- Email: [email protected]
The analysis of naturally-occurring radionuclides from uranium and thorium decay series in table mineral waters
The Science of the Total Environment, 130/131 (1993) 253-259 Elsevier Science Publishers B.V., .Amsterdam
253
The analysis of naturally-occurring radionuciides from uranium and thorium decay series in table mineral waters T.C. Aellen*, O. Umbricht and W. Goerlich Radiation Hygiene Division, Paul Scherrer Instituw, CH-5232 VtTiigen PSI, Switzerland ABSTRACT This project required the highly sensitive analysis of low-level alpha-and beta-emitters naturally occurring in table mineral water sold on the Swiss market. These radionuclides occur in the three major decay series-uranium-238, uranium-235, and thorium-232. The radionudides analysed were 238U, 235U, 234U, 232Th, 230Th' 228Th' 210po' 210pb' and 226Ra. Many other radionuclides were determinable as a result of their equilibrium with an analysed nuclide. Efficient, element specific separation techniques were developed, allowing for the spectral analysis of each element without interference from other radioactive elements. Radioactive tracers, 232U, 23°Th, and 2°9p0, were necessary to determine the percentage yield. These yie)ds often varied greatly between different mineral waters, especially for thorium, ranging from 30 to 100%. Uranium, thorium and polonium isotopes could be directly analysed for by alpha-spectrometry. 226Rawas determined through the ingrowth of its daughter 222Rn by liquid scintillation counting. From the sample~ remaining after 21°po removal, the isotope's re-ingrowth from 21°pb determined the original 21°pb content. Limits of detection ranged from 0.1 to 2.0 mBq/l. The following contents were determined 234U + 23SU30-720 mBq/l; 232Th + 23°Th < 1-5 mBq/I; 22STh 2-40 mBq/l; 226Ra 5-370 mBq/l; 21°po 1-90 mBq/l; 2'°pb 1-90 mBq/I Key words: mineral water; radionuclide; uranium; thorium; radium
INTRODUCTION
In Switzerland up to now, no data about the occurrence of natural radionuclides in mineral and drinking water except for uranium were available. As a result of the discovery of relatively high concentrations of uranium in some waters, and in view of the assessment of these findings, the question about the presence of other natural radionuclides was raised (E. Bosshard and B. Zimmerli, Uranium in the diet, unpublished). For comparison purposes their limit of detection was set so that a resulting theoretical cancer risk Present address: Ciba-Geigy Ltd., 4002 Basel, Switzerland.
254
T.C. AELLEN ET AL.
of 10 -7, from life-long consumption of 1 litre of water per day, could be estimated. This extremely low detection limit required especially sensitive methods. Since the water samples were plentiful, and sample preparation was without serious difficulties, a sequential method was considered unnecessary, concentrating instead on more sensitive element-specific separations. The mineral content of the waters often played an important role in, for example, the formation of precipitates. As a result of the sometimes extreme differences in the ionic concentrations of the waters, the ionic equilibrium was often imbalanced thus hindering precipitate formation. In such cases the separation techniques required modification. MATERIALS AND METHODS
Most chemicals were purchased from Merck, W. Germany. Aliquat-336 was purchased from Serva, New York. Tracers: 232U Standard solution from Medipro AG, Switzerland; 2~°Th and 2°9po standard solutions from PTB (Physikalisch-Technische Bundesanstalt), W. Germany. Scintillation liquid was purchased from Instagel, Packard. Silver discs (28 mm diameter, 0.1 mm thickness) were purchased from Johnson Matthey and Brandenberger AG, Switzerland. Silic<~L~:i~ium diodes were used in the alpha-spectrometers and the liquid scintillation analyser was a Tri-carb 2000 from Canberra/Packard. Uranium method
The determinati~.-. 6f uranium in mineral water was carried out with the fcr~nafion of an ammoniacale precipitate followed by extraction of uranium and thorium with 6% aliquat-336 (triaprylmethylammoniumchloride) (personal communication from H.R. Bachli, Radioanalytical Group, PSI-Ost, CH-5234, ViUigen, Switzerland). Uranium was re-extracted from the aliquat with dernineralised water. A 6% solution of aliquat-336 in xylene was prepared. Each litre of this solution was washed three times with 250 ml of 4 M nitric acid, where vigorous shaking was required to convert the quaternary ammonium salt to the nitrate form. One litre of mineral water was acidified with 200 ml cone. nitric acid and 5 ml calcium phosphate (300 mg) were added. The volume was reduced by boiling to -.- 500 ml, and allowed to cool. 400 ml of cone. ammonium hydroxide was added and the resulting precipitate allowed to settle overnight. The supernatant was removed by suction, the precipitate centrifuged and finally dissolved in --- 10 ml nitric acid. The solution was evaporated to dryness with a little perhydrol. Twenty-five millilitres hydrochloric acid were required to
NATURALLY-OCCURRING U AND TH IN TABLE MINERAL WATERS
~5
dissolve the precipitate. Uranium was then extracted using 3 x 10 ml portions of 6% aliquat-336 solution. Re-extraction of uranium from aliquat required 3 x 10 ml portions of demineralised water. Mineralisation of any remaining organics was achieved by heating to dryness with conc. nitric acid and perhydrol until a white precipitate was obtained. The sample was prepared for electrolysis by dissolving it in 0.4 ml conc. sulphuric acid, heated until no further white fumes appeared, and then left to cool. Four millilitres of demineralised water were added, along with two drops of methyl-red indicator. Concentrated ammonium hydroxide solution and 2 N sulphuric acid were used to adjust the pH of the solution to 4. The electrolysis was carded out at 300 mA, for 4 h, with a platinum anode and a steel plate (circular with 28 mm diameter, 0.1 mm thickness) as the cathode. A second electrolysis was carried out after the re-setting of the pH with the 2N sulphuric acid solution. Counting of the steel plates was carded out in an alpha-spectrometer.
Thorium method The determination of thorium in water was carried out using the method descrirJed by M. Cospilo and L. Rigali [1] with the following modifications. The aliquat-336 extractions were carded out three times (3 x 30 ml), as were the 3 M hydrochloric acid extractions (3 x 50 ml). No washing of the aliquat-336 portions with 40 ml of 4 M nitric acid was necessary. On evaporating the acid extracts to dryness, perhydrol and conc. nitric acid were used to mineralise any organic material. The determination was carried out by preparing the sample for electrolysis in the same way used in the uranium method.
Radium analysis The determination of 226Ra ;Ln mineral water was carried out by measuring the ingrowth of its daughter 222Rn into n-hexane~ using a liquid scintillation counter (personal communication from H.R. Bachli, Radioanalytical Group, PSI-Ost, CH-5234, Villigen, Switzerland). To 1 i of water were added 5 ml conc. nitric acid and 1 ml barium chloride tracer (where 1 rnl contains 10 mg of Ba2+). Ammonium acetate (5 g) was added and the solution adjusted to pH 5 with conc. ammonium hydroxide. Ammonium sulphate (5 g) was added and the solution was heated at -80°C for 1 h, and allowed to cool overnight. The precipitate was filtered and then washed with 0.5% sulphuric acid. The filter was placed in a platinum cup and heated in an oven for one hour at 900°C. After cooling the paper was ashed with 2 ml (1:1) sulphuric acid and
256
T.C. AELLEN ET AL.
5 ml 40% hydrofluoric acid until no further sulphur dioxide vapours could be seen. After cooling, the solution was washed with 30 ml demineralised water into a centrifuge glass and left standing for 2-3 h. The precipitate was then centrifuged and washed twice with demineralised water. Five millilitres of 0.25 M ammoniacle EDTA solution and 5 ml demineralised water were used to dissolve the precipitate. The solution was placed into a glass scintillation bottle and the bottle filled with n-hexane so that no room for air bubbles existed. The sample was left - 2 0 days with occasional shaking before removing the n-hexane for analysis in a liquid scintillation counter. Polonium method The determination of polonium in water was carried out by autoelectrolysis, adapted from the method published by J.C. Laul eta|., [3]. The method describes the co-precipitation of iron(III), polonium and lead hydroxide. Polonium is separated from an acidic solution by autoelectrolysis onto one side of a silver disc. Several alterations had to be made. Since mineral water in general contains little or no iron(III) ion required for the formation of the iron(III) hydroxide precipitate, addition of 100 mg of the ion during the sample preparation was necessary. A specially designed teflon holder for the silver discs was developed (Fig. 1), so that only one side of the disc was free for electrolysis and that the disc could be vertically supported to prevent air bubble formation on its surface. The silver discs were measured in an alpha-spectrometer.
Lead analysis The existence of an equilibrium between 2~°pb and 2~°po was tested by
m
Fig. 1. Teflon holder developed side of the disc only.
to
vertically suspend a silver disc in solution, revealing one
NATURALLY-OCCURRING U AND TH IN TABLE MINERAL WATERS
257
allowing the aqueous samples used in the polonium analysis to stand long enough (at least one half-life of 21°po) for 2m°po to grow into the sample from 21°pb. The sample was then re-analysed for 21°po through a second autoelectrolysis in order to determine the original 21°pb content. RESULTS AND DISCUSSION
The range of results determined for 19 mineral waters in mBq 1-~ are shown below in Table 1. Uranium
A few mineral waters produced little or no ammoniacal precipitate so the solution had to be evaporated down to a volume of 500 ml in order to increase the ion concentration. During electrolysis the second electrolysis often produced a higher uranium yield than the first. No explanation has yet been found. 232U was used as tracer. 235U, although often present in the spectrum (Fig. 2), proved easier to calculate as 4.76% of the 238U content. Activity Ratios (A.R.) of 234U/238U were mostly approximate to 1 with few exceptions reaching as high as 4.6. Percentage yields were good, averaging 80%, and consistent. Thorium
Since analysis of 23°Th would be affected by the 23°Th tracer, percentage yields were determined separately. 232Th and 23°Th contents were very low level so that the tracer was unaffected by the waters actual content. The time required for equilibrium of 22STh and its daughters was 2-4 weeks, lncreas-
TABLE I Range of results in mBq I-' for nineteen mineral waters Isotope
Range of Results in mBq L-'
Uranium-234 and 238 Thorium-232 and 230 Thorium-228 Radium-226 Polonium-210 Lead-210
30-720 < 1-5 2-40 5-370 1-90 1-90
258
T.C. AELLEN ET AL.
U-234 U-238 U-232 Counts
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
Energy / MeV
Fig. 2. Alpha spectrum of uranium co~tent of mineral water with 2J2U tracer peak.
ing the sample volume to 2 litres affected reproducibility of the results and reduced the percentage yield. Ammonium ions are added at ~.hebeginning of sample preparation disturbing the equilibrium constant of those mineral waters with an already high NH4 + concentration, resulting in no precipitate formation. Addition of a few milligrammes of oxalic acid readjusted the equilibrium constant. Each mineral water produced a different percentage yield which was assumed to have resulted from differing mineral contents. Percentage yields ranged from 30-100%, i.e. complete recovery. Thorium's general lack of solubility in water explains the overall low thorium content of mineral water. A.R. of 22STlff232Th ratages from 0.05-80; 23°Th/234Ufrom 0.001 to 1. Both ratios require further examination. Radium
Percentage yield was determined by the weight of the final precipitate before dissolution in EDTA. The percentage yield often exceeded 100%, thought to be due to co-precipitation of calcium and strontium sulphate. An average yield of 80% was accepted. No major problems occurred as a result of differing mineral contents. 226Ra was occasionally present in large quantities. Since radium is extremely similar to calcium, it was thought that a high calcium content would indicate a high radium content. This was examined and could not be confirmed [2].
NATURALLY-OCCURRING U AND TH IN TABLE MINERAL WATERS
259
Polonium and lead
The 2°9p0 tracer peak did not interfere with that of 21°po, and therefore, tracer could be added to each sample. Only the length of time of autoelectrolysis affected the percentage yield. Four hours produced yields of 80--90%. Sample volume was increased ~o 2 litres with no effect on percentage yield, and indeed doubled results to well above the limit of detection. Polonium results could not be averaged due to occasional sudden increases in content. A very likely explanation is that 21°po and 21°pb are not in equilibrium at the time of analysis. Since the 2l°pb results are mostly higher than those for 2'°po it can be said that the mineral water on removal from source contained little or no 2l°po, instead it grew into the sample from 2~°pb. This is also suggested by the average length of time between removal from source and consumption (-- 3 months). Equilibrium is therefore almost always never reached. Ground rock is, of course, a very important factor in the natural radioactivity content of mineral water. However, the extreme values determined for some waters can perhaps be explained by a lack of homogeneity in the water when taken from its source. ACKNOWLEDGEMENTS
This project was proposed and financially supported by the Federal Office of Public Health, Division of Food Science. REFERENCES 1 2 3
M. Cospilo and L. Rigali, Determination of thorium in natural waters after extraction with aliquat-336. Anal. Chim. Acta. 106 (2) (1979) 385-388. O. Hogl, Die Mineral-und Heilquellen der Schweiz, Haupt, Berne, Switzerland. J.C.Laul et al., Analysis of natural radionuclides from uranium and thorium decay series in briny groundwaters. J. Radioanal. Nucl. Chem., I I0 (l)(1987) 101-112.
253
The analysis of naturally-occurring radionuciides from uranium and thorium decay series in table mineral waters T.C. Aellen*, O. Umbricht and W. Goerlich Radiation Hygiene Division, Paul Scherrer Instituw, CH-5232 VtTiigen PSI, Switzerland ABSTRACT This project required the highly sensitive analysis of low-level alpha-and beta-emitters naturally occurring in table mineral water sold on the Swiss market. These radionuclides occur in the three major decay series-uranium-238, uranium-235, and thorium-232. The radionudides analysed were 238U, 235U, 234U, 232Th, 230Th' 228Th' 210po' 210pb' and 226Ra. Many other radionuclides were determinable as a result of their equilibrium with an analysed nuclide. Efficient, element specific separation techniques were developed, allowing for the spectral analysis of each element without interference from other radioactive elements. Radioactive tracers, 232U, 23°Th, and 2°9p0, were necessary to determine the percentage yield. These yie)ds often varied greatly between different mineral waters, especially for thorium, ranging from 30 to 100%. Uranium, thorium and polonium isotopes could be directly analysed for by alpha-spectrometry. 226Rawas determined through the ingrowth of its daughter 222Rn by liquid scintillation counting. From the sample~ remaining after 21°po removal, the isotope's re-ingrowth from 21°pb determined the original 21°pb content. Limits of detection ranged from 0.1 to 2.0 mBq/l. The following contents were determined 234U + 23SU30-720 mBq/l; 232Th + 23°Th < 1-5 mBq/I; 22STh 2-40 mBq/l; 226Ra 5-370 mBq/l; 21°po 1-90 mBq/l; 2'°pb 1-90 mBq/I Key words: mineral water; radionuclide; uranium; thorium; radium
INTRODUCTION
In Switzerland up to now, no data about the occurrence of natural radionuclides in mineral and drinking water except for uranium were available. As a result of the discovery of relatively high concentrations of uranium in some waters, and in view of the assessment of these findings, the question about the presence of other natural radionuclides was raised (E. Bosshard and B. Zimmerli, Uranium in the diet, unpublished). For comparison purposes their limit of detection was set so that a resulting theoretical cancer risk Present address: Ciba-Geigy Ltd., 4002 Basel, Switzerland.
254
T.C. AELLEN ET AL.
of 10 -7, from life-long consumption of 1 litre of water per day, could be estimated. This extremely low detection limit required especially sensitive methods. Since the water samples were plentiful, and sample preparation was without serious difficulties, a sequential method was considered unnecessary, concentrating instead on more sensitive element-specific separations. The mineral content of the waters often played an important role in, for example, the formation of precipitates. As a result of the sometimes extreme differences in the ionic concentrations of the waters, the ionic equilibrium was often imbalanced thus hindering precipitate formation. In such cases the separation techniques required modification. MATERIALS AND METHODS
Most chemicals were purchased from Merck, W. Germany. Aliquat-336 was purchased from Serva, New York. Tracers: 232U Standard solution from Medipro AG, Switzerland; 2~°Th and 2°9po standard solutions from PTB (Physikalisch-Technische Bundesanstalt), W. Germany. Scintillation liquid was purchased from Instagel, Packard. Silver discs (28 mm diameter, 0.1 mm thickness) were purchased from Johnson Matthey and Brandenberger AG, Switzerland. Silic<~L~:i~ium diodes were used in the alpha-spectrometers and the liquid scintillation analyser was a Tri-carb 2000 from Canberra/Packard. Uranium method
The determinati~.-. 6f uranium in mineral water was carried out with the fcr~nafion of an ammoniacale precipitate followed by extraction of uranium and thorium with 6% aliquat-336 (triaprylmethylammoniumchloride) (personal communication from H.R. Bachli, Radioanalytical Group, PSI-Ost, CH-5234, ViUigen, Switzerland). Uranium was re-extracted from the aliquat with dernineralised water. A 6% solution of aliquat-336 in xylene was prepared. Each litre of this solution was washed three times with 250 ml of 4 M nitric acid, where vigorous shaking was required to convert the quaternary ammonium salt to the nitrate form. One litre of mineral water was acidified with 200 ml cone. nitric acid and 5 ml calcium phosphate (300 mg) were added. The volume was reduced by boiling to -.- 500 ml, and allowed to cool. 400 ml of cone. ammonium hydroxide was added and the resulting precipitate allowed to settle overnight. The supernatant was removed by suction, the precipitate centrifuged and finally dissolved in --- 10 ml nitric acid. The solution was evaporated to dryness with a little perhydrol. Twenty-five millilitres hydrochloric acid were required to
NATURALLY-OCCURRING U AND TH IN TABLE MINERAL WATERS
~5
dissolve the precipitate. Uranium was then extracted using 3 x 10 ml portions of 6% aliquat-336 solution. Re-extraction of uranium from aliquat required 3 x 10 ml portions of demineralised water. Mineralisation of any remaining organics was achieved by heating to dryness with conc. nitric acid and perhydrol until a white precipitate was obtained. The sample was prepared for electrolysis by dissolving it in 0.4 ml conc. sulphuric acid, heated until no further white fumes appeared, and then left to cool. Four millilitres of demineralised water were added, along with two drops of methyl-red indicator. Concentrated ammonium hydroxide solution and 2 N sulphuric acid were used to adjust the pH of the solution to 4. The electrolysis was carded out at 300 mA, for 4 h, with a platinum anode and a steel plate (circular with 28 mm diameter, 0.1 mm thickness) as the cathode. A second electrolysis was carried out after the re-setting of the pH with the 2N sulphuric acid solution. Counting of the steel plates was carded out in an alpha-spectrometer.
Thorium method The determination of thorium in water was carried out using the method descrirJed by M. Cospilo and L. Rigali [1] with the following modifications. The aliquat-336 extractions were carded out three times (3 x 30 ml), as were the 3 M hydrochloric acid extractions (3 x 50 ml). No washing of the aliquat-336 portions with 40 ml of 4 M nitric acid was necessary. On evaporating the acid extracts to dryness, perhydrol and conc. nitric acid were used to mineralise any organic material. The determination was carried out by preparing the sample for electrolysis in the same way used in the uranium method.
Radium analysis The determination of 226Ra ;Ln mineral water was carried out by measuring the ingrowth of its daughter 222Rn into n-hexane~ using a liquid scintillation counter (personal communication from H.R. Bachli, Radioanalytical Group, PSI-Ost, CH-5234, Villigen, Switzerland). To 1 i of water were added 5 ml conc. nitric acid and 1 ml barium chloride tracer (where 1 rnl contains 10 mg of Ba2+). Ammonium acetate (5 g) was added and the solution adjusted to pH 5 with conc. ammonium hydroxide. Ammonium sulphate (5 g) was added and the solution was heated at -80°C for 1 h, and allowed to cool overnight. The precipitate was filtered and then washed with 0.5% sulphuric acid. The filter was placed in a platinum cup and heated in an oven for one hour at 900°C. After cooling the paper was ashed with 2 ml (1:1) sulphuric acid and
256
T.C. AELLEN ET AL.
5 ml 40% hydrofluoric acid until no further sulphur dioxide vapours could be seen. After cooling, the solution was washed with 30 ml demineralised water into a centrifuge glass and left standing for 2-3 h. The precipitate was then centrifuged and washed twice with demineralised water. Five millilitres of 0.25 M ammoniacle EDTA solution and 5 ml demineralised water were used to dissolve the precipitate. The solution was placed into a glass scintillation bottle and the bottle filled with n-hexane so that no room for air bubbles existed. The sample was left - 2 0 days with occasional shaking before removing the n-hexane for analysis in a liquid scintillation counter. Polonium method The determination of polonium in water was carried out by autoelectrolysis, adapted from the method published by J.C. Laul eta|., [3]. The method describes the co-precipitation of iron(III), polonium and lead hydroxide. Polonium is separated from an acidic solution by autoelectrolysis onto one side of a silver disc. Several alterations had to be made. Since mineral water in general contains little or no iron(III) ion required for the formation of the iron(III) hydroxide precipitate, addition of 100 mg of the ion during the sample preparation was necessary. A specially designed teflon holder for the silver discs was developed (Fig. 1), so that only one side of the disc was free for electrolysis and that the disc could be vertically supported to prevent air bubble formation on its surface. The silver discs were measured in an alpha-spectrometer.
Lead analysis The existence of an equilibrium between 2~°pb and 2~°po was tested by
m
Fig. 1. Teflon holder developed side of the disc only.
to
vertically suspend a silver disc in solution, revealing one
NATURALLY-OCCURRING U AND TH IN TABLE MINERAL WATERS
257
allowing the aqueous samples used in the polonium analysis to stand long enough (at least one half-life of 21°po) for 2m°po to grow into the sample from 21°pb. The sample was then re-analysed for 21°po through a second autoelectrolysis in order to determine the original 21°pb content. RESULTS AND DISCUSSION
The range of results determined for 19 mineral waters in mBq 1-~ are shown below in Table 1. Uranium
A few mineral waters produced little or no ammoniacal precipitate so the solution had to be evaporated down to a volume of 500 ml in order to increase the ion concentration. During electrolysis the second electrolysis often produced a higher uranium yield than the first. No explanation has yet been found. 232U was used as tracer. 235U, although often present in the spectrum (Fig. 2), proved easier to calculate as 4.76% of the 238U content. Activity Ratios (A.R.) of 234U/238U were mostly approximate to 1 with few exceptions reaching as high as 4.6. Percentage yields were good, averaging 80%, and consistent. Thorium
Since analysis of 23°Th would be affected by the 23°Th tracer, percentage yields were determined separately. 232Th and 23°Th contents were very low level so that the tracer was unaffected by the waters actual content. The time required for equilibrium of 22STh and its daughters was 2-4 weeks, lncreas-
TABLE I Range of results in mBq I-' for nineteen mineral waters Isotope
Range of Results in mBq L-'
Uranium-234 and 238 Thorium-232 and 230 Thorium-228 Radium-226 Polonium-210 Lead-210
30-720 < 1-5 2-40 5-370 1-90 1-90
258
T.C. AELLEN ET AL.
U-234 U-238 U-232 Counts
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
Energy / MeV
Fig. 2. Alpha spectrum of uranium co~tent of mineral water with 2J2U tracer peak.
ing the sample volume to 2 litres affected reproducibility of the results and reduced the percentage yield. Ammonium ions are added at ~.hebeginning of sample preparation disturbing the equilibrium constant of those mineral waters with an already high NH4 + concentration, resulting in no precipitate formation. Addition of a few milligrammes of oxalic acid readjusted the equilibrium constant. Each mineral water produced a different percentage yield which was assumed to have resulted from differing mineral contents. Percentage yields ranged from 30-100%, i.e. complete recovery. Thorium's general lack of solubility in water explains the overall low thorium content of mineral water. A.R. of 22STlff232Th ratages from 0.05-80; 23°Th/234Ufrom 0.001 to 1. Both ratios require further examination. Radium
Percentage yield was determined by the weight of the final precipitate before dissolution in EDTA. The percentage yield often exceeded 100%, thought to be due to co-precipitation of calcium and strontium sulphate. An average yield of 80% was accepted. No major problems occurred as a result of differing mineral contents. 226Ra was occasionally present in large quantities. Since radium is extremely similar to calcium, it was thought that a high calcium content would indicate a high radium content. This was examined and could not be confirmed [2].
NATURALLY-OCCURRING U AND TH IN TABLE MINERAL WATERS
259
Polonium and lead
The 2°9p0 tracer peak did not interfere with that of 21°po, and therefore, tracer could be added to each sample. Only the length of time of autoelectrolysis affected the percentage yield. Four hours produced yields of 80--90%. Sample volume was increased ~o 2 litres with no effect on percentage yield, and indeed doubled results to well above the limit of detection. Polonium results could not be averaged due to occasional sudden increases in content. A very likely explanation is that 21°po and 21°pb are not in equilibrium at the time of analysis. Since the 2l°pb results are mostly higher than those for 2'°po it can be said that the mineral water on removal from source contained little or no 2l°po, instead it grew into the sample from 2~°pb. This is also suggested by the average length of time between removal from source and consumption (-- 3 months). Equilibrium is therefore almost always never reached. Ground rock is, of course, a very important factor in the natural radioactivity content of mineral water. However, the extreme values determined for some waters can perhaps be explained by a lack of homogeneity in the water when taken from its source. ACKNOWLEDGEMENTS
This project was proposed and financially supported by the Federal Office of Public Health, Division of Food Science. REFERENCES 1 2 3
M. Cospilo and L. Rigali, Determination of thorium in natural waters after extraction with aliquat-336. Anal. Chim. Acta. 106 (2) (1979) 385-388. O. Hogl, Die Mineral-und Heilquellen der Schweiz, Haupt, Berne, Switzerland. J.C.Laul et al., Analysis of natural radionuclides from uranium and thorium decay series in briny groundwaters. J. Radioanal. Nucl. Chem., I I0 (l)(1987) 101-112.