Science of the Total Environment

Transcription

1 Science of the Total Environment 437 (2012) Contents lists available at SciVerse ScienceDirect Science of the Total Environment journal homepage: Natural radionuclides in bottled drinking waters produced in Croatia and their contribution to radiation dose Martina Rožmarić a,, Matea Rogić a, Ljudmila Benedik b, Marko Štrok b a Ruđer Bošković Institute, Bijenička cesta 54, Zagreb, Croatia b Jožef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia HIGHLIGHTS Levels of 234,238 U, 226,228 Ra, 210 Po and 210 Pb in Croatian bottled drinking waters Dose assessment for three groups of population (infants, children and adults) The most critical group of population are children (they assessed the highest dose) Contribution of each radionuclide to total annual effective ingestion doses A base for further investigation of radioactivity in Croatian drinking water article info abstract Article history: Received 1 June 2012 Received in revised form 6 July 2012 Accepted 6 July 2012 Available online xxxx Keywords: Bottled drinking water Natural radionuclides Dose assessment Activity concentrations of 234 U, 238 U, 226 Ra, 228 Ra, 210 Po and 210 Pb in all Croatian bottled drinking natural spring and natural mineral water products, commercially available on the market, were determined. The samples originated from various geological regions of Croatia. Activity concentrations of measured radionuclides are in general decreasing in this order: 234 U> 238 U> 226 Ra> 228 Ra> 210 Pb> 210 Po and 226 Ra> 228 Ra> 234 U> 238 U> 210 Pb> 210 Po for natural spring and mineral waters, respectively. Based on the radionuclide activity concentrations average total annual effective ingestion doses for infants, children and adults, as well as contribution of each particular radionuclide to total dose, were assessed and discussed. The highest doses were calculated for children from 7 to 12 years of age, which makes them the most critical group of population. All values for each type of water, as well as for each population group, were well below the recommended reference dose level (RDL) of 0.1 msv from one year's consumption of drinking water according to the European Commission recommendations from Contribution of each particular radionuclide to total doses varied among different water types and within each water type, as well as between different age groups, where the lowest contribution was found for uranium isotopes and the highest for 228 Ra Elsevier B.V. All rights reserved. 1. Introduction With the increasing global problems of fresh water supplies, humans are becoming more aware of how precious water is nowadays. Commercial bottled water has become one of the main drinking sources, especially in developing countries, but an increasing trend of its consumption is also visible in European countries. As Croatia is rich in natural and mineral water springs many domestic bottled water brands are available on the market, categorized in two types: natural spring and natural mineral water (depending on the mineral content). By the Croatian legislation (Official Gazette, NN 58/98) both natural spring and mineral waters are obtained from underground Corresponding author. Tel.: ; fax: address: rozmar@irb.hr (M. Rožmarić). beds from one or more natural or drilled springs, protected from all kinds of pollution, where the difference between natural spring and mineral waters is in total dissolved solid content (mineral waters contain more than 500 mg L 1 ). Humans are constantly exposed to radiation from external sources as well as from internal sources through inhalation and ingestion by food and water consumption (UNSCEAR, 2000). The main contribution to dose is largely due to the presence of naturally occurring radionuclides of both uranium and thorium decay chains, which are omnipresent in the Earth's crust and leached by surface waters and especially groundwaters. For this reason, the geological setting strongly influences the occurrence of natural radionuclides in drinking water. Mineral waters stemming from very deep aquifers usually show higher loads of natural radionuclides compared to surface water and water from shallower wells /$ see front matter 2012 Elsevier B.V. All rights reserved.

2 54 M. Rožmarić et al. / Science of the Total Environment 437 (2012) Monitoring of natural radioactivity in drinking water is important from radioecological aspect because of relatively high radiotoxicity of some of natural radionuclides and their importance in the study of cumulative radiation effects on human health. The occurrence of radionuclides in drinking water can cause health hazards to humans as a result of internal radiation exposure from their direct ingestion and absorption in human tissues. Due to their high radiotoxicity (Harrison et al., 2007; Leggett and Eckerman, 2001; Mays et al., 1985), the contributions of 210 Po and 228 Ra to the dose are more pronounced. For accurate dose assessment it is necessary that specific radionuclides in drinking water are identified and their individual activity concentrations measured since the dose coefficients are always related to the specific radionuclides. There are a variety of techniques for the specific radionuclide identification and determination, such as: alpha spectrometry (Garcia-Torano, 2006), inductively coupled plasma-mass spectrometry (ICP-MS) (Becker, 2005; Lariviere et al., 2006), liquid scintillation counting (LSC) (Forte et al., 2001, 2003), gamma spectrometry (Jia and Jia, 2012), etc. Based on their advantages and disadvantages, these techniques can be used for different purposes in the exposure monitoring procedure. Due to the vital importance of water in human life and the necessity for the achievement of the lowest possible radiation exposure of the population, systematic studies on the radiological characterization of drinking water started in 1990s after the recommendations were published in Guidelines for Drinking Water Quality, issued by the World Health Organisation (WHO, 1998). These guidelines, as well as the European Directive 98/83/EC, published in 1998 (EC, 1998), state that the reference dose level of the committed annual effective dose due to drinking water consumption is 0.1 msv. The Directive points out that the total indicative dose must be evaluated, excluding tritium, 40 K, 14 C, radon and its decay products, but including all other radionuclides of the natural decay chains. The Directive 98/83/EC was followed by the European Commission Recommendation 2001/928/EURATOM (2001/928/Euratom, 2001) concerning radiological quality of drinking water supplies regarding radon and long-lived radon decay products, such as 210 Pb and 210 Po for which the recommended maximum activity concentrations are set to 0.2 Bq L 1 and 0.1 Bq L 1, respectively. Uranium is also covered by the WHO Guidelines, although its contribution to the ingestion dose is minor due to its low dose conversion factor. However, uranium is a toxic heavy metal, especially hazardous for kidneys where it is highly accumulated (Wrenn et al., 1985), and therefore has to be regulated and controlled even though maximum allowed activity concentrations for its isotopes are higher than for other radionuclides. The WHO in Chemical aspects of drinking water quality set the most stringent limitation with 2 μg L 1 in the 1998 report (WHO, 1998)butlaterchanged this limit to 15 μg L 1 (WHO, 2004) andfinally to 30 μg L 1 (WHO, 2011) based on its chemical toxicity, which is predominant compared with its radiological toxicity. USA set limits on the maximum uranium mass concentrations of 30 μg L 1 in drinking water (US EPA, 2000) while Germany set a maximum level of 2 μg L 1 for mineral water considered to be suitable for infants (Bundesgesetzblatt, 2006). The new WHO Guidelines for Drinking Water Quality from 2011 (WHO, 2011) in Radiological aspects gave the same recommendations for activity concentrations for a wide range of radionuclides as in 2004, with guidance levels of 0.1 Bq L 1 for 210 Po, 228 Ra and 210 Pb, 1 Bq L 1 for 234 Uand 226 Ra, and 10 Bq L 1 for 238 U. The values of various water quality parameters recommended by WHO and European Commission are the general guidelines. Therefore, many countries have established their own water quality standards to meet their national priorities. No such Regulation, giving the highest allowed activity concentration of particular radionuclide in drinking water, has yet been issued in Croatia, but the highest received annual dose for water consumption is currently set to value of 0.1 msv, along with the maximum allowed activity concentration for tritium, being 100 Bq L 1 (which must only be determined when a new spring location is being formed and spring water used for human consumption) (Official Gazette, NN 47/08). As the annual consumption of bottled water increases, and due to the fact that more than 98% of total population exposure to radiation comes from natural sources (excluding medical exposure) (UNSCEAR, 2000), estimation of total annual effective dose received from consumption of these products is necessary, along with the assessment of the contribution of particular radionuclides to the total dose. Several studies to assess the natural radioactivity levels in bottled drinking and mineral water were performed across Europe (Benedik and Jeran, 2012; Beyermann et al., 2010; Chau and Michalec, 2009; Desideri et al., 2007a, 2007b; Duenas et al., 1997; Forte et al., 2007; Gruber et al., 2009; Jia et al., 2009; Jia and Torri, 2007; Jobbágy et al., 2010; Karamanis et al., 2007; Kehagia et al., 2007; Kendall, 2004; Kozlowska et al., 2007; Palomo et al., 2007; Skwarzec et al., 2003, 2001; Vesterbacka et al., 2005; Wallner and Jabbar, 2010; Wallner et al., 2008), while data on bottled drinking water quality in Croatia are very scarce (Peh et al., 2010) and data on radiological quality are still not available. In the last decades, nationwide studies have also been performed in some developing countries where tap water is not approved for human consumption and is replaced by bottled water (Ajayi and Achuka, 2009; Amrani, 2002; Ben Fredj et al., 2005; Fatima et al., 2007; Jha et al., 2009; Osman et al., 2008; Seghour and Seghour, 2009). The main aim of our study was to obtain a representative estimation of the activity concentration levels of natural radionuclides in bottled drinking water and the corresponding radiation doses for people consuming these products. This paper presents data for the activity concentrations of 238 U, 234 U, 226 Ra, 228 Ra, 210 Po and 210 Pb in Croatian commercially available bottled drinking natural spring and mineral waters and assessment of total annual effective doses received from mentioned radionuclides for three groups of Croatian population (infants, children and adults). The data generated in this research may contribute to determine the base-line levels of natural radioactivity in drinking water and help in the development of future guidelines for radiological protection of the Croatian population. 2. Materials and methods 2.1. Sampling The drinking water samples of twelve different brands were purchased «from the shelf» on the market in Zagreb. Three samples of natural mineral carbonated water (MW1, MW2, MW3) and nine samples of natural spring noncarbonated (NW1, NW2, NW3, NW5, NW6, NW7, NW8, NW9) and carbonated (NW4) water were analyzed. Springs locations of the analyzed bottled natural and mineral waters are shown in Fig. 1. The collected samples are produced by Croatian companies and originated from various geological regions of Croatia with bedrock aquifers of various depths (90 800) m. Table and tap waters were not included in this study Instruments An alpha spectrometer (Alpha Analyst, Canberra, USA) with a passivated implanted planar silicon (PIPS) semiconductor detectors, with an active area of 450 mm 2 and 28% efficiency for a 25 mm diameter disk, was used for alpha particle spectrometric measurements of 234 U, 238 U, 226 Ra and 210 Po. The calibration of the detector was made with a standard radionuclide source, containing 238 U, 234 U, 239 Pu and 241 Am (code: ), obtained from Analytics, Inc. Data acquisition and analysis were performed with the Genie 2000 spectroscopy system software. The counting time varied from 3 to 8 d depending on the radionuclide activities in the samples. A low background gas flow proportional counter (TENNELEC LB4100-W) was used for measurement of 210 Pb daughter radionuclide

3 M. Rožmarić et al. / Science of the Total Environment 437 (2012) Fig. 1. Geographical springs locations of the analyzed natural spring (NW) and mineral (MW) waters. 210 Bi, which was calibrated to account for corrections due to in-growth of 210 Bi and different self-absorption for samples with different radiochemical yields (Štrok et al., 2008). A coaxial HPGe gamma spectrometer (ORTEC) was used for 133 Ba (Eγ= kev, Pγ=62.05%) measurement in 226 Ra and 228 Ra yield determination and for 228 Ra (via 228 Ac; Eγ= kev, Pγ=25.8%) measurement. For evaluation of gamma spectra, the Hyperlab (HyperLab, 2005) program was used Chemicals Radioactive tracers were used for evaluation of radiochemical yields in determination of 234 U, 238 U, 226 Ra, 228 Ra and 210 Po. Working solutions of 232 U, 133 Ba and 209 Po were prepared from standard solutions purchased from Analytics, Inc. (Atlanta, GA, USA). Standard of stable lead solution, prepared by dissolution of Pb(NO 3 ) 2 (Merck, Germany), was used for gravimetrical determination of radiochemical yield of 210 Pb separation. All other used chemicals were of analytical grade Radiochemical methods Analyses of natural radionuclides ( 234,238 U, 226,228 Ra, 210 Po and 210 Pb) were performed by radiochemical procedures summarized in Table 1 and described in detail in Sections Quality control was performed by participation in many international interlaboratory comparisons organized by different institutions (IAEA, NPL, BFS, EC) in period from 2006 to 2011, with all results being acceptable Determination of uranium isotopes For determination of uranium radioisotopes a known amount of 232 U tracer was added to water sample previously acidified with concentrated HNO 3 (3 ml of acid per 1 L of the sample). Uranium was preconcentrated from the water sample by Fe(OH) 3 coprecipitation at ph 9 10 using an ammonia solution. The precipitate was centrifuged, rinsed with water (to ph=7) and dissolved in 3 M HNO 3.Uranium separation was performed on UTEVA resin (Eichrom) (Eichrom Technologies, 2001) whichwaspreconditionedin3mhno 3.Interfering elements were separated by washing the column with 3 M HNO 3,9MHCland0.5MH 2 C 2 O 4 /5 M HCl. Uranium radioisotopes were stripped from the column with 1 M HCl. The source for alpha spectrometric measurement was prepared by microcoprecipitation with NdF 3 (Hindman, 1983; Sill and Williams, 1981). The neodymium fluoride suspension was filtered through a 25 mm diameter, 0.1 μm polypropylene filter, the microcoprecipitate was dried Table 1 Summary of methods used for bottled drinking water analysis. Radionuclide Tracer Analytical method V (L) L D * (mbq L 1 ) 234,238 U 232 U Fe(OH) 3 precipitation, extraction chromatography, microcoprecipitation, alpha spectrometry 210 Po 209 Po MnO 2 precipitation, extraction chromatography, self deposition, alpha spectrometry 210 Pb Pb 2+ MnO 2 precipitation, extraction chromatography, PbSO 4 precipitation, low-level beta counting ( 210 Bi) 226 Ra 133 Ba Pb(Ra)(Ba)SO 4 precipitation, alpha spectrometry ( 226 Ra), gamma spectrometry ( 133 Ba) 228 Ra 133 Ba Pb(Ra)(Ba)SO 4 precipitation, gamma spectrometry ( 133 Ba, 228 Ac) V sample volume; *L D detection limit (Currie, 1968)

4 56 M. Rožmarić et al. / Science of the Total Environment 437 (2012) under an infrared lamp, mounted on an aluminum disk and measured on alpha spectrometer Ra determination Procedure for determination of 226 Ra was based on co-precipitation of Ra(Ba)(Pb)SO 4 (Lozano et al., 1997). The water sample was acidified with concentrated H 2 SO 4 (10 ml of acid per 1 L of the sample) and after addition of 133 Ba tracer the sample was stirred for 30 min. 50 mg Pb 2+ carrier solution was added to allow good co-precipitation of radium and barium. After settling, the suspension was centrifuged, rinsed with water (to ph=7) and the obtained PbSO 4 precipitate, containing radium and barium, was dissolved in 0.1 M EDTA prepared in 0.5 M NaOH. The counting source was prepared by microcoprecipitation of Ra(Ba)SO 4 upon addition of Ba carrier solution (0.3 mg ml 1 ), 1:1 acetic acid (ph=4.5 5), saturated Na 2 SO 4 solution and Ba seeding solution. The microcoprecipitate (Ra(Ba)SO 4 ) was filtered through a 0.1 μm polypropylene filter, dried, mounted on an aluminum disk and measured on a gamma spectrometer for 133 Ba (yield determination of 226 Ra) and on alpha spectrometer for 226 Ra determination Ra determination For 228 Ra determination a larger sample volume (10 L) was taken and acidified with concentrated H 2 SO 4 (10 ml of acid per 1 L of the sample). Upon the addition of 133 Ba tracer solution, the sample was stirred for 30 min. (Ra)(Ba)PbSO 4 was co-precipitated by adding of Pb 2+ carrier solution. The suspension was stirred for 4 h and left to settle down over night. The remaining supernatant was decanted, the precipitate centrifuged and washed till neutral ph, then transferred by centrifugation onto a counting planchet. The obtained (Ra)(Ba)PbSO 4 was measured on a gamma spectrometer for yield determination (via 133 Ba) and for activity measurement of 228 Ra (via its in-growth daughter 228 Ac) Po and 210 Pb determination For 210 Po and 210 Pb determination in water, tracer solutions of 209 Po and stable Pb 2+ were added to 9 L of water. After sample acidification with concentrated HCl (2 ml of acid per 1 L of the sample) the radionuclides were co-precipitated with MnO 2. Precipitation of MnO 2 was achieved by adding KMnO 4, MnCl 2 and adjusting the ph to 9 with ammonia solution. The precipitate was centrifuged, rinsed with water (to ph=7) and dissolved in 2 M HCl. Separation of polonium from lead was performed on Sr specific resin (Eichrom) (Eichrom Technologies, 2009) preconditioned with 2 M HCl. Polonium was stripped from the column with 6 M HNO 3 and lead with 6 M HCl. Obtained fractions were evaporated to dryness. Polonium source was prepared by self-deposition on a silver disk from HCl solution (ph=1 2) with addition of 0.5 g of ascorbic acid to prevent deposition of interfering elements. The Ag disk, covered on one side, was fixed in a holder and immersed in the solution (Benedik et al., 2009). Spontaneous deposition of polonium was carried out at 90 C for 4 h. The disk was rinsed with water and ethanol, dried at room temperature and polonium radioisotopes were measured on alpha spectrometer. 210 Pb source was prepared by precipitation of PbSO 4. Dry lead fraction residue was dissolved in water and concentrated H 2 SO 4 was added. The precipitate was transferred to a counting planchet and measured after in-growth of its daughter radionuclide 210 Bi on a low background gas flow proportional counter. 3. Results and discussion 3.1. Radionuclide activity concentrations The results of the activity levels of 238 U, 234 U, 226 Ra, 228 Ra, 210 Po and 210 Pb in two different groups of drinking bottled water (natural spring and natural mineral) produced in Croatia are given in Table 2. The results show that activity concentrations of uranium in water samples are in range ( )mBq L 1 and ( )mBq L 1 for 238 Uand 234 U, respectively. There are no significant differences between values measured in natural spring and mineral waters. These values are relatively low compared to some literature data (Beyermann et al., 2010; Jia et al., 2009; Jia and Torri, 2007; Jobbágy et al., 2010) and well below the limit values given by European Commission and World Health Organization (EC, 1998; World Health Organisation WHO, 2011). However, some elevated uranium activity levels (higher than 10 mbq L 1 ) were found in three bottled natural spring waters (NW2, NW5 and NW8) and one mineral water (MW2) originating from the north-west part of Croatia. Radium activity concentrations are in range ( )mBq L 1 and (b )mBq L 1 for 226 Ra and 228 Ra, respectively, and can be considered low and comparable with data reported for Europe (Benedik and Jeran, 2012; Beyermann et al., 2010; Chau and Michalec, 2009; Jia et al., 2009; Jia and Torri, 2007). The highest absolute activity concentrations were found in mineral waters (MW1, MW2 and MW3) and some natural waters (NW2, NW5 and NW8) originating from the region in the north-west part of the country (Fig. 1), known for its thermal and mineral springs. 210 Po activity concentrations are in range ( )mBq L 1 and can be regarded as the lowest among all analyzed radionuclides, while the range for 210 Pb is ( )mBq L 1, which is also comparable with already published literature data for surrounding countries (Benedik and Jeran, 2012; Jia et al., 2009; Jia and Torri, 2007; Wallner et al., 2008). There are no significant differences, taking into account combined standard uncertainties, between 210 Po and 210 Pb activity concentration values measured in both types of waters. If the results are compared with other studies, radionuclide activity concentrations in Croatian commercial bottled waters are significantly lower or equal. Based on the results given in Table 2 it is obvious that the highest average activity concentrations in natural spring waters come from uranium isotopes, while mineral waters have the highest average activity concentrations of radium isotopes. Activity concentrations are in general decreasing in this order: 234 U> 238 U> 226 Ra> 228 Ra> 210 Pb> 210 Po and 226 Ra> 228 Ra> 234 U> 238 U> 210 Pb> 210 Po for natural spring and mineral waters, respectively Assessment of combined annual effective ingestion doses from bottled natural spring and mineral waters Based on the results of activity concentrations of already mentioned radioisotopes in different drinking water samples, presented in Table 2, the total annual effective ingestion doses for the members Table 2 Radionuclide activity concentrations A(mBq L 1 ) in bottled natural spring (NW) and mineral (MW) waters produced in Croatia with combined standard uncertainties (k=2). Sample A( 238 U) A( 234 U) A( 226 Ra) A( 228 Ra) A( 210 Po) A( 210 Pb) NW1 6.5± ± ± ± ± ±0.6 NW2 11.0± ± ± ± ± ±1.2 NW3 5.1± ± ± ± ± ±0.6 NW4 5.1± ± ± ± ± ±0.4 NW5 13.1± ± ± ± ± ±0.7 NW6 6.6± ± ± ± ± ±0.6 NW7 3.4± ± ±0.3 b ± ±1.2 NW8 14.6± ± ± ± ± ±0.9 NW9 6.3± ± ± ± ± ±0.7 Mean 8.0± ± ± ± ± ±2.4 Range b MW1 6.7± ± ± ± ± ±0.6 MW2 10.7± ± ± ± ± ±1.3 MW3 2.1± ± ± ± ± ±0.3 Mean 6.5± ± ± ± ± ±1.5 Range

5 M. Rožmarić et al. / Science of the Total Environment 437 (2012) Table 3 Dose coefficients of the relevant radionuclides (C f )inμsv Bq 1 for three age groups (IAEA, 1996). of the public were estimated. Contributions of each radionuclide to the total dose for 3 critical age groups of the population, infants (1 2 years), children (7 12 years) and adults(>17 years), are also given. For the total annual effective dose calculation, Eqs. (1) and (2) and the data from Table 2 were used: E d ¼ A c qc f D ¼ Σ A c qc f 238 U 234 U 226 Ra 228 Ra 210 Pb 210 Po Age group 1 2 years years >17 years where: E d is the annual effective ingestion dose due to relevant radionuclide in μsv, A c is radionuclide activity concentration in the water sample in Bq L 1, q is the annual consumption rate of 150 L for infants, 350 L for children and 500 L for adults according to the UNSCEAR 2000 (UNSCEAR, 2000), C f is the dose coefficient of the relevant radionuclide for each age group given in the «International Basic Safety Standards for Protection against Ionizing Radiation and for Safety of Radiation Sources» (IAEA, 1996) and shown in Table 3 while D is the total annual effective ingestion dose due to all determined radionuclides in μsv. Numerous studies have thus far been made for estimation of average annual water intake for different age groups of the population. These studies provide a wide range of data where estimates of water intake sources and amounts vary substantially across studies. Some estimate annual water intake by both food and water consumptions (Levallois et al., 1998), bottled and tap water combined (Levallois et al., 1998) or separately (UNSCEAR, 2000; US EPA, 2000), or just give recommendations for a healthy daily water intake which is, in some cases, assumed to be overestimated and non realistic (WHO, 2011). Data used in this study was taken from UNSCEAR (2000), which is in accordance with EPA (US EPA, 2000), and refers to a more realistic estimation of average annual bottled water consumption rate for different age groups of the population. In addition it is important to emphasize that tap water is still the main source of daily water consumption in Croatia, which is ð1þ ð2þ the main reason why it was relevant to use the data referring to bottled water intake only. Total annual effective ingestion doses (μsv) assessed from already mentioned radionuclides for three groups of population by bottled natural spring and mineral water consumptions are summarized in Fig. 2 as minimum to maximum range values, median (Me) and average values, and first and third quartiles. It is evident that the range of doses is rather wide for each type of waters as well as for each population group, but all values are well below the recommended reference dose level (RDL) of 0.1 msv from one year's consumption of drinking water (EC, 1998). The doses received from the consumption of mineral waters (Me: 25.7 μsv, 42.2 μsv and 15.2 μsv for infants, children and adults, respectively) are higher than those received from the consumption of natural spring waters (Me: 7.9 μsv, 11.4 μsv and 5.0 μsv for infants, children and adults, respectively) for all three age groups. The highest doses were calculated for children from 7 to 12 years of age, which makes them the most critical population group. Further on, the obtained results of total annual effective doses (μsv) for 3 age groups and 2 types of bottled water are presented in Fig. 3. It can be seen that doses assessed from natural spring water consumption are very low for all age groups and ranged (2.9± ±2.0)μSv, (3.8± ±2.7)μSv and (1.6± ±1.1) μsv for infants, children and adults, respectively. If a conservative assumption is made that annual water intake consisted only of one particular water brand (NW2), the highest effective doses received would be approximately 15 μsv, 25 μsv and 10 μsv for infants, children and adults, respectively. These values are representing only up to 25% of the recommended reference dose level (WHO, 2011) for the most sensitive population group (7 to 12 years old children) (UNSCEAR, 2000), but can be considered as highly improbable. On the other hand, higher total effective doses are received by mineral water consumption and are in range (17.8± ±4.9)μSv, (28.9± ±7.8)μSv and (10.9± ±2.0)μSv for infants, children and adults, respectively. Doses higher than 35 μsv, 60 μsv and 15 μsv for infants, children and adults, respectively, would be received only by annual consumption of one mineral water brand (MW3), which is unlikely in case of infants and children since they rarely consume mineral water in higher amounts. The adults would in such a case receive less than 20% of the maximum recommended reference dose level of 0.1 msv (WHO, 2011). As seen from the obtained results, also contrary to some of other similar studies made in European countries (Beyermann et al., 2010; Chau and Michalec, 2009; Desideri et al., 2007a,b), all of the assessed Fig. 2. Total annual effective ingestion doses (μsv) for infants (1 2 years), children (7 12 years) and adults (>17 years) due to radionuclide intake by consumption of bottled natural spring (NW) and mineral (MW) waters presented as box plot diagram.

6 58 M. Rožmarić et al. / Science of the Total Environment 437 (2012) (up to 18% for infants) compared to mineral waters where it is below 3%. 210 Pb dose contribution is around 20% for natural spring waters for all three age groups, unlike for mineral waters where it is up to only 5%. The highest dose received from 210 Pb is for the adult population regardless of the water type. If 226 Ra and 228 Ra are considered together, high dose contributions are estimated for both types of waters (up to 95% and 69% for mineral and natural spring waters, respectively) for all age groups. Proportions of the one of the most radiotoxic isotope, 228 Ra, to total annual effective ingestion doses are higher than for 226 Ra for all age groups, and considerably higher than all other determined radionuclides (up to 71% and 68% for infants and children, respectively). 4. Conclusion Fig. 3. Total annual effective ingestion doses (μsv) for three groups of Croatian population assessed by drinking different types of natural spring and mineral waters with combined standard uncertainties (k=2). doses due to consumption of bottled waters produced and sold in Croatia are well below the dose limit of 0.1 msv/year. The average values of assessed doses obtained for individual radionuclides for three different age groups by consumption of natural spring and mineral waters are presented in Table 4. If the individual doses received from the presented radionuclides are compared only for natural spring water consumption, for children and adults, their contribution to the total effective dose is in this order: 228 Ra> 210 Pb> 226 Ra> 210 Po> 234 U> 238 U. The situation is similar for infants but 210 Po is in this case the grater contributor to total dose than 226 Ra. As for mineral water consumption, all three age groups exhibit the same decreasing order ( 228 Ra> 226 Ra> 210 Pb> 210 Po> 234 U> 238 U) for radionuclide's contribution to the total effective dose. Investigation of the contribution of particular radionuclides to the assessed total annual effective ingestion doses for three different age groups, drinking two different types of water, showed high variability and mostly reflected the variability of radionuclide activity concentrations in different water sample types and variability of dose coefficients among different age groups (Table 3). As seen from Fig. 4 contribution of both uranium isotopes ( 234 U and 238 U) is very low for both types of waters and age groups (due to relatively low radiotoxicity compared to other radionuclides of interest, Table 3). Their contribution to total dose is also lower for mineral waters than for natural spring waters (due to lower activity concentration in mineral waters, compared to natural spring waters), where the highest is found for adult population, but even in that case being less than 10% of total effective dose. 210 Po is a highly radiotoxic isotope with a high dose conversion factor. Its dose contribution is thus higher than from uranium isotopes despite its low activity concentration in all samples. Significantly higher contribution was noted for natural spring waters Presented study is the first one providing information on activity concentrations of natural alpha ( 234 U, 238 U, 226 Ra and 210 Po) and beta ( 228 Ra and 210 Pb) emitting radionuclides in all bottled natural spring and natural mineral waters produced and sold in Croatia. The results showed that the highest activities in natural spring waters come from uranium isotopes, while mineral waters have the highest concentrations of radium isotopes. Activity concentrations of measured radionuclides are in general decreasing in this order: 234 U> 238 U> 226 Ra> 228 Ra> 210 Pb> 210 Po and 226 Ra> 228 Ra> 234 U> 238 U> 210 Pb> 210 Po for natural spring and mineral waters, respectively. The results are comparable with researches from other European countries. Based on the activity concentration results obtained in this study, total annual effective ingestion doses for three different age groups (infants, children and adults) and contribution of each radionuclide to the assessed doses were estimated. The highest doses were calculated for children from 7 to 12 years of age, which makes them the most critical group of population, but all values for each type of water, as well as for each population group, are well below the recommended reference dose level (RDL) of 0.1 msv from 1 year's consumption of drinking water (EC, 1998). Contribution of both uranium isotopes ( 234 U and 238 U) is very low for both types of waters and age groups and followed by contribution of 210 Po and 210 Pb. If 226 Ra and 228 Ra are considered together, high dose contributions are estimated for both types of waters (up to 95% and 69% for mineral and natural spring waters, respectively) for all age groups. However, due to its high radiotoxicity, proportions of 228 Ra are higher than 226 Ra and considerably higher than all other determined radionuclides. This study further showed that radionuclides' activity concentrations, and thus estimated annual effective ingestion doses assessed from natural bottled water consumption, whose spring locations are in different regions of Croatia, vary considerably. Since considerable amounts of tap water are used in Croatia for drinking purposes, an extension of this survey to analysis of drinking tap water is needed to provide additional information and to get an overall picture of the combined annual effective ingestion doses due to water consumption. Table 4 Total annual effective ingestion doses (μsv) assessed from individual radionuclides by water consumption for different age groups with combined standard uncertainties (k=2). Type of water Age group Annual effective dose/μsv 238 U 234 U 226 Ra 228 Ra 210 Pb 210 Po NW Infants (1 2) 0.14± ± ± ± ± ±0.3 Children (7 12) 0.19± ± ± ± ± ±0.2 Adults (>17) 0.18± ± ± ± ± ±0.1 MW Infants (1 2) 0.12± ± ± ± ± ±0.1 Children (7 12) 0.15± ± ± ± ± ±0.1 Adults (>17) 0.15± ± ± ± ± ±0.1

7 M. Rožmarić et al. / Science of the Total Environment 437 (2012) Fig. 4. Contribution of each analyzed radionuclide to the total annual effective ingestion dose among different types of age groups in natural spring and mineral waters with combined standard uncertainties (k=2). Acknowledgments This work was financially supported by the Ministry of Science, Education and Sports of the Republic of Croatia (bilateral project Radiochemical methods for determination of radionuclides in water samples ) and Ministry of Higher Education, Science and Technology of the Republic of Slovenia (bilateral project BI-RH/ ). References 2001/928/Euratom. Commission recommendation on the protection of the public against exposure to radon in drinking water supplies. Off J Eur Communities 2001:4580. [C]. Ajayi OS, Achuka J. Radioactivity in drilled and dug well drinking water of Ogun state Southwestern Nigeria and consequent dose estimates. Radiat Prot Dosimetry 2009;135(1): Amrani D. Natural radioactivity in Algerian mineral waters. J Radioanal Nucl Chem 2002;252: Becker JS. Inductively coupled plasma mass spectrometry (ICP-MS) and laser ablation ICP-MS for isotope analysis of long-lived radionuclides. Int J Mass Spectrom 2005;242: Ben Fredj A, Hizem N, Chelbi M, Ghedira L. Quantitative analysis of gamma-ray emitters radioisotopes in commercialised bottled water in Tunisia. Radiat Prot Dosimetry 2005;117(4): Benedik L, Jeran Z. Radiological characterization of natural and mineral drinking waters in Slovenia. Radiat Prot Dosimetry 2012; Benedik L, Vasile M, Spasova Y, Wätjen U. Sequential determination of 210 Po and uranium radioisotopes in drinking water by alpha-particle spectrometry. Appl Radiat Isot 2009;67: Beyermann M, Bünger T, Schmidt K, Obrikat D. Occurrence of natural radioactivity in public water supplies in Germany: 238 U, 234 U, 235 U, 228 Ra, 226 Ra, 222 Rn, 210 Pb, 210 Po and gross α activity concentrations. Radiat Prot Dosimetry 2010;141: Bundesgesetzblatt Nr. 56 Seite 2762, Deutsche Mineral and Tafelwasserverordnung. Chau ND, Michalec B. Natural radioactivity in bottled natural spring, mineral and therapeutic waters in Poland. J Radioanal Nucl Chem 2009;279: Currie LA. Limits for qualitative detection and quantitative determination, application to radiochemistry. Anal Chem 1968;40: Desideri D, Meli MA, Feduzi L, Roselli C, Rongoni A, Saetta D. 238 U, 234 U, 226 Ra, 210 Po concentrations of bottled mineral waters in Italy and their dose contribution. J Environ Radioact 2007a;94: Desideri D, Roselli C, Feduzi L, Meli MA. Radiological characterization of drinking waters in Central Italy. Microchem J 2007b;87:13 9. Duenas C, Fernandez MC, Liger E, Carretero J. Natural radioactivity levels in bottled water in Spain. Water Res 1997;31: Eichrom Technologies Inc. Analytical procedures. Lead-210 and Polonium-210 in Water, Rev 2.0; Eichrom Techologies Inc. Analytical procedures. Uranium and Thorium in Water, Rev 1.7; European Commission. Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption. Official Journal L 330; p Fatima I, Zaidi JH, Arif M, Tahir SNA. Measurement of natural radioactivity in bottled drinking water in Pakistan and consequent dose estimates. Radiat Prot Dosimetry 2007;123(2): Forte M, Rusconi R, Margini C, Abbate G, Maltese S, Badalamenti P, et al. Determination of uranium isotopes in food and environmental samples by different techniques: a comparison. Radiat Prot Dosimetry 2001;97: Forte M, Rusconi R, Di Caprio E, Bellinzona S, Sgorbati G. Natural radionuclides measurements in drinking water by liquid scintillation counting. Methods and results. In: Warwick P, editor. Environmental Chemical Analysis, vol. II p RSC Cambridge. Forte M, Rusconi R, Cazzaniga MT, Sgorbati G. The measurement of radioactivity in Italian drinking waters. Microchem J 2007;85: Garcia-Torano E. Current status of alpha-particle spectrometry. Appl Radiat Isot 2006;64: Gruber V, Maringer FJ, Landstetter C. Radon and other radionuclides in drinking water in Austria: measurement and assessment. Appl Radiat Isot 2009;67: Harrison J, Leggett R, Lloyd D, Phipps A, Scott B. Polonium-210 as a poison. J Radiol Prot 2007;27: Hindman FD. Neodymium fluoride mounting for alpha spectrometric determination of uranium, plutonium and americium. Anal Chem 1983;55: HyperLab. System. Installation and Quick Start Guide. Budapest, Hungary: HyperLabs Software; International Atomic Energy Agency (IAEA). International basic safety standards for protection against ionizing radiation and for the safety of radiation sources. Safety Report Series, No. 115, Vienna; Jha SK, Lenka P, Gothankar S, Tripathi RM, Puranik VD, Khating DT. Radiological assessment of surface water quality around proposed uranium mining site in India. J Environ Radioact 2009;100: Jia G, Jia J. Determination of radium isotopes in environmental samples by gamma spectrometry, liquid scintillation counting and alpha spectrometry: a review of analytical methodology. J Environ Radioact 2012;106: Jia G, Torri G. Estimation of radiation doses to members of the public in Italy from intakes of some important naturally occurring radionuclides ( 238 U, 234 U, 235 U, 226 Ra, 228 Ra, 224 Ra and 210 Po) in drinking water. Appl Radiat Isot 2007;65: Jia G, Torri G, Magro L. Concentrations of 238 U, 234 U, 235 U, 232 Th, 230 Th, 228 Th, 226 Ra, 228 Ra, 224 Ra, 210 Po, 210 Pb and 212 Pb in drinking water in Italy: reconciling safety standards based on measurements of gross α and β. J Environ Radioact 2009;100:

8 60 M. Rožmarić et al. / Science of the Total Environment 437 (2012) Jobbágy V, Kávási N, Somlai J, Máté B, Kovács T. Radiochemical characterization of spring waters in Balaton Upland, Hungary, estimation of radiation dose to members of public. Microchem J 2010;94: Karamanis D, Stamoulis K, Ioannides KG. Natural radionuclides and heavy metals in bottled water in Greece. Desalination 2007;213:90 7. Kehagia K, Koukouliou V, Bratakos S, Seferlis S, Tzoumerkas F, Potiriadis C. Radioactivity monitoring in drinking water of Attika Greece. Desalination 2007;213: Kendall GM. Controls on radioactivity in water supplies in England and Wales, with especial reference to radon. J Radiol Prot 2004;24: Kozlowska B, Walencik A, Dorda J, Przylibski TA. Uranium, radium and K-40 isotopes in bottled mineral waters from Outer Carpathians, Poland. Radiat Meas 2007;42: Lariviere D, Taylor VF, Evans RD, Cornett RJ. Radionuclide determination in environmental samples by inductively coupled plasma mass spectrometry. Spectrochim Acta B 2006;61: Leggett RW, Eckerman KF. A systemic biokinetic model for polonium. Sci Total Environ 2001;275: Levallois P, Guevin N, Gingras S, Levesque B, Weber JP, Letarte R. New patterns of drinking water consumption: results of a pilot study. Sci Total Environ 1998;209: Lozano JC, Fernandez F, Gomez JMG. Determination of radium isotopes by BaSO 4 coprecipitation for the preparation of alpha-spectrometric sources. J Radioanal Nucl Chem 1997;223: Mays CW, Rowland RE, Stehney AF. Cancer risk from the lifetime intake of Ra and U isotopes. Health Phys 1985;48(5): Official Gazette, 47. (NN 47/08) Pravilnik o zdravstvenoj ispravnosti vode za piće(national regulation); Official Gazette, 58. (NN 58/98) Pravilnik o temeljnim zahtjevima za prirodne mineralne, izvorske i stolne vode (national regulation); Osman AAA, Salih I, Shaddad IA, El Din S, Siddeeg MB, Eltayeb H, et al. Investigation of natural radioactivity levels in water around Kadugli, Sudan. Appl Radiat Isot 2008;66: Palomo M, Penalver A, Borrull F, Aguilar C. Measurement of radioactivity in bottled drinking water in Spain. Appl Radiat Isot 2007;65: Peh Z, Šorša A, Halamić J. Composition and variation of major and trace elements in Croatian bottled waters. J Geochem Explor 2010;107: Seghour A, Seghour FZ. Radium and 40 K in Algerian bottled mineral waters and consequent doses. Radiat Prot Dosimetry 2009;133(1):50 7. Sill CW, Williams RL. Preparation of actinides for alpha spectrometry without electrodeposition. Anal Chem 1981;53: Skwarzec B, Strumińska DI, Borylo A. The radionuclides 234 U, 238 Uand 210 Po in drinking water in Gdańsk agglomeration (Poland). J Radioanal Nucl Chem 2001;250: Skwarzec B, Strumińska DI, Borylo A. Radionuclides of Po-210, U-234 and U-238 in drinking bottled mineral water in Poland. J Radioanal Nucl Chem 2003;256: Štrok M, Repinc U, Smodiš B. Calibration and validation of a proportional counter for determining beta emitters. Int J Power Energ Syst 2008;2: United Nations Scientific Committee on the Effects of Atomic Radiation UNSCEAR. Report to the general assembly, with scientific annexes, vol. I. Sources and Effects of Ionizing Radiation; United States Environmental Protection Agency (EPA). contaminants/basicinformation/radionuclides.cfm#what%20are%20epa%e2%80%99s %20drinking%20water%20regulations%20for%20radionuclides. Vesterbacka P, Makelainen I, Arvela H. Natural radioactivity in drinking water in private wells in Finland. Radiat Prot Dosimetry 2005;113: Wallner G, Jabbar T. Natural radionuclides in Austrian bottled mineral waters. J Radioanal Nucl Chem 2010;286: Wallner G, Wagner R, Katzlberger C. Natural radionuclides in Austrian mineral water and their sequential measurement by fast methods. J Environ Radioact 2008;99: World Health Organisation WHO. Guidelines for drinking water quality. Recommendations, with Addendum to Volume 1 on selected chemical substances, Geneva., Second edition, Volume World Health Organisation WHO. Guidelines for drinking water quality. Third edition. Geneva, Switzerland; World Health Organisation WHO. Guidelines for drinking water quality. 4th edition. Geneva, Switzerland; Wrenn ME, Durbin PW, Howard B, Lipsztein J, Rundo J, Still ET, et al. Metabolism of ingested U and Ra. Health Phys 1985;48: