Radioactive impact of phosphate ore processing in a wet marshland in southwestern Spain

Radioactive Impact of Phosphate Ore Processing in a Wet Marshland in Southwestern Spain A. Martínez-Aguirre & M. García-León Facultad de Física, Universidad de Sevilla, Apdo. 1065, 41080 Sevilla, Spain ABSTRACT In this paper an extensive study of the presence of natural radionuclides in a marshland area located in the vicinity of a phosphoric acid production complex in southwestern Spain is presented. This marsh is a natural reservation where significan! biological activity takes place. The marsh is highly affected by the infl.ow of water from the Odie/ river which is enhanced with natural radionuclides from industries which use phosphate rocks as raw material for fertilizer production. Sediment samples, collected from the intertidal zone, showed that wastes from the phosphate industries have resulted in enhanced radioactivities in the marsh. INTRODUCTION Studies and surveys of natural environmental radiation and radioactivity are of great importance and interest for health physics as well as many other disciplines. The progressive development of the nuclear industry and other contaminating technologies and the widespread and everincreasing use of radiation and radioactive isotopes requires the evaluation of the background of natural radionuclides and radioactivity in order to detect anthropogenic contamination, assess its impact, and implement appropriate countermeasures for population and environmental protection. The radionuclides which contribute most to environmental radiation are the members of the natural radioactive series, 238U and 232Th, and 4° K. They are present in rocks and minerals in concentrations that can vary by severa] orders of magnitude, depending on the geochemical processes to which the material they carne from have been subjected. Any soil would have a concentration of natural radionuclides that is deter mined by the concentration in the parent material and by the physico chemical phenomena associated with its weathering. Thus, soils on top of sedimentary materials present diverse concentrations of these elements depending on those present in the eroded substrate and on the chemical and physical phenomena associated with the sedimentation. Phosphate ores contain important amounts of natural radioactive elements, specially 238U and descendants, in concentrations that depend on its geographical and geological origin, but which can reach up to 300 ppm for U. Sedimentary and igneous phosphate ores are used as raw material for the production of phosphoric acid, and, consequently, fertili zers for agricultura} purposes. Large-scale production of phosphoric acid results in the redistribution of huge amounts of natural radioactivity. Depending on the technology used to produce fertilizer or phosphoric acid, the distribution of radioactive elements among the various products would be different (Roessler et al., 1979). Sometimes fertilizers and a common by-product, phosphogypsum, contain such high activities of natural radioactive elements that the resulting radiological impact should be considered carefully (Paul et al., 1984; Martínez-Aguirre et al., 1994; Koster et al., 1992; Períañez & García-León, 1993). The radioactive impact that the industrial complex, which includes two fertilizer plants, located at the south of Spain (Huelva), has over the Odiel and Tinto river basins has been widely studied (Martínez-Aguirre & García-León, 1991; Martínez-Aguirre et al., 1994; Periañez & García León, 1993). Part of the solid wastes (phosphogypsum) are directly released into the Odiel. Moreover, acid liquid wastes (process waters) resulting from the phosphoric acid production, with high concentration of natural radionuclides in solution, are also released into the Odie! river channel. These releases have affected the concentration level of natural radionuclides in water and bottom sediment of the Odiel river (Martínez Aguirre et al., 1994). The factories are located in an estuarine system highly aff ected by the tides. The movement of the water results in a wide distribution of the contamination upstream and downstream the indus tries. In this paper we present the results obtained in the study of the presence of natural radionuclides in soils from the Odiel marshland area. The marsh is located at the western margin of the Odiel river, in front of the phosphoric acid production complex. It is highly aff ected by tidal move ments of the Odiel riverwaters and, consequently, its formation regime should be considered as sedimentary. Significant biological activity takes place at the marsh which is a natural reservation. Four small islands, crossed by four channel form the marsh. We will show that the radio activity released by the phosphate fertilizer factories wasting activities reaches the marsh and describe how such transport is produced. SAMPLES ANO EXPERIMENTAL METHOOS A total of 13 and 23 surface soil samples were taken from the Odiel marsh located at the western margin of the Odiel river opposite to the fertilizer factories, during February and November 1993, respectively. The samples were collected during the low tide from areas covered with water during the high tide. The density and organic matter content were determined for each soil sample. Approximately 1 g of dried and powdered soil sample were spiked with well known 208Po activities. After digestion of the sample with HN03 and agua regia, the residue is dissolved in 8M HN03. A measured volume (5 ml) of TBP (tributilphosphate) was then added to the nitric solution. After shaking for about 5 min, Po is retained in the HN03 8M phase whereas the rest of radionuclides move to the TBP phase. Once both phases are separated, the solution which contain the Po is evaporated to dryness and self-deposited (Morón et al., 1988) onto 2·2cm diameter silver planchets. Po was finally measured by cx-spectrometry with surface barrier or ion implanted Si detectors. Activities of 210Pb, 226Ra, 228Ra and 4°K were measured by y-spectro metry by using a reverse-electrode coaxial Ge detector. Lead-210 was determined by its y emission of 46·5 keV, 226Ra by its emission of 186·2 keV, after the contribution of the y emission of 235U to this peak was subtracted. Radium-228 was measured via the 991 ·07 keV y emission of 228 Ac, once secular equilibríum was achieved and, finally, 4°K was measured by its own emission of 1460·8 keV. The efficiency curve was determined by adding well-known activities of 152Eu and 133Ba to two different soil samples which had similar densities than most of the rest of soil samples. They were then air dried and mixed until correctly homogenized. Coincidence decay effects were corrected by comparing the count rate with that from a point source of the same radionuclides {152Eu and 133Ba). RESULTS ANO OISCUSSION Table 1 gives the radioactive concentrations in mBq/g (dry weight) measured for 210Po, 210Pb, 226Ra, 228Ra and 4°K together with the dry TABLE l Activity Concentrations in mBq/g ofNatural Radionuclides in Soil Samples Collected at the Odie! Marsh in February* and November ¡993t Code D (g/cm3) % organ 210Po 2/0pb 226Ra228Ra 4oK Ml 0·64 9.05 199 ± 10 236 ± 16 166 ± 25 56·0 ± 5.9 562 ± 27 M2 0·58 9.27 778 ± 41 820 ± 27 711 ±45 62·5 ± 7.1 439 ± 34 M3 0-60 6-36 643 ± 28 795 ± 23 681 ± 36 84-4 ± 7-3 351 ± 32 M6* 0·38 23-6 264 ± 12 454 ± 27 321 ± 40 24·6± 11·1 466 ± 44 M6 0·65 14·2 580 ± 26 673 ± 16 598 ± 25 53.7 ± 4-4 470 ± 24 M5 0-56 14-7 222 ± 11 156 ± 14 161 ± 24 42-8 ± 4-6 511 ± 26 N7* 0-36 20·9 290 ± 31 424 ± 29 280 ± 43 31·3 ± 11·5 504 ± 48 M7 0•73 14·2 213 ± 10 287 ± 12 213 ± 21 48·8 ± 4·0 556 ± 22 M9* 0-43 18·7 341 ± 19 514 ± 36 326 ± 56 24•7 ± 12·2 318 ± 54 M9 0-63 10·5 141 ± 7 161 ± 12 149 ± 20 34·0 ± 3.4 549 ± 20 M8 0-74 15·2 130 ± 7 212 ± 14 130 ± 21 35·8 ± 4·1 520 ± 23 Mil* 0·49 18-5 276 ± 14 435 ± 30 304 ± 39 32·2 ± I0-7 448 ± 41 Mll 0·96 7.9 64·6 ± 4.0 76·2 ± 12·3 49·0 ± 18·6 28·2 ± 3-6 589 ± 21 MIO 0-84 8-6 188 ± 10 161 ± 13 149 ± 22 31·9 ± 4·0 566 ± 24 Ml2* 0·58 12·9 25·1 ± 1.7 37-4 ± 9.1 26-4 ± 14·8 34·1 ± 2·9 615 ± 18 Ml2 0·81 12-8 27·9 ± 2-0 24·8 ± 12·3 46·0 ± 22.3 34.3 ± 4.2 538 ± 23 M4 0-71 13.1 210 ± 11 233 ± 13 242 ± 23 52·0 ± 4.3 581 ± 23 Ml9 0-728-6 233 ± 11 268 ± 15 245 ± 26 57-4±5·1 475 ± 25 Ml3 0-59 15·6 565 ± 23 681 ± 23 673 ± 39 57-4 ± 6·5 370 ± 34 Ml4* 0·70 9-4 108 ± 6 160 ± 25 143 ± 35 39.5 ± 9.7 490 ± 39 Ml4 1·33 39 41-0±3·2 18-4± 10·1 30-4 ± 16-5 27·9±3·1 648 ± 20 MIS* 0-54 13.7 367 ± 34 658 ± 30 505 ± 44 41·5 ± 9.9 289 ± 40 MIS 0·63 14· l441 ± 20 589 ± 21 484 ± 32 56·5 ± 6-6 359 ± 32 Ml6* 0-49 10-7 117 ± 7 137 ± 18 65-4 ± 25-4 34.7 ± 7.9 492 ± 29 Ml 6 ¡ .¡5 4.5 131 ± 7 162 ± 11 134 ± 16 29·1 ± 3-2 338 ± 17 TABLE 1-contd. Code D (g/cm3) % organ 210Po2/0pb226Ra22sRa4oK M22* 0-40 15-6 169 ± 9 236 ± 25 136 ± 36 33-6 ± 10·2 443 ±40 M22 0-93 9.5 44·2 ± 2-5 26-6 ± 11-1 31-7±17-2 33.5 ± 3-2 654 ± 20 M23* 0-39 13-3 124±8 205 ± 23 76-6 ± 35.9 28-6 ± 9.5 545 ± 39 M23 0-67 6-1 194±9 121 ± 13 97.3 ± 20-7 30-8 ± 3·8 445 ± 22 M24* 0-63 11-3 44.9 ± 2-9 34-8 ± 6-8 33-2 ± 11-4 25-2 ± 2-0 604 ± 14 Ml8* 0-53 11-9 29·0 ± 1·8 34-0 ± 8-4 34·8 ± 13-7 25-7 ± 2·6 703 ± 16 Ml8 1-12 5-6 54.4 ± 3.7 51-9± 11-1 35.3 ± 17-6 21-3±3·1 401 ± 19 Ml7* 0-52 12-5 40-5 ± 2-3 53-2 ± 6-5 43.3 ± 10-5 28-6 ± 2-1 567 ± 13 M17 0-74 8-8 54.7 ± 3.9 80·6 ± 10-5 29·8 ± 17-8 32-0 ± 3-1 606 ± 18 M20 1-54 1 · l 16·2 ± 2-5 15·4± 10·1 21-3 ± 16-0 13-2 ± 2-9 287 ± 18 M21 1·54 2·1 37·2 ± 2-5 22-9 ± 10-9 20-3±17-7 16-0 ± 3.3 422 ± 20 1Density (dry weight) and organic content (%) in each soil sample are also given. "" � ����� ��� "' * • e:) ! i "D ;:: Ul e ¡¡; C') 1 m ! z ; �<.._./, � . . Fig. l. Map of the Odie! marsh area where the locations of the sampling stations in February and November 1993 are shown. bulk density and per cent of organic matter in each soil for both sampling campaigns. Location of each sampling station is displayed in Fig. l for both sampling campaigns. Thirteen soil samples were collected at the end of February 1993. lt is clear that the range of radioactivity concentrations for all the 238U decay series descendants is wide. Thus, concentrations below 45, 53 and 43 mBq/g for 210Po, 21ºPb and 226Ra, respectively, are found in the further part of the marsh, the west side of the Saltes island and the southwest of the Enmedio island. Higher levels, from 100 to 170, 130 to 240 and 60 to 150 mBq/g for the same radionuclides, respectively were measured at the northwest of Enmedio island and east side of the Saltés island. These samples are not clearly enhanced but the levels are higher than those above. Besides that it appears a set of samples where the influence of the industries is clearly seen without further data analysis. Thus, along the Mojarrera channel concentrations higher than 250, 400 and 250 mBq/g are found for 210Po, 210Pb and 226Ra. The pattern distribution of activity concentrations across the marsh must be logically governed by the tidal intrusion of waters from the Odiel river which transport radioactivity released by the faetones. Probably, during high tides, the contaminated Odiel waters invade the marsh through the north of the Bacuta island. Most of this water flows into the Mojarrera channel which is the main sink of such contamination. Besides that, part of this water would flow into the Estero de Cajarias. That is the reason why the northern part of the Enmedio island is clearly enhanced for 210Po, 210Pb and 226Ra. Thus the activity decreases all along the El Chate channel and the Punta Umbría river. The eastern part of the Saltés island is affected by the radioactivity released from the factories. lt is transported inland by the waters at low tide. Differences in activity concentrations between stations at the east of the island must therefore be related to the differences in organic content (see Table 1). A more extensive study was carried out during November 1993. During this campaign, 20 soil samples were collected from different (but close to those of the February study) sampling stations in the Odiel marsh. Three additional soil samples were taken from the Northern marsh in order to wider our study (see Fig. 1). Results are also presented in Table 1, together with the density and per cent of organic matter. Very similar conclusions to those of February can be derived from the analysis of 238U descendant radioactivity data. Thus, it seems clear again that the flow of Odiel waters through the Mojarrera channel is responsible of the natural radioactivity enhancement found at the Odiel marsh. Thus, background levels, below 70, 80 and 50 mBq/g for 210Pa, 21ºPg and 226Ra were found along El Chate channel and the Punta Umbría river. Up to 600, 700 and 800mBq/g activity concentrations for the same radionuclides were found in different stations across the Odiel marsh. It is also true that the general activity levels for 210Pa, 210Pb or 226Ra are lower than those found in February. This is not strange if one takes into account the complexity of the dynamics of the system which is being studied. Again, it is important to notice that the sampling stations were not exactly the same. In fact, the majority of samples collected in November were located at places which are covered by water for less time during high tides than those of February. The cases of samples 20 and 21 deserve sorne special comments. Being situated clase to the source input of radioactivity, but on the other side of the river, the radioactivity levels found are certainly low. Two reasons could explain such effect. Firstly, their density and organic content (see Table 1) reflect the presence of large amount of sand particles: the direct consequence is a lowering of radioactivity concentrations. Secondly, it is known (Periañez et al., 1994) that the transversal velocity across the Odiel river for the dispersion of a given pollutant is, at least, one order of magnitude lower than the longitudinal velocity along the Odiel river. This meaning that once an input of radioactivity has been introduced into the river waters it moves faster in the longitudinal direction, reaching the opposite shoreline highly diluted. The differences in density and organic matter could also explain the special activity pattern found at the east side of the Saltés island for samples 14 to 16 (see Table 1). Very high levels of radioactivity were found at the Northern marsh (see Fig. 1), especially for the case of samples N2 and N3. lt is apparent that this place is also affected by the phosphoric acid factories. In fact, it has been found that the marsh is an important sink for many other pollutants (Martín et al., 1995). This finding requires further investigation, although it seems clear that the reason for such high radioactivity levels should be related to tidal transport by the Odiel waters. This agrees with the high activity concentrations found in sample N4, located at the Odiel river margin upstream the factories. The 228Ra radioactivity concentrations are fairly uniform across the studied area in February. An average value of 31-3 ± 2·4 mBq/g could revea! that the area under study is not contaminated by 232Th descen dants. This is not surprising since it has been found that the Odiel waters are little contaminated by 232Th and daughters. Concentrations of 228Ra in samples collected in November showed more dispersed values than those of February, ranging from about 13 to 85mBq/g. From this viewpoint, the behaviour of the 228Ra/226Ra activity ratio for instance, (see column 4 of Table 2), could provide information on the pattern of contamination. However, if we exclude the northern marsh data, the average value of 37 ± 13 mBq/g, agrees with that found in February. In general, higher activity concentrations of 228Ra are found in those samples with high activity concentrations of 238U descendants. Indeed, the ratio increases for low 226Ra activities and decreases when the 226Ra radioactivity concentration increases, for both samplings, following the geographical trend discussed above (see Fig. 2). A poten tial fitting of the curve for each sampling, separately, gives similar results R = 8·75 A-0·71 R being the 228Ra/226Ra activity ratio and A the 226Ra activity in mBq/g. The regression coefficient is -0·909. The relative distribution of 4°K along the marsh also gives interesting information of the enhancement of the studied area. The 4°K radioactivity concentration ranges from 290 to 700 mBq/g, which agrees well with the world 4°K background range for soils (UNSCEAR, 1988). Nevertheless, it is clear that the 4°K activity decreases when the activities from any 238U descendants increase. For instance, in Fig. 3, we show this trend for the case of the pair 4°K-21ºPb. The only reason for such interesting behaviour could reside in the fact that the phosphate wastes, as the phosphate ores, are very poor in 4°K. For instance, Hussein (1994) obtained 19 ± 2·1 mBq/ g for 4°K in phosphate ores. Consequently the relative lowering of 4°K could mark the intrusion of contaminated material. A similar effect, but on sediment samples, was known in Gallardo et al. (1993). The relative distribution of 4°K is quite similar for both sampling campaign, as can be seen in Fig. 3, where we have excluded those samples with very different density values. A fitting of this line gives, with a regression coefficient of -0·8035. It is interesting to discuss sorne isotopic ratios. In Table 2, column 3, we present the 21 ºPb/226Ra radioactivity concentration ratio for both sampling campaigns. This ratio is higher than unity revealing an excess of 210Pb across the marsh related to 226Ra. The 210Pb/226Ra activity ratio in the Odie! waters (Martínez-Aguirre & García-León, 1995) shows a differ ent trend. In fact, values from 0-4 to O· 7 can be found close to the source input. The fact that in the marsh, the activity ratio is higher than unity must be related to the higher mobility of 226Ra in nature. Thus, the 210Pb