranges in the average value of world rivers reported by Palmer and Edmon (24). The method was then applied to study the behavior of U-Th series radionuclides in the freshwater of a former uranium mining area near Limoges (France) including two lakes connected by a brook that collects mine waters (Figure 1). Three sampling sites, denoted A, B, and C (Figure 1), were selected. Site A (first lake) is upstream the input of disused mine water on the river, while sites B (brook) and C (second lake) are downstream. Rainwater samples were also analyzed. Chemical Treatment. Twenty liters of shallow water samples was collected with a beaker (polyethylene) and stored in a tank (polyethylene) for transportation to the laboratory. 10.1021/es034333i CCC: $25.00
2003 American Chemical Society Published on Web 09/20/2003
Back to the laboratory (less than 1 h after sampling), the sample was vigorously mixed. Nine liters of unfiltered water was kept for analysis, and the rest (11 L) was filtered though 1.2 µm and 0.45 µm Millipore membrane filters to remove particles larger than 0.45 µm. Filtration through 1.2 µm filters is recommended to limit the clogging of the 0.45 µm membrane, which could decrease the filters nominal pore size (25). Membranes (1.2 µm) were changed regularly (about every 4 L). Unfiltered and 0.45 µm filtered samples were acidified with nitric acid to pH ) 1 to avoid adsorption risks on the beaker’s walls. Particles retained on the filters were directly analyzed by γ spectrometry. For the validation experiment, unfiltered and 0.45 µm filtered samples were separated in three aliquots (2 L each): one for the evaporation experiment and the two others for the analysis based on conventional methods. For samples from Limoges only evaporation was used. Evaporation. Carrierss2 g of Al (as AlCl3), 10 mg of Pb (as Pb(NO3)2), 20 mg of Be as (Be(NO3)2), and 20 mg of Fe (as FeCl3)swere added to the sample to obtain a solid residue at the end of evaporation. Spike (84 mBq of 229Th in equilibrium with its daughter 225Ra) was added to determine the evaporation yield of Th and Ra (17). The volume was reduced to 100 mL by evaporation at 70 °C in a glass beaker (26). Then, to avoid adsorption risks, evaporation to dryness was carried out in a PTFE beaker. Finally, the residue was crushed and sealed in a polypropylene tube for γ spectrometry. Conventional Methods. The procedure proposed by Ku (20) for the determination of U in sediments has been adapted to water samples (27). A 232U/228Th spike was added to a 3 L water sample. Uranium was purified by passing the sample through an anion exchange column and by extraction in Thenoyl-Trifluor-Aceton (TTA). It was finally deposited on an aluminum disk for U determination in an R grid chamber. 226Ra and 228Ra were coprecipitated with BaSO and analyzed 4 by γ spectrometry (28). Gamma Spectrometry. High efficiency, very low background, well-type Ge detectors of 430 cm3 and 930 cm3 [EURYSIS Mesure (presently Canberra-Eurysis)] were used for the measurements. Diameters and depths of the wells are 16 mm and 70 mm, respectively, for the first detector and 21 mm and 100 mm, respectively, for the second one. All materials used in the two detectors have been subjected to a stringent selection process. The cover, the rod, and all metallic components close to the crystal are built of radiopure Al alloy, specially made by the Pechiney company, with less than 0.3 ppb of U and Th isotopes (29). Six standards were used to calibrate the gamma detectors: three from the International Atomic Energy Agency (IAEA135, IAEA375, IAEAsoil6), two from the National Institute of Standards and Technology (NIST, formerly NBS) (NBS4350B and NBS4353), and a mockup of marine sediment with U and Th US standards from NBS at 1000 ppm. The Underground Laboratory of Modane(LSM), jointly operated by the Centre National de la Recherche Scientifique (CNRS) and the Commissariat a` l’Energie Atomique (CEA), is located in the Alps at the midpoint of the Frejus road tunnel, close to the Italian-French border. The elevation is 1260 m, and a 1700 m of rocks overburden reduces the cosmic ray muon fluence rate to 4 m-2‚d-1. The ventilation system which renews the air twice per hour in the γ spectrometry room keeps the radioactivity of radon and its daughter products down to 5-15 Bq‚m-3. The integrated background counting rate from 30 to 3000 keV is as low as 792 counts per day for the 930 cm3 detector and 850 for the 430 cm3, about 3 orders of magnitude lower than at ground level. Generally a 1-day counting stage of the residue of a 3 L evaporated water sample leads to a statistical counting uncertainty less than 10% for most of the radionuclides.
TABLE 1: 229Th-225Ra Spike Activities (mBq‚L-1) Determined in Unfiltered and Filtered Water Samplesa
unfiltered filtered
229Th 100 keV
229Th 194 keV
225Ra 40 keV
26 ( 2 27 ( 2
27 ( 1 26 ( 1
28 ( 1 25 ( 2
a Uncertainty corresponds to combined standard uncertainties. Introduced activity was 28 mBq‚L-1.
226Ra was analyzed by measuring its granddaughters: 214Pb (295, 352 keV) and 214Bi (609 keV) 20 days after evaporation to ensure secular equilibrium. 228Ra was analyzed by measuring its daughter 228Ac (338, 911, 968 keV). A second measurement was conducted 3 months later to determine the activity of 238U by measuring its short half-life daughter 234 Th (T1/2 ) 24 d, 63 and 92 keV). Uranium can also be radiochemicaly purified and analyzed by R spectrometry of the residue of evaporation. 229Th spike activity was determined on both 100 and 193.6 keV γ rays, while 225Ra spike activity was determined at 40 keV. Atomic Absorption Spectrometry. Pb and Be, added before evaporation as carriers (3.3 and 6.7 mg‚L-1, respectively), were used to determine evaporation yield of these two elements. They both were determined by Flameless Atomic Absorption in the residue of evaporation after dissolution.
Results and Discussion Validation. Results of the validation experiment are presented in Table 1 (radioactive spike) and Table 2 (conventional methods). Gamma spectrometry of the samples spiked with 229Th and 225Ra leads to activities between 25 ( 2 and 28 ( 1 mBq‚L-1. These results are in good agreement with the activity of 28 mBq‚L-1 introduced before evaporation. Therefore an evaporation yield better than 90% can be assumed for Ra and Th. Usual determinations of 238U by R spectrometry (20, 27) and of 226Ra,228Ra by coprecipitation with BaSO4 followed by γ spectrometry (28) conduct activities in agreement with those obtained by γ spectrometry of the evaporated sample (Table 2). It can therefore be assumed that the evaporation is quantitative for U and Ra. For Pb and Be, Flameless Atomic Absorption Analysis also shows a very good evaporation yield. Similar validations conducted on six additional samples of water from different origins confirmed a recovery close to 100% for U, Th, Ra, Pb, and Be. Therefore, the evaporation method presented for determination of natural radionuclides in freshwaters by multielement γ spectrometry can be considered as quantitative for U, Th, Ra, Pb, and Be. Advantages/Disavantages of the Method. Determination of γ emitters in diluted liquid samples after concentration by evaporation is a fast and simple method. Compared to previous methods, this analytical procedure spares time and manpower. Such a method is then particularly adapted for the investigation of some U-Th series radionuclides (238U, 234-228Th, 226-228Ra, 210Pb) as well as of atmospheric inputs (210Pb and 7Be) in continental waters. Actually, health-related surveillance of natural and anthropogenic radioactivity increases (is going up), and a mandated method of preparation and measurement is required (16, 30). The use of big, low background, well-type Ge detectors settled in an underground low level laboratory leads to a sufficient precision with 2 L of common freshwater sample. With classical low background, well-type Ge detectors in standard laboratory, the method can be applied for measurements of low activities by increasing the sample initial volume. For example, the measurement of 5 L of water containing 10 mBq‚L-1 of 226Ra would be achieved with 10% error (2σ) after VOL. 37, NO. 21, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4991
TABLE 2: 238U, 226Ra, and 228Ra Contents (mBq‚L-1) in Unfiltered and 0.45 µm Filtered Water Samples Determined by Conventional Methods (r Spectrometry for U and γ Spectrometry on a BaSO4 Precipitate) and by γ Spectrometry on Evaporated Samplesa 238U
conventional unfiltered particles (γ) filtered
226Ra
evaporation
12.5 ( 0.6 12 ( 1 1.3 ( 0.3 10.3 ( 1.3 9(1
228Ra
conventional
evaporation
conventional
evaporation
3.6 ( 0.3
3.2 ( 0.1
1.9 ( 0.2
2.2 ( 0.2
2.3 ( 0.2
1.3 ( 0.1
2.1 ( 0.2
0.7 ( 0.1
0.8 ( 0.1
1.3 ( 0.2
a
Particles were also analyzed by γ spectrometry in order to compare the sum of particulate and dissolved activities to unfiltered water activity and then to detect major losses or contamination.
TABLE 3: Activities (mBq‚L-1) Measured by γ Spectrometry on Evaporated Lake Water Samples and Filterable Particles site A
particles filtered sum unfiltered particles filtered sum unfiltered particles filtered sum unfiltered
B
C
238U
234Th
226Ra
210Pb
228Ra
228Th
7Be
3.9 ( 0.3 8.2 ( 0.2 12.1 ( 0.4 11.3 ( 0.3 168 ( 10 80 ( 6 248 ( 12 256 ( 15 16 ( 1 24 ( 3 40 ( 3 41 ( 4
1.0 ( 0.2 8(3 9(3 11 ( 3 152 ( 8 34 ( 6 186 ( 10 186 ( 19 15 ( 1 17 ( 5 32 ( 5 33 ( 6
1.8 ( 0.1 15.1 ( 2.0 16.9 ( 2.0 16.5 ( 0.5 85 ( 4 63 ( 6 148 ( 7 162 ( 15 4.0 ( 0.5 30 ( 3 34 ( 3 29 ( 2
13 ( 1 23 ( 2 36 ( 2 34 ( 3 53 ( 3 15 ( 3 68 ( 4 63 ( 6 7.4 ( 0.7 5(2 12 ( 2 18 ( 4
0.2 ( 0.1 0.8 ( 0.2 1.0 ( 0.2 1.6 ( 0.3 2.0 ( 0.4 2.5 ( 0.4 4.5 ( 0.6 5.8 ( 0.8 0.32 ( 0.05 1.7 ( 0.3 2.0 ( 0.3 1.2 ( 0.4
0.6 ( 0.1 0.3 ( 0.1 0.9 ( 0.1 0.9 ( 0.2 1.9 ( 0.2 1.5 ( 0.3 3.4 ( 0.7 3.3 ( 0.6 0.27 ( 0.04 1.1 ( 0.2 1.3 ( 0.2 1.5 ( 0.3
5(1 16 ( 2 21 ( 2 23 ( 2 4.0 ( 0.6 14 ( 3 18 ( 3 19 ( 3 4.4 ( 0.3 17 ( 2 21 ( 2 20 ( 2
FIGURE 2. A 1-day counting stage spectrum obtained from 3 L of rainwater. Square root of number of counts versus energy (0.336 keV/channel); the square root scale enhances the visibility. a 24-h counting stage. When the sample is processed in a short time (a few days) after sampling, radioactive equilibrium between short half-lived γ emitter nuclides (234Th and 222Rn decay products) and the long half-lived nuclide parents (238U and 226Ra, respectively) is not reached. Therefore, the date of the measurement must be selected in regard to the radionuclides investigated. When disequilibrium (234Th/238U and 222Rn/226Ra) is studied, two counting stages, separated by an appropriated period, are required. Application of the Method. The γ spectrum of the evaporated rainwater sample is presented in Figure 2; only 210Pb and 7Be are observed. Corresponding activities are 66 ( 3 mBq‚L-1 for 210Pb and 1163 ( 19 mBq‚L-1 for 7Be. 210 Pb is delivered in watershed by erosion and precipitations. 7Be originates only from the atmosphere. Turekian et al. (31) suggested that 7Be and 210Pb are two distinct tracers, with 7Be tracing materials of a stratospheric origin, while 210Pb traces materials of a tropospheric origin. Todd et al. (32) investigated the atmospheric depositional characteristics of 7Be and 210Pb and concluded that the physical processes of washout and deposition overwhelm the differences in the chemistries and input functions of these radionuclides. Therefore, in rainwater, 7Be and 210Pb concentrations are correlated, and, therefore, 7Be activities can be used to trace 4992
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 21, 2003
the amount of 210Pb delivered by precipitations in lakes and streams. At the three locations A, B, and C (Figure 1), γ spectrometry was performed on particles and evaporated samples (unfiltered and filtered water) for 238U, 228-234Th, 226-228Ra, 210Pb, and 7Be activities determination. Results are presented in Table 3. Within the uncertainty range, mass balance calculations (sum of activities measured in filters and filtered water) are consistent with activities determined in unfiltered water. The sampling point A is located to the east of the input of mine water. Since the direction of the flow is E-W, it can be assumed that it is unaffected by the disused mine water. 234Th and 238U are almost at secular equilibrium, and the activity ratio 226Ra/238U is greater than unity. Higher activities of the radionuclides and changes in these ratios encountered downstream at site B reflect the influence of the mine or rainwater input. 7Be activity in unfiltered samples is the same on the three sites: 20 mBq‚L-1. The corresponding atmospheric 210Pb activity is about 1 mBq‚L-1 which is negligible. Consequently, U-series long-lived radionuclides inputs from rain can be considered as insignificant. Higher activities and changes of activity ratios are therefore likely to reflect the influence of the mine. The amount of radionuclides to the brook is different between U, Th, Ra, Pb, and Be. The origin of these differences must lie in the solubility of elements. During chemical weatering, Ra is more mobile than Th (1). 238U, 234Th, and 210Pb seem to be delivered mainly associated with particles (Table 3). The secular equilibrium between activities of 228Th and 228Ra (daughters of 232Th) observed in particles suggests that these radionuclides are tightly bound into detritial material which is very resistant to chemical weathering. The increase of activities of these two radionuclides in particles downstream the input of water from disused mines confirms the supply of particles larger than 0.45 µm. The association of Th (and U) with colloids in the 0.45 µm filtered fraction has been reported by several authors (1, 2, 33) and can be inferred to justify the 228Th activities determined in this soluble fraction. For 226Ra, a different behavior is observed: inputs associated with particles are almost equal to the ones in the less than 0.45 µm fraction. This indicates a greater solubility of Ra than U, Th, and Pb
which can also be deduced from the greater activity ratios (226Ra/238U, 226Ra/234Th, and 226Ra/210Pb) in filtered fraction compared with those from particles. The higher solubility of Ra compared with U, Th, and Pb has been mentioned in saline environments but, in freshwater, U is usually considered to be more soluble than Ra, Pb, and Th (1). Such a discrepancy with previous studies remains unexplained. It might result in the low concentration of SO4 or Ba (5). A decrease in the activities is observed downstream in the second lake (site C). When the stream enters the lake, the velocity of the current decreases drastically which amplifies the sedimentation processes. Radioelements associated with particles are then exported to the sediment. This is confirmed by the high activities measured in the sediments of this area of the lake: 18 600 Bq‚kg-1 of 238U, 1510 Bq‚kg-1 for 226Ra, and 3144 Bq‚kg-1 for 210Pb (34).
Acknowledgments We are grateful to C. Riccio for technical support at the LSM. We also thank the “Mairie de Limoges” and the “Conseil Ge´ne´ral de la Haute-Vienne” for help during the field work. We are grateful to Dr. P. van Beek for his helpful comments on an earlier draft of this manuscript. The comments of the reviewers were much appreciated.
Literature Cited (1) Ivanovich, M.; Harmon, R. S. Uranium-series Disequilibrium: Applications to Earth, Marine, and Environmental Sciences, 2nd ed.; Clarendon Press: Oxford, 1992. (2) Porcelli, D.; Andersson, P. S.; Wasserbourg, G. J.; Ingri, J.; Baskaran, M. Geochim. Cosmochim. Acta 1997, 61, 4095. (3) Andersson, P. S.; Porcelli, D.; Wasserburg, G. J.; Ingri, J. Geochim. Cosmochim. Acta 1998, 62(3), 385. (4) Feng, H.; Cochran, J. K.; Hirschberg, D. J. Geochim. Cosmochim. Acta 1999, 63(17), 2487. (5) Porcelli, D.; Andersson, P. S.; Baskaran, M.; Wasserburg, G. J. Geochim. Cosmochim. Acta 2001, 65(15), 2439. (6) Gonzalez-Labajo, J.; Bolivar, J. P.; Garcia-Tenorio, R. Rad. Phys. Chem. 2001, 61, 643. (7) Lozano, J. C.; Blanco Rodriguez, P.; Vera Tome´, F. J. Environ. Radioact. 2002, 63, 153. (8) Thomson, J.; Dyer, F. M.; Croudace, I. W. Geochim. Cosmochim. Acta 2002, 66(6), 1011. (9) Dominik, J.; Schuler, Ch.; Santschi, P. H. Earth Planet. Sci. Lett. 1989, 93, 345. (10) Dominik, J.; Burrus, D.; Vernet, J. P. Earth Planet. Sci. Lett. 1987, 84, 165.
(11) Wallbrink, P. J.; Murray, A. S. Water Resour. Res. 1996, 32(2), 467. (12) Trautmannsheimer, M.; Schramel, P.; Winkler, R.; Bunzl, K. Environ. Sci. Technol. 1998, 32, 238. (13) Dowdall, M.; O’Dea, J. Radiochim. Acta 1999, 87, 109. (14) Ko¨hler, M.; Gleisberg, B.; Niese, S. Appl. Radiat. Isot. 2000, 53, 203. (15) Waples, J. T.; Orl; ini, K. A.; Weckerly, K. M.; Edgington, D. N.; Klump, J. V. Mar. Chem. 2003, 80, 265. (16) Semkow, T. M.; Parekh, P. P.; Schwenker, C. D.; Khan, A. J.; Bari, A.; Colaresi, J. F.; Tench, O. K.; David, G.; Guryn, W. Appl. Radiat. Isot. 2002, 213. (17) Schmidt, S.; Reyss, J. L. Radiochim. Acta 2000, 88, 459. (18) Santschi, P. H.; Li, Y. H.; Bell, J. Earth Planet. Sci. Lett. 1979, 45, 201. (19) Moore, W. S. J. Geophys. Res. 1984, 89, 2061. (20) Ku, T. L. Uranium series disequilibrium in deep sea sediments. Ph.D. Thesis, Colombia University, 1966. (21) Brown, L.; Stensl, G. J.; Klein, J.; Middleton, R. Geochim. Cosmochim. Acta 1989, 53, 135. (22) Olsen, C. R.; Larsen, I. L.; Lowry, P. D.; Cutshall, N. H.; Todd, J. F.; Wong, G. T. F.; Casey, W. H. J. Geophys. Res. 1985, 91, 896. (23) Cutshall, N. H.; Larsen, I. L.; Olsen, C. R. Nucl. Instrum. Methods 1983, 206, 309. (24) Palmer, M. R.; Edmond, J. M. Geochim. Cosmochim. Acta 1993, 57, 4947. (25) Eyrolle, F. La fraction colloı¨dale organique dans les processus de transport des me´taux dans les eaux de surface: application aux syste`mes d’alte´ration en milieu tropical. Thesis University of Grenoble I France, 1994. (26) Norme AFNOR: NF M 60-801 Dec 1997. (27) Hakam, O. K.; Choukri, A.; Reyss, J. L.; Lferde, M. Radiochim. Acta 2000, 88(1), 55. (28) Schmidt, S.; Reyss, J. L. J. Geophys. Res. 1996, 101, 3589. (29) Reyss, J.-L.; Schmidt, S.; Legeleux, F.; Bonte´, Ph. Nucl. Instrum. Methods 1995, 357, 391. (30) Decret 2001-1220, 20 dec 2001 relatif aux eaux destine´es a` la consommation humaine a` l’exclusion des eaux mine´rales. J. O. 22 dec 2001, 20381. (31) Turekian, K. K.; Benninger, L. K.; Dion, E. P. J. Geophys. Res. 1983, 88, 5411. (32) Todd, J. F.; Wong, G. T. F. J. Geophys. Res. 1989, 94(8), 11106. (33) Honeyman, B. D.; Santschi, P. H. J. Mar. Res. 1989, 47, 951. (34) Royer, A.; Reyss, J. L.; Decossas, J. L. Radioprotection 2002, 31(C1), 69.
Received for review April 11, 2003. Revised manuscript received August 2, 2003. Accepted August 18, 2003. ES034333I
VOL. 37, NO. 21, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4993
