Applied Radiation and Isotopes 87 (2014) 429–434
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
Critical remarks on gross alpha/beta activity analysis in drinking waters: Conclusions from a European interlaboratory comparison V. Jobbágy n, J. Merešová, U. Wätjen European Commission, Joint Research Centre, Institute for Reference Materials and Measurements, Retieseweg 111, B-2440 Geel, Belgium
H I G H L I G H T S
Gross alpha/beta standard methods for drinking water analysis are discussed. Large spread of results (up to 2 orders of magnitude) observed in comparisons. Sources of interferences are reviewed. We propose to use true standardized methods to obtain better measurement results.
art ic l e i nf o
a b s t r a c t
Available online 1 December 2013
The most common gross alpha/beta standard methods used for drinking water analysis are discussed, and sources of interferences are reviewed from a metrological point of view. Our study reveals serious drawbacks of gross methods on the basis of an interlaboratory comparison analyzing commercial mineral water samples with the participation of 71 laboratories. A proposal is made to obtain comparable measurement results using true standardized methods. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Gross alpha/beta activity Drinking water Interlaboratory comparison Radioactivity Limits of gross methods
1. Introduction Gross alpha/beta activity measurement is applied widely as a screening technique in the field of radioecology, environmental monitoring and industrial applications. Water intended for drinking purposes has to be analyzed first for gross alpha/beta activity content according to many different national and international standards and recommendations. Anticipating the new European Union (EU) Drinking Water Directive (EC, 2012) which incorporates gross alpha/beta activity screening levels, the Institute for Reference Materials and Measurements (IRMM) organized an interlaboratory comparison (ILC) to check the fitness for purpose of this method and the performance of European monitoring laboratories. On the basis of the reported values from the 71 participating laboratories, we review some of the most influential parameters on gross measurements. The claimed main advantages of the gross alpha/beta methods are the relatively low costs, rapidity and simplicity. Although it is one of the simplest, it is also one of the most disputed radioanalytical methods because the determination of gross alpha and beta activities
n Corresponding author. Present address: SCKCEN, Belgian Nuclear Research Centre, Boeretang 200, B-2400 Mol, Belgium. Tel.: þ 32 14 33 32 38. E-mail address: [email protected] (V. Jobbágy).
0969-8043/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apradiso.2013.11.073
faces some specific problems that may refute the afore-mentioned claims. There are many sources of interferences in gross alpha/beta measurement that may corrupt the comparability of the measurement results (Arndt and West, 2004; Rusconi et al., 2006; Semkow et al., 2004; Montaña et al., 2012). First one is related to the radionuclide composition of the sample. During gross alpha/beta activity measurement, a mixed radionuclide composition must be simultaneously measured. Drinking water samples may contain different naturally occurring alpha (238U, 234U, 232Th, 226Ra and 210Po) and beta (40K, 228Ra and 210Pb) emitters, and artificial radionuclides (241Am, 90Sr) in various concentrations (UNSCEAR, 2000). Moreover, most of these are members of a complex decay chain, therefore the ingrowths of the daughter products influence the measurement result. The second important source of interference is due to the final source thickness that causes self absorption of the emitted particles already in the source itself. In this respect, it is crucial to use standardized methods (Jobbágy et al., 2010). The WHO recommends that, "Where possible, standardized methods should be used to determine concentrations of gross alpha and beta activities" (WHO, 2011). For this reason, the most common standard methods – based on direct evaporation, co-precipitation and liquid scintillation counting – are discussed in this paper with respect to the sample preparation and counting techniques. Regarding the measurement methods, the influence of the following parameters must be considered: counting efficiency, self absorption, moisture absorption,
430
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
chemical recovery and the interferences due to the isotopic composition of the water sample. Experimental comparisons of the gross alpha/beta standard methods were done using real drinking water samples with different salinity and radionuclide activity concentration. Besides the pitfalls of the gross measurements, the paper also tries to give examples where gross measurements can be used as a rapid alternative technique to the radionuclide-specific analysis.
2. Materials and methods 2.1. Reagents All the chemicals (conc. H2SO4, FeCl3, BaCl2) were analytical grade and all the stock solutions were prepared using de-ionized water. Three water samples, with different salinity and radionuclide activity concentrations, were used for the described experiments and in the interlaboratory comparison. Two water samples (Water A and B) were commercially available natural mineral waters. The third one (Water C) was prepared by spiking de-ionized water with a known activity of 241Am and 90Sr/90Y solutions standardized at IRMM by liquid scintillation counting using the CIEMAT/NIST method. The total dissolved solid was set to a final total concentration of about 10 mg/L by adding the following inactive inorganic salts: NaCl, CaCl2 and Sr(NO3)2. The major sources of gross alpha activity for Water A are 234U and 238U ( 40 mBq/L each), for Water B this is 226Ra ( 330 mBq/L) and for Water C 241Am ( 950 mBq/L).
2.2. Sample preparation Sources were prepared in accordance with ISO 9696/9697 (“thick source method”) or ISO 10704 (“thin source method”). The surface density of a source prepared under ISO 9696/9697 must exceed 10 mg/cm2. An aliquot of sample was evaporated to dryness, and the dried residue converted to sulfate form by sulfuric acid and ignited at 350 1C for an hour (ISO 9696, 2007; ISO 9697, 2008). The surface density of the source in the thin source method must be below 5 mg/cm2, and has been described previously (ISO 10704, 2009; Montaña et al, 2012; Suarez-Navarro et al., 2002). The pH of the filtered water sample was adjusted with sulfuric acid and heated to purge radon and CO2. Then the radium isotopes were co-precipitated as Ba(Ra)SO4, whereas uranium, thorium and polonium isotopes are co-precipitated with Fe(OH)3 after pH adjustment (pH E7–8) (ISO 10704, 2009; Suarez-Navarro et al., 2002).
2.3. Gross counting system and activity calculation For the gross alpha/beta measurement, a 10-detector, lowbackground gas-flow proportional counting system was used. The high voltage was set to 1450 V and the counting gas (Ar/ CH4, 90/10) flow was kept stable with a flow rate of 25 mL/min. The gross alpha/beta count of the filtered/dried precipitate was measured in 5 h cycles repeating it several times (from three to twelve cycles). For the counting efficiency and self absorption experiments, 241 Am and 90Sr standard solutions were used, since these are the most frequently used radionuclides for this purpose. Gross alpha/ beta activity concentrations were calculated from count rates by following the formulae of the corresponding ISO standard (ISO 9696, 2007; ISO 9697, 2008; ISO 10704, 2009; ISO 11704, 2010).
3. Results and discussion 3.1. Summary of the interlaboratory exercise After radiochemical characterization of the three water samples at IRMM, they were sent to the European monitoring laboratories for gross alpha/beta analysis. The gross alpha/beta reference values were determined independently from the laboratory comparison from the results determined by the three laboratories involved in the reference value determination (Table 1). Water C is, in principle, the easiest sample to measure since its gross alpha/beta activity concentration is the highest among the ILC samples. However, from a measurement point of view, the gross alpha activity is not the only key factor, but the alpha/beta emitting radionuclides and the total dissolved solid content have to be considered as well. Taking into account all three factors one can make an order of difficulty in terms of measurement as follows: Water C oWater A rWater B. As is evident from the reported results (Fig. 1 and 2), the outcome of the laboratory comparison exercise is far from satisfactory. The measurement results span a wide range, e.g. for Water C the maximum reported gross beta activity was more than 3000 times higher than the minimum reported gross beta activity. Furthermore, several laboratories (no. 49, 50 etc.) present for one type of sample a measurement result several times higher than the reference, whilst for another type of sample the same laboratory has a result several times lower. The gross alpha activity results for Water C (spiked sample, i.e. the best case scenario) are sorted according to the applied counting techniques in Fig. 2. For Water A and B (natural waters) we could observe a similar data spread but with a higher degree of variation for each technique. Possible reasons for the diversity of results are given in the next sections, but a detailed evaluation of the laboratory comparison itself will be published later in a separate report. 3.2. Method comparison It is well-known that the sample form or geometry play a significant role in producing reliable results from thicker sources, requiring a uniform thickness and a homogeneous layer of residue material on a planchet (60 mm in diameter). The surface density, influencing the self absorption of alpha and beta particles, should be controlled and determined very carefully (ISO 9696/9697; ISO 10704; Montaña et al., 2012; Suarez-Navarro et al., 2002; Semkow et al., 2004; Parsa, 1998). A surface density higher than 10 mg/cm2 (ISO 9696/9697) achieves satisfying counting statistics and constant self absorption. The self absorption of the alpha and beta particles limits their counting efficiency to usually less than 50%. It should be mentioned that no energy resolution is possible with proportional counters. Since the sample is evaporated and later heat-treated, ISO 9696/9697 does not allow determination of the volatile radionuclides (e.g. 3H, 210Po, 137Cs), which escape from the sample and Table 1 The reference activity concentration values (Aref) in the ILC water samples with their expanded uncertainties (Uref) with a coverage factor k ¼ 2. Reference values determined by three laboratories external to the laboratory comparison. Parameter
Reference value with expanded uncertainty Aref 7 Uref (mBq/L) Water A
Gross alpha activity 47.5 7 22.8 Gross beta activity 309.8 7 57.4 Total dissolved solids (mg/L) 955 7 44
Water B
Water C
434.7 7 56.6 190.4 7 32.6 3647 27
954.5 7 77.3 1037.3 7 83.0 10.2 7 0.1
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
431
Fig. 2. Laboratory results of gross alpha activity concentration sorted by techniques for Water C. The solid line indicates the reference activity concentration (Aref) of gross alpha activity. Its corresponding expanded uncertainty 7 Uref (k ¼2) is plotted as dashed lines.
determination possible, no conclusion can be drawn on the chemical recovery. It should be stressed that in the case of real water samples it is difficult to quantify this without further analysis as the original chemical composition is not always known accurately. From the chemical composition of the sample, the excess precipitate formed from the original inorganic cations could be quantified. Second, 40K cannot be included in the gross beta result because it stays in the final supernatant. For this reason, the gross beta results obtained by co-precipitation methods are always lower than the true gross beta values. The 40K content has to be separately determined with either a radiometric method (low background gamma-ray spectrometry) or with a non-radiometric method (emission flame photometry).
3.2.1. Liquid scintillation counting (ISO 11704) Ultra low-level α/β-discrimination LSC, due to its high detection efficiency (up to 100%) and low background count rate, is a useful tool for the determination of alpha- and beta emitting radionuclides (Schönhofer, 1995). The basic steps of the sample preparation for liquid scintillation counting are very simple. It consists of a thermal pre-concentration, pH adjustment and sample mixing with scintillation cocktail (ISO 11704, 2010). Advantages are: Fig. 1. Laboratory results for gross alpha/beta activity concentration for different water samples. The solid lines indicate the reference activity concentrations (Aref) of gross alpha and gross beta activity, respectively. Their corresponding expanded uncertainties 7 Uref (k ¼2) are plotted in dashed lines.
the residue during heat treatment. Therefore, the gross alpha and beta activity concentration might be underestimated. For instance, 210 Po losses, varying with the chemical form of the element, have been reported beginning above 100 1C (Momoshima et al., 2002; Matthews et al., 2007). However, ISO 9696 claims that losses of polonium are not expected for samples “which have been acidified with nitric acid and subjected to sulfation” and treated at 350 1C (nitrate salts have been converted to the sulfate form). Some of these drawbacks of the thick source approach can be avoided by using co-precipitation of the water sample (SuarezNavarro et al., 2002; ISO 10704). On one hand, this method needs more chemical treatment than the evaporation methods. On the other, more uniform and homogeneous residues can be obtained, especially for higher salt concentrations (e.g. in mineral waters). However, it suffers from two very severe drawbacks. Firstly, since neither is a radiotracer applied nor is gravimetric yield
no self absorption occurs in the sample, LSC methods can be adapted to monitor low-energy beta
emitters such as 3H and 14C (Schönhofer, 1995; Gruber et al., 2009), and radon in water can be determined separately using liquid scintillation methods (Salonen, 2010).
One of the disadvantages is quenching (chemical, color and physical) which reduces the counting efficiency; therefore, a quench correction must be made. Furthermore, the separation of beta energies from all alpha energies can be difficult, because it depends on many factors in LSC counting (Rusconi et al., 2006; Schönhofer, 2012). For this reason, the optimal setting of the different parameters (type of the vial, cocktail, α/β pulse discrimination, counting efficiency) is essential for gross alpha/beta measurements by LSC. Many other factors like the physical or chemical behaviour of the radionuclides, dissolved ions and the energies of the emitted particles may influence the detection, too. As stated by Rusconi et al. (2006) the same settings should be used during measurement and calibration.
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
3.3. Sources of interferences 3.3.1. Self absorption Most gross alpha/beta standard methods apply proportional counter or solid state scintillation counters as detector (ISO 9696, 2007; ISO 9697, 2008; ISO 10704, 2009). The counting efficiency of these methods is strongly affected by the total dissolved solids and the surface density of the sample. The direct evaporation approach of ISO 10704 allows the use of a binding agent that contains a polymer “to produce evenly spread counting sources” (ISO 10704, 2009). This polymer stays on the planchet, increasing the surface density which cannot exceed 5 mg/cm2. To get equivalent alpha/ beta self-absorption for the sample and the calibration source in a multi-component microchemical environment is not easy to accomplish. The sample and the calibration source must be as similar as possible in terms of surface density, homogeneity, and distribution of radioactivity in the precipitate. The sequence of precipitation during evaporation, which is related to the solubility of the dissolved compounds, also influences self-absorption (Zikovsky, 2000). During our experiments, the typical surface density for the ISO 10704 method varied in the range of 0.8– 1.3 mg/cm2. The self absorption factor for gross alpha counting was 0.927 0.03 at 0.5 mg/cm2 and 0.57 70.02 at 1.5 mg/cm2.
3.3.2. Time delay and radon ingrowth Radon ingrowth can result in variations of count rate and measurement result. Many natural waters contain 226Ra at up to several Bq/L concentration. From a measurement point of view, progenies of 226Ra play a significant role in gross alpha/beta measurements since they are produced from 226Ra continuously after the source preparation and contribute to the gross count rates which can vary as a function of time. A measurement started immediately after source preparation, as required in ISO standards, renders lower gross alpha activity than a delayed measurement as shown in Fig. 3. Careful control of timing, such as the elapsed time between source preparation and start of measurement and measurement duration, are absolutely required. One approach is to wait for secular equilibrium between 226 Ra/222Rn and its progenies ( 30 days) before measurement. To reach equilibrium with certainty is, however, difficult to assess, since 222Rn can escape from the source itself. To quantify the escaped/emanated radon fraction (i.e. the level of equilibrium) is difficult because the microstructure and density of the dry residue influence the emanation rate. Parsa (1998) pointed out that short-lived 224Ra (T1/2 ¼3.64 d) can remain undetected because of the time delays between sampling and routine gross alpha analysis. In order to assess the 224 Ra contribution, the gross alpha activity should be performed as
soon as possible after the time of sampling. In addition, the duration of the drying step may vary, leading to an unknown increase in gross count rates. 3.3.3. Calibration Since determination of self-absorption is a crucial step, calibration standards must be appropriate for the measurement samples (e.g. electrodeposited standard sources cannot be used for calibration of dry precipitate samples). The counting efficiency of certain counters shows energy dependence (ISO 9696; Semkow et al., 2004; Montaña et al., 2012). Therefore radionuclides used for efficiency calibration have to be carefully selected. Finally, spatial variation of the counting efficiency may also occur for proportional counters as shown in Fig. 4, which means that a calibration source with the same diameter as the sample should be used and both sources must be positioned relative to the counter in the same way. 3.3.4. Moisture Moisture has to be controlled and kept constant during the measurement to maintain stable measurement conditions. Some of the participants did not take this into account. Applying the direct evaporation approach, the adsorption of moisture from ambient air can be comparable (up to 40%) to the net mass of dry residue. If nitrates are still present, most dry residues are hygroscopic. Mass change of the sample can continue during measurement affecting significantly the self absorption. 3.3.5. Sources of bias and uncertainties Details on sources of uncertainties for gross alpha/beta measurements are given in Table 2. 40 35 30
Alpha count rate (cpm)
432
25 20 15 10 5 -3
-2
-1
0
0 1 Distance from the detector centre (cm)
2
3
Fig. 4. Spatial variation of the alpha count rate for a proportional counter (error bars represent the uncertainty from the counting statistics, with k ¼ 2).
Alpha count rate (cpm)
6.0 5.5
Table 2 Uncertainty budget for ISO 10704 gross alpha/beta activity concentration measurements in water samples (Water A-B-C), giving the standard uncertainties (1 s) for a single measurement of a single sample (5 h counting time) measured at IRMM.
5.0 4.5
y = 0.036x + 2.317
4.0
R2 = 0.967
Uncertainty component
Gross alpha (%)
Gross beta (%)
Counting statistics (min–max) Counting efficiency/Self absorption Volume of the test sample Calibration source Weighing Chemical yield Partialn combined standard uncertainty (uc) (min–max)
2.5–12.4 4.5 0.5 1 0.2 Not known 5.3–13.2
1.8–6.5 3.1 0.5 1 0.2 Not known 3.8–7.3
3.5 3.0 2.5 2.0
0
20
40
60
80
100
Elapsed time (hours) Fig. 3. Variation of alpha count rate due to the ingrowth of 226Ra progenies in the co-precipitated source (error bars represent the uncertainty from the counting statistics, with k ¼ 2).
♯
Excluding chemical yield.
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
Not unexpectedly, the major contributions are counting statistics and self absorption. The uncertainty due to sample preparation (sample volume, weighing) contributes only 1.1% to the combined standard uncertainty. There can be a significant bias due to the variation of counting efficiency as a function of alpha particle energies and from the fitting of the self-absorption curve. The bias from the counting efficiency can be up to 75% (ISO 9696), making the appropriate calibration and determination of selfabsorption absolutely crucial. We obtained approximately 20% bias using different electrodeposited and drop deposited calibration sources (for alpha energies of 4 MeV to 7 MeV). However, positioning of these sources was more important as shown in Fig. 4. The uncertainty from chemical yield cannot be quantified, since the determination of the chemical yield itself is difficult in the case of the co-precipitation approach of ISO 10704. Due to chemical manipulations, the yield can never be assumed to be 100%. Loss of precipitate occurs during the filtration step and some precipitate can adsorb to the walls of the glassware as was observed in these experiments.
433
are analyzed. It should be noted, however, that there is one parameter that cannot be fixed as analysts simply cannot control it: the original radiochemical composition of the samples. Replacing gross methods of drinking water analysis with radionuclide-specific methods would not take a lot of effort or expense. Radionuclide measurements can be done using the same instrumentation (gas-flow proportional counter, liquid—and solid scintillation counter) that are used for gross alpha/beta analysis. Only more expertise/proficiency and validated methods are needed, but in a routine radiochemistry laboratory they should already exist. Routine laboratory measurement protocols must be up-dated to the new food safety-, health- and environmentrelated challenges where rapid, specific and accurate responses are needed. There are simultaneous methods that allow performing multi-radionuclide analysis focusing on the main naturally occurring radionuclides of interest (226Ra, 228Ra, 210Po, 210Pb, and the uranium isotopes) with a comparable speed to gross alpha/ beta methods such as described by Chalupnik and Lebecka, 1992; Vajda et al., 1997; Schönhofer and Wallner, 2001; Eikenberg et al., 2004; Benedik et al. 2009; Jia et al., 2009.
4. Conclusions Acknowledgments 4.1. Recommendations for the gross alpha/beta method applied to drinking water analysis It was shown that gross methods are not as simple as usually stated (Montaña et al., 2012; Semkow et al., 2004), and they are far from accurate. In some cases they fail to determine certain radionuclides, therefore they give only an “activity index” rather than an approximate activity concentration, as shown by Schönhofer (2012) and confirmed by the data spread from this laboratory comparison. The difference between laboratory results in this ILC is sometimes two or three orders of magnitude, which is far beyond the measurement uncertainties. These findings lead us to conclude that gross alpha/beta methods are not fit to be used as an independent method to assess activity concentration. Gross measurement should be used for monitoring only after the radionuclide composition is known from radionuclide specific analysis of representative samples. It can be used as a complementary or substitute method for radionuclide-specific measurement only with important restrictions: (1) no temporary change is expected in the radiochemical composition (no significant ingrowth of progenies during the measurement); (2) no complex decay chains are present; (3) a true standardized method is used; and (4) the measurement parameters are fixed. Radionuclide specific analysis should be repeated on a regular basis in accordance with the drinking water directive concerning check and audit monitoring (EC, 1998). Any suspected change in parameters requires more frequent nuclide specific analysis. 4.2. True standardization of methods The main reason for unreliability of gross methods and the noncomparability of gross alpha/beta measurement results is the lack of knowledge of the real radionuclide composition of the water. Since there are many variables playing a key role in gross measurement (Arndt and West, 2004; Semkow et al., 2004; Jobbágy et al, 2010; Montaña et al., 2012), it is important to fix as many parameters as possible. The radionuclides used in the calibration, the geometry of the source, quenching parameters, the chemical form and any time delay must be clearly defined. Acceptable time delays for each step (e.g. between sampling and sample preparation, source preparation and start of measurement) must be set in the written measurement standards, which is particularly important when 226Ra- and 224Ra-containing waters
The authors wish to thank Pieter Kwakman (RIVM, The Netherlands), Edmond Dupuis (SCKCEN, Belgium), Timotheos Altzitzoglou, Mikael Hult and Andrej Rozkov (EC-JRC-IRMM, Belgium) for their contributions to the reference value determination. References Arndt, M.F., West, L.E., 2004. A study of the factors affecting the gross alpha measurement, and a radiochemical analysis of some groundwater samples from the state of Wisconsin exhibiting an elevated gross alpha activity. In: Proceedings of the 51st Annual Radiobioassay and Radiochemical Measurements Conference, Cincinatti, October 30–November 4. Benedik, L., Vasile, M., Spasova, Y., Watjen, U., 2009. Sequential determination of 210 Po and uranium radioisotopes in drinking water by alpha-particle spectrometry. Appl. Radiat. Isot 67, 770–775. Chalupnik, S., Lebecka, J., 1992. Determination of 226Ra, 228Ra and 224Ra in water and aqueous solutions by liquid scintillation counting. In: Proceedings of Liquid Scintillation Conference on Radiocarbon, Vienna, 14–18 September 1992, 397–403. Eikenberg, J., Bajo, S., Beer, H., Hitz, J., Ruethi, M., Zumsteg, I., Letessier, P., 2004. Fast methods for determination of antropogenic actinides and U/Th-series isotopes in aqueous samples. Appl. Radiat. Isot 61, 101–106. EC, 1998. European Union Council Directive 98/83/EC of 3 November 1998 on the Quality of Water Intended for Human Consumption. EC, 2012. COM/2012/147 Final of 28 March 2012: Proposal for a Council Directive laying Down Requirements for the Protection of the Health of the General Public with Regard to Radioactive Substances in Water Intended for Human Consumption. Gruber, V., Maringer, F.J., Landstetter, C., 2009. Radon and other natural radionuclides in drinking water in Austria: measurement and assessment. Appl. Radiat. Isot 67, 913–917. ISO 9696, 2007. Water Quality: Measurement of Gross Alpha Activity in Non-Saline Water—Thick Source Method. International Organization for Standardization, Geneva. ISO 9697, 2008. Water Quality: Measurement of Gross Beta Activity in Non-Saline Water. International Organization for Standardization, Geneva. ISO 10704, 2009. Water quality: Measurement of Gross Alpha and Gross Beta Activity in Non-saline Water—Thin Source Deposit Method. International Organization for Standardization, Geneva. ISO 11704, 2010. Water Quality: measurement of Gross Alpha and Beta Activity Concentration in Non-saline Water—Liquid Scintillation Counting Method. International Organization for Standardization, Geneva. Jia, G., Torri, G., Magro, L., 2009. Concentrations of 238U, 234U, 235U, 232Th, 230Th, 228 Th, 226Ra, 228Ra, 224Ra, 210Po, 210Pb and 212Pb in drinking water in Italy: reconciling safety standards based on measurements of gross α and β. J. Environ. Radioact 100, 941–949. Jobbágy, V., Wätjen, U., Merešova, J., 2010. Current status of gross alpha/beta activity analysis in water samples: a short overview of methods. J. Radioanal. Nucl. Chem. 286, 393–399. Matthews, K.M., Kim, C.K., Martin, P., 2007. Determination of 210Po in environmental materials: a review of analytical methodology. Appl. Radiat. Isot 65, 267–279.
434
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
Momoshima, N., Song, L.X., Osaki, S., Maeda, Y., 2002. Biologically induced Po emission from fresh water. J. Environ. Radioact. 63, 187–197. Montaña, M., Camacho, A., Vallés, I., Serrano, I., 2012. Experimental analysis of the mass efficiency curve for gross alpha activity and morphological study of the residue obtained by the co-precipitation methodt. Appl. Radiat. Isot 70, 1541–1548. Parsa, B., 1998. Contribution of short-lived radionuclides to alpha-particle radioactivity in drinking water and their impact on the Safe Drinking Water Act Regulations. Radioact. Radiochem 9, 41–50. Rusconi, R., Forte, M., Caresana, M., Bellinzona, S., Cazzaniga, M.T., Sgorbati, G., 2006. The evaluation of uncertainty in low-level LSC measurements of water samples. Appl. Radiat. Isot 64, 1124–1129. Salonen, L., 2010. Comparison of two direct LS methods for measuring 222Rn in drinking water using α/Β liquid scintillation spectrometry. Appl. Radiat. Isot 68, 1970–1979. Schönhofer, F., 1995. Liquid scintillation spectrometry in environmental measurements. Sci. Total Environ. 173-174, 29–40. Schönhofer, F., Wallner, G., 2001. Very rapid determination of Ra-226, Ra-228, and Pb-210 by selective adsorption and liquid scintillation spectrometry. Radioact. Radiochem 12, 33–38. Schönhofer, F., 2012. An LSC approach to improve (???) or better replace gross alpha and/or beta measurements. In: III. Terrestrial Radionuclides in
Environment International Conference on Environmental Protection, Veszprém (Hungary), 16–18 May. Semkow, T.M., Bari, A., Parekh, P.P., Haines, D.K., Gao, H., Bolden, A.N., Dahms, K.S., Scarpitta, S.C., Thern, R.E., Velazquez, S., 2004. Experimental investigation of mass efficiency curve for alpha radioactivity counting using a gas-proportional detector. Appl. Radiat. Isot 60, 879–886. Suarez-Navarro, J.A., Pujol, L., de Pablo, M.A., 2002. Rapid determination of gross alpha-activity in sea water by coprecipitation. J. Radioanal. Nucl. Chem. 253, 47–52. UNSCEAR, 2000. United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), Sources and Effects of Ionising Radiation. United Nations, New York. Vajda, N., LaRosa, J., Zeisler, R., Danesi, P., Kis-Benedek, G., 1997. A novel technique for the simultaneous determination of 210Pb and 210Po using a crown ether. J. Environ. Radioact. 37, 355–372. WHO, 2011. World Health Organization, Guidelines for Drinking-water Quality, fourth ed. World Health Organization, Geneva. Zikovsky, L., 2000. Uranyl ions, nitrate ions and some major cations as causes of anomalous results when determining gross alpha radioactivity in ground water. Radioact. Radiochem 11, 26–28.
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
Critical remarks on gross alpha/beta activity analysis in drinking waters: Conclusions from a European interlaboratory comparison V. Jobbágy n, J. Merešová, U. Wätjen European Commission, Joint Research Centre, Institute for Reference Materials and Measurements, Retieseweg 111, B-2440 Geel, Belgium
H I G H L I G H T S
Gross alpha/beta standard methods for drinking water analysis are discussed. Large spread of results (up to 2 orders of magnitude) observed in comparisons. Sources of interferences are reviewed. We propose to use true standardized methods to obtain better measurement results.
art ic l e i nf o
a b s t r a c t
Available online 1 December 2013
The most common gross alpha/beta standard methods used for drinking water analysis are discussed, and sources of interferences are reviewed from a metrological point of view. Our study reveals serious drawbacks of gross methods on the basis of an interlaboratory comparison analyzing commercial mineral water samples with the participation of 71 laboratories. A proposal is made to obtain comparable measurement results using true standardized methods. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Gross alpha/beta activity Drinking water Interlaboratory comparison Radioactivity Limits of gross methods
1. Introduction Gross alpha/beta activity measurement is applied widely as a screening technique in the field of radioecology, environmental monitoring and industrial applications. Water intended for drinking purposes has to be analyzed first for gross alpha/beta activity content according to many different national and international standards and recommendations. Anticipating the new European Union (EU) Drinking Water Directive (EC, 2012) which incorporates gross alpha/beta activity screening levels, the Institute for Reference Materials and Measurements (IRMM) organized an interlaboratory comparison (ILC) to check the fitness for purpose of this method and the performance of European monitoring laboratories. On the basis of the reported values from the 71 participating laboratories, we review some of the most influential parameters on gross measurements. The claimed main advantages of the gross alpha/beta methods are the relatively low costs, rapidity and simplicity. Although it is one of the simplest, it is also one of the most disputed radioanalytical methods because the determination of gross alpha and beta activities
n Corresponding author. Present address: SCKCEN, Belgian Nuclear Research Centre, Boeretang 200, B-2400 Mol, Belgium. Tel.: þ 32 14 33 32 38. E-mail address: [email protected] (V. Jobbágy).
0969-8043/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apradiso.2013.11.073
faces some specific problems that may refute the afore-mentioned claims. There are many sources of interferences in gross alpha/beta measurement that may corrupt the comparability of the measurement results (Arndt and West, 2004; Rusconi et al., 2006; Semkow et al., 2004; Montaña et al., 2012). First one is related to the radionuclide composition of the sample. During gross alpha/beta activity measurement, a mixed radionuclide composition must be simultaneously measured. Drinking water samples may contain different naturally occurring alpha (238U, 234U, 232Th, 226Ra and 210Po) and beta (40K, 228Ra and 210Pb) emitters, and artificial radionuclides (241Am, 90Sr) in various concentrations (UNSCEAR, 2000). Moreover, most of these are members of a complex decay chain, therefore the ingrowths of the daughter products influence the measurement result. The second important source of interference is due to the final source thickness that causes self absorption of the emitted particles already in the source itself. In this respect, it is crucial to use standardized methods (Jobbágy et al., 2010). The WHO recommends that, "Where possible, standardized methods should be used to determine concentrations of gross alpha and beta activities" (WHO, 2011). For this reason, the most common standard methods – based on direct evaporation, co-precipitation and liquid scintillation counting – are discussed in this paper with respect to the sample preparation and counting techniques. Regarding the measurement methods, the influence of the following parameters must be considered: counting efficiency, self absorption, moisture absorption,
430
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
chemical recovery and the interferences due to the isotopic composition of the water sample. Experimental comparisons of the gross alpha/beta standard methods were done using real drinking water samples with different salinity and radionuclide activity concentration. Besides the pitfalls of the gross measurements, the paper also tries to give examples where gross measurements can be used as a rapid alternative technique to the radionuclide-specific analysis.
2. Materials and methods 2.1. Reagents All the chemicals (conc. H2SO4, FeCl3, BaCl2) were analytical grade and all the stock solutions were prepared using de-ionized water. Three water samples, with different salinity and radionuclide activity concentrations, were used for the described experiments and in the interlaboratory comparison. Two water samples (Water A and B) were commercially available natural mineral waters. The third one (Water C) was prepared by spiking de-ionized water with a known activity of 241Am and 90Sr/90Y solutions standardized at IRMM by liquid scintillation counting using the CIEMAT/NIST method. The total dissolved solid was set to a final total concentration of about 10 mg/L by adding the following inactive inorganic salts: NaCl, CaCl2 and Sr(NO3)2. The major sources of gross alpha activity for Water A are 234U and 238U ( 40 mBq/L each), for Water B this is 226Ra ( 330 mBq/L) and for Water C 241Am ( 950 mBq/L).
2.2. Sample preparation Sources were prepared in accordance with ISO 9696/9697 (“thick source method”) or ISO 10704 (“thin source method”). The surface density of a source prepared under ISO 9696/9697 must exceed 10 mg/cm2. An aliquot of sample was evaporated to dryness, and the dried residue converted to sulfate form by sulfuric acid and ignited at 350 1C for an hour (ISO 9696, 2007; ISO 9697, 2008). The surface density of the source in the thin source method must be below 5 mg/cm2, and has been described previously (ISO 10704, 2009; Montaña et al, 2012; Suarez-Navarro et al., 2002). The pH of the filtered water sample was adjusted with sulfuric acid and heated to purge radon and CO2. Then the radium isotopes were co-precipitated as Ba(Ra)SO4, whereas uranium, thorium and polonium isotopes are co-precipitated with Fe(OH)3 after pH adjustment (pH E7–8) (ISO 10704, 2009; Suarez-Navarro et al., 2002).
2.3. Gross counting system and activity calculation For the gross alpha/beta measurement, a 10-detector, lowbackground gas-flow proportional counting system was used. The high voltage was set to 1450 V and the counting gas (Ar/ CH4, 90/10) flow was kept stable with a flow rate of 25 mL/min. The gross alpha/beta count of the filtered/dried precipitate was measured in 5 h cycles repeating it several times (from three to twelve cycles). For the counting efficiency and self absorption experiments, 241 Am and 90Sr standard solutions were used, since these are the most frequently used radionuclides for this purpose. Gross alpha/ beta activity concentrations were calculated from count rates by following the formulae of the corresponding ISO standard (ISO 9696, 2007; ISO 9697, 2008; ISO 10704, 2009; ISO 11704, 2010).
3. Results and discussion 3.1. Summary of the interlaboratory exercise After radiochemical characterization of the three water samples at IRMM, they were sent to the European monitoring laboratories for gross alpha/beta analysis. The gross alpha/beta reference values were determined independently from the laboratory comparison from the results determined by the three laboratories involved in the reference value determination (Table 1). Water C is, in principle, the easiest sample to measure since its gross alpha/beta activity concentration is the highest among the ILC samples. However, from a measurement point of view, the gross alpha activity is not the only key factor, but the alpha/beta emitting radionuclides and the total dissolved solid content have to be considered as well. Taking into account all three factors one can make an order of difficulty in terms of measurement as follows: Water C oWater A rWater B. As is evident from the reported results (Fig. 1 and 2), the outcome of the laboratory comparison exercise is far from satisfactory. The measurement results span a wide range, e.g. for Water C the maximum reported gross beta activity was more than 3000 times higher than the minimum reported gross beta activity. Furthermore, several laboratories (no. 49, 50 etc.) present for one type of sample a measurement result several times higher than the reference, whilst for another type of sample the same laboratory has a result several times lower. The gross alpha activity results for Water C (spiked sample, i.e. the best case scenario) are sorted according to the applied counting techniques in Fig. 2. For Water A and B (natural waters) we could observe a similar data spread but with a higher degree of variation for each technique. Possible reasons for the diversity of results are given in the next sections, but a detailed evaluation of the laboratory comparison itself will be published later in a separate report. 3.2. Method comparison It is well-known that the sample form or geometry play a significant role in producing reliable results from thicker sources, requiring a uniform thickness and a homogeneous layer of residue material on a planchet (60 mm in diameter). The surface density, influencing the self absorption of alpha and beta particles, should be controlled and determined very carefully (ISO 9696/9697; ISO 10704; Montaña et al., 2012; Suarez-Navarro et al., 2002; Semkow et al., 2004; Parsa, 1998). A surface density higher than 10 mg/cm2 (ISO 9696/9697) achieves satisfying counting statistics and constant self absorption. The self absorption of the alpha and beta particles limits their counting efficiency to usually less than 50%. It should be mentioned that no energy resolution is possible with proportional counters. Since the sample is evaporated and later heat-treated, ISO 9696/9697 does not allow determination of the volatile radionuclides (e.g. 3H, 210Po, 137Cs), which escape from the sample and Table 1 The reference activity concentration values (Aref) in the ILC water samples with their expanded uncertainties (Uref) with a coverage factor k ¼ 2. Reference values determined by three laboratories external to the laboratory comparison. Parameter
Reference value with expanded uncertainty Aref 7 Uref (mBq/L) Water A
Gross alpha activity 47.5 7 22.8 Gross beta activity 309.8 7 57.4 Total dissolved solids (mg/L) 955 7 44
Water B
Water C
434.7 7 56.6 190.4 7 32.6 3647 27
954.5 7 77.3 1037.3 7 83.0 10.2 7 0.1
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
431
Fig. 2. Laboratory results of gross alpha activity concentration sorted by techniques for Water C. The solid line indicates the reference activity concentration (Aref) of gross alpha activity. Its corresponding expanded uncertainty 7 Uref (k ¼2) is plotted as dashed lines.
determination possible, no conclusion can be drawn on the chemical recovery. It should be stressed that in the case of real water samples it is difficult to quantify this without further analysis as the original chemical composition is not always known accurately. From the chemical composition of the sample, the excess precipitate formed from the original inorganic cations could be quantified. Second, 40K cannot be included in the gross beta result because it stays in the final supernatant. For this reason, the gross beta results obtained by co-precipitation methods are always lower than the true gross beta values. The 40K content has to be separately determined with either a radiometric method (low background gamma-ray spectrometry) or with a non-radiometric method (emission flame photometry).
3.2.1. Liquid scintillation counting (ISO 11704) Ultra low-level α/β-discrimination LSC, due to its high detection efficiency (up to 100%) and low background count rate, is a useful tool for the determination of alpha- and beta emitting radionuclides (Schönhofer, 1995). The basic steps of the sample preparation for liquid scintillation counting are very simple. It consists of a thermal pre-concentration, pH adjustment and sample mixing with scintillation cocktail (ISO 11704, 2010). Advantages are: Fig. 1. Laboratory results for gross alpha/beta activity concentration for different water samples. The solid lines indicate the reference activity concentrations (Aref) of gross alpha and gross beta activity, respectively. Their corresponding expanded uncertainties 7 Uref (k ¼2) are plotted in dashed lines.
the residue during heat treatment. Therefore, the gross alpha and beta activity concentration might be underestimated. For instance, 210 Po losses, varying with the chemical form of the element, have been reported beginning above 100 1C (Momoshima et al., 2002; Matthews et al., 2007). However, ISO 9696 claims that losses of polonium are not expected for samples “which have been acidified with nitric acid and subjected to sulfation” and treated at 350 1C (nitrate salts have been converted to the sulfate form). Some of these drawbacks of the thick source approach can be avoided by using co-precipitation of the water sample (SuarezNavarro et al., 2002; ISO 10704). On one hand, this method needs more chemical treatment than the evaporation methods. On the other, more uniform and homogeneous residues can be obtained, especially for higher salt concentrations (e.g. in mineral waters). However, it suffers from two very severe drawbacks. Firstly, since neither is a radiotracer applied nor is gravimetric yield
no self absorption occurs in the sample, LSC methods can be adapted to monitor low-energy beta
emitters such as 3H and 14C (Schönhofer, 1995; Gruber et al., 2009), and radon in water can be determined separately using liquid scintillation methods (Salonen, 2010).
One of the disadvantages is quenching (chemical, color and physical) which reduces the counting efficiency; therefore, a quench correction must be made. Furthermore, the separation of beta energies from all alpha energies can be difficult, because it depends on many factors in LSC counting (Rusconi et al., 2006; Schönhofer, 2012). For this reason, the optimal setting of the different parameters (type of the vial, cocktail, α/β pulse discrimination, counting efficiency) is essential for gross alpha/beta measurements by LSC. Many other factors like the physical or chemical behaviour of the radionuclides, dissolved ions and the energies of the emitted particles may influence the detection, too. As stated by Rusconi et al. (2006) the same settings should be used during measurement and calibration.
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
3.3. Sources of interferences 3.3.1. Self absorption Most gross alpha/beta standard methods apply proportional counter or solid state scintillation counters as detector (ISO 9696, 2007; ISO 9697, 2008; ISO 10704, 2009). The counting efficiency of these methods is strongly affected by the total dissolved solids and the surface density of the sample. The direct evaporation approach of ISO 10704 allows the use of a binding agent that contains a polymer “to produce evenly spread counting sources” (ISO 10704, 2009). This polymer stays on the planchet, increasing the surface density which cannot exceed 5 mg/cm2. To get equivalent alpha/ beta self-absorption for the sample and the calibration source in a multi-component microchemical environment is not easy to accomplish. The sample and the calibration source must be as similar as possible in terms of surface density, homogeneity, and distribution of radioactivity in the precipitate. The sequence of precipitation during evaporation, which is related to the solubility of the dissolved compounds, also influences self-absorption (Zikovsky, 2000). During our experiments, the typical surface density for the ISO 10704 method varied in the range of 0.8– 1.3 mg/cm2. The self absorption factor for gross alpha counting was 0.927 0.03 at 0.5 mg/cm2 and 0.57 70.02 at 1.5 mg/cm2.
3.3.2. Time delay and radon ingrowth Radon ingrowth can result in variations of count rate and measurement result. Many natural waters contain 226Ra at up to several Bq/L concentration. From a measurement point of view, progenies of 226Ra play a significant role in gross alpha/beta measurements since they are produced from 226Ra continuously after the source preparation and contribute to the gross count rates which can vary as a function of time. A measurement started immediately after source preparation, as required in ISO standards, renders lower gross alpha activity than a delayed measurement as shown in Fig. 3. Careful control of timing, such as the elapsed time between source preparation and start of measurement and measurement duration, are absolutely required. One approach is to wait for secular equilibrium between 226 Ra/222Rn and its progenies ( 30 days) before measurement. To reach equilibrium with certainty is, however, difficult to assess, since 222Rn can escape from the source itself. To quantify the escaped/emanated radon fraction (i.e. the level of equilibrium) is difficult because the microstructure and density of the dry residue influence the emanation rate. Parsa (1998) pointed out that short-lived 224Ra (T1/2 ¼3.64 d) can remain undetected because of the time delays between sampling and routine gross alpha analysis. In order to assess the 224 Ra contribution, the gross alpha activity should be performed as
soon as possible after the time of sampling. In addition, the duration of the drying step may vary, leading to an unknown increase in gross count rates. 3.3.3. Calibration Since determination of self-absorption is a crucial step, calibration standards must be appropriate for the measurement samples (e.g. electrodeposited standard sources cannot be used for calibration of dry precipitate samples). The counting efficiency of certain counters shows energy dependence (ISO 9696; Semkow et al., 2004; Montaña et al., 2012). Therefore radionuclides used for efficiency calibration have to be carefully selected. Finally, spatial variation of the counting efficiency may also occur for proportional counters as shown in Fig. 4, which means that a calibration source with the same diameter as the sample should be used and both sources must be positioned relative to the counter in the same way. 3.3.4. Moisture Moisture has to be controlled and kept constant during the measurement to maintain stable measurement conditions. Some of the participants did not take this into account. Applying the direct evaporation approach, the adsorption of moisture from ambient air can be comparable (up to 40%) to the net mass of dry residue. If nitrates are still present, most dry residues are hygroscopic. Mass change of the sample can continue during measurement affecting significantly the self absorption. 3.3.5. Sources of bias and uncertainties Details on sources of uncertainties for gross alpha/beta measurements are given in Table 2. 40 35 30
Alpha count rate (cpm)
432
25 20 15 10 5 -3
-2
-1
0
0 1 Distance from the detector centre (cm)
2
3
Fig. 4. Spatial variation of the alpha count rate for a proportional counter (error bars represent the uncertainty from the counting statistics, with k ¼ 2).
Alpha count rate (cpm)
6.0 5.5
Table 2 Uncertainty budget for ISO 10704 gross alpha/beta activity concentration measurements in water samples (Water A-B-C), giving the standard uncertainties (1 s) for a single measurement of a single sample (5 h counting time) measured at IRMM.
5.0 4.5
y = 0.036x + 2.317
4.0
R2 = 0.967
Uncertainty component
Gross alpha (%)
Gross beta (%)
Counting statistics (min–max) Counting efficiency/Self absorption Volume of the test sample Calibration source Weighing Chemical yield Partialn combined standard uncertainty (uc) (min–max)
2.5–12.4 4.5 0.5 1 0.2 Not known 5.3–13.2
1.8–6.5 3.1 0.5 1 0.2 Not known 3.8–7.3
3.5 3.0 2.5 2.0
0
20
40
60
80
100
Elapsed time (hours) Fig. 3. Variation of alpha count rate due to the ingrowth of 226Ra progenies in the co-precipitated source (error bars represent the uncertainty from the counting statistics, with k ¼ 2).
♯
Excluding chemical yield.
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
Not unexpectedly, the major contributions are counting statistics and self absorption. The uncertainty due to sample preparation (sample volume, weighing) contributes only 1.1% to the combined standard uncertainty. There can be a significant bias due to the variation of counting efficiency as a function of alpha particle energies and from the fitting of the self-absorption curve. The bias from the counting efficiency can be up to 75% (ISO 9696), making the appropriate calibration and determination of selfabsorption absolutely crucial. We obtained approximately 20% bias using different electrodeposited and drop deposited calibration sources (for alpha energies of 4 MeV to 7 MeV). However, positioning of these sources was more important as shown in Fig. 4. The uncertainty from chemical yield cannot be quantified, since the determination of the chemical yield itself is difficult in the case of the co-precipitation approach of ISO 10704. Due to chemical manipulations, the yield can never be assumed to be 100%. Loss of precipitate occurs during the filtration step and some precipitate can adsorb to the walls of the glassware as was observed in these experiments.
433
are analyzed. It should be noted, however, that there is one parameter that cannot be fixed as analysts simply cannot control it: the original radiochemical composition of the samples. Replacing gross methods of drinking water analysis with radionuclide-specific methods would not take a lot of effort or expense. Radionuclide measurements can be done using the same instrumentation (gas-flow proportional counter, liquid—and solid scintillation counter) that are used for gross alpha/beta analysis. Only more expertise/proficiency and validated methods are needed, but in a routine radiochemistry laboratory they should already exist. Routine laboratory measurement protocols must be up-dated to the new food safety-, health- and environmentrelated challenges where rapid, specific and accurate responses are needed. There are simultaneous methods that allow performing multi-radionuclide analysis focusing on the main naturally occurring radionuclides of interest (226Ra, 228Ra, 210Po, 210Pb, and the uranium isotopes) with a comparable speed to gross alpha/ beta methods such as described by Chalupnik and Lebecka, 1992; Vajda et al., 1997; Schönhofer and Wallner, 2001; Eikenberg et al., 2004; Benedik et al. 2009; Jia et al., 2009.
4. Conclusions Acknowledgments 4.1. Recommendations for the gross alpha/beta method applied to drinking water analysis It was shown that gross methods are not as simple as usually stated (Montaña et al., 2012; Semkow et al., 2004), and they are far from accurate. In some cases they fail to determine certain radionuclides, therefore they give only an “activity index” rather than an approximate activity concentration, as shown by Schönhofer (2012) and confirmed by the data spread from this laboratory comparison. The difference between laboratory results in this ILC is sometimes two or three orders of magnitude, which is far beyond the measurement uncertainties. These findings lead us to conclude that gross alpha/beta methods are not fit to be used as an independent method to assess activity concentration. Gross measurement should be used for monitoring only after the radionuclide composition is known from radionuclide specific analysis of representative samples. It can be used as a complementary or substitute method for radionuclide-specific measurement only with important restrictions: (1) no temporary change is expected in the radiochemical composition (no significant ingrowth of progenies during the measurement); (2) no complex decay chains are present; (3) a true standardized method is used; and (4) the measurement parameters are fixed. Radionuclide specific analysis should be repeated on a regular basis in accordance with the drinking water directive concerning check and audit monitoring (EC, 1998). Any suspected change in parameters requires more frequent nuclide specific analysis. 4.2. True standardization of methods The main reason for unreliability of gross methods and the noncomparability of gross alpha/beta measurement results is the lack of knowledge of the real radionuclide composition of the water. Since there are many variables playing a key role in gross measurement (Arndt and West, 2004; Semkow et al., 2004; Jobbágy et al, 2010; Montaña et al., 2012), it is important to fix as many parameters as possible. The radionuclides used in the calibration, the geometry of the source, quenching parameters, the chemical form and any time delay must be clearly defined. Acceptable time delays for each step (e.g. between sampling and sample preparation, source preparation and start of measurement) must be set in the written measurement standards, which is particularly important when 226Ra- and 224Ra-containing waters
The authors wish to thank Pieter Kwakman (RIVM, The Netherlands), Edmond Dupuis (SCKCEN, Belgium), Timotheos Altzitzoglou, Mikael Hult and Andrej Rozkov (EC-JRC-IRMM, Belgium) for their contributions to the reference value determination. References Arndt, M.F., West, L.E., 2004. A study of the factors affecting the gross alpha measurement, and a radiochemical analysis of some groundwater samples from the state of Wisconsin exhibiting an elevated gross alpha activity. In: Proceedings of the 51st Annual Radiobioassay and Radiochemical Measurements Conference, Cincinatti, October 30–November 4. Benedik, L., Vasile, M., Spasova, Y., Watjen, U., 2009. Sequential determination of 210 Po and uranium radioisotopes in drinking water by alpha-particle spectrometry. Appl. Radiat. Isot 67, 770–775. Chalupnik, S., Lebecka, J., 1992. Determination of 226Ra, 228Ra and 224Ra in water and aqueous solutions by liquid scintillation counting. In: Proceedings of Liquid Scintillation Conference on Radiocarbon, Vienna, 14–18 September 1992, 397–403. Eikenberg, J., Bajo, S., Beer, H., Hitz, J., Ruethi, M., Zumsteg, I., Letessier, P., 2004. Fast methods for determination of antropogenic actinides and U/Th-series isotopes in aqueous samples. Appl. Radiat. Isot 61, 101–106. EC, 1998. European Union Council Directive 98/83/EC of 3 November 1998 on the Quality of Water Intended for Human Consumption. EC, 2012. COM/2012/147 Final of 28 March 2012: Proposal for a Council Directive laying Down Requirements for the Protection of the Health of the General Public with Regard to Radioactive Substances in Water Intended for Human Consumption. Gruber, V., Maringer, F.J., Landstetter, C., 2009. Radon and other natural radionuclides in drinking water in Austria: measurement and assessment. Appl. Radiat. Isot 67, 913–917. ISO 9696, 2007. Water Quality: Measurement of Gross Alpha Activity in Non-Saline Water—Thick Source Method. International Organization for Standardization, Geneva. ISO 9697, 2008. Water Quality: Measurement of Gross Beta Activity in Non-Saline Water. International Organization for Standardization, Geneva. ISO 10704, 2009. Water quality: Measurement of Gross Alpha and Gross Beta Activity in Non-saline Water—Thin Source Deposit Method. International Organization for Standardization, Geneva. ISO 11704, 2010. Water Quality: measurement of Gross Alpha and Beta Activity Concentration in Non-saline Water—Liquid Scintillation Counting Method. International Organization for Standardization, Geneva. Jia, G., Torri, G., Magro, L., 2009. Concentrations of 238U, 234U, 235U, 232Th, 230Th, 228 Th, 226Ra, 228Ra, 224Ra, 210Po, 210Pb and 212Pb in drinking water in Italy: reconciling safety standards based on measurements of gross α and β. J. Environ. Radioact 100, 941–949. Jobbágy, V., Wätjen, U., Merešova, J., 2010. Current status of gross alpha/beta activity analysis in water samples: a short overview of methods. J. Radioanal. Nucl. Chem. 286, 393–399. Matthews, K.M., Kim, C.K., Martin, P., 2007. Determination of 210Po in environmental materials: a review of analytical methodology. Appl. Radiat. Isot 65, 267–279.
434
V. Jobbágy et al. / Applied Radiation and Isotopes 87 (2014) 429–434
Momoshima, N., Song, L.X., Osaki, S., Maeda, Y., 2002. Biologically induced Po emission from fresh water. J. Environ. Radioact. 63, 187–197. Montaña, M., Camacho, A., Vallés, I., Serrano, I., 2012. Experimental analysis of the mass efficiency curve for gross alpha activity and morphological study of the residue obtained by the co-precipitation methodt. Appl. Radiat. Isot 70, 1541–1548. Parsa, B., 1998. Contribution of short-lived radionuclides to alpha-particle radioactivity in drinking water and their impact on the Safe Drinking Water Act Regulations. Radioact. Radiochem 9, 41–50. Rusconi, R., Forte, M., Caresana, M., Bellinzona, S., Cazzaniga, M.T., Sgorbati, G., 2006. The evaluation of uncertainty in low-level LSC measurements of water samples. Appl. Radiat. Isot 64, 1124–1129. Salonen, L., 2010. Comparison of two direct LS methods for measuring 222Rn in drinking water using α/Β liquid scintillation spectrometry. Appl. Radiat. Isot 68, 1970–1979. Schönhofer, F., 1995. Liquid scintillation spectrometry in environmental measurements. Sci. Total Environ. 173-174, 29–40. Schönhofer, F., Wallner, G., 2001. Very rapid determination of Ra-226, Ra-228, and Pb-210 by selective adsorption and liquid scintillation spectrometry. Radioact. Radiochem 12, 33–38. Schönhofer, F., 2012. An LSC approach to improve (???) or better replace gross alpha and/or beta measurements. In: III. Terrestrial Radionuclides in
Environment International Conference on Environmental Protection, Veszprém (Hungary), 16–18 May. Semkow, T.M., Bari, A., Parekh, P.P., Haines, D.K., Gao, H., Bolden, A.N., Dahms, K.S., Scarpitta, S.C., Thern, R.E., Velazquez, S., 2004. Experimental investigation of mass efficiency curve for alpha radioactivity counting using a gas-proportional detector. Appl. Radiat. Isot 60, 879–886. Suarez-Navarro, J.A., Pujol, L., de Pablo, M.A., 2002. Rapid determination of gross alpha-activity in sea water by coprecipitation. J. Radioanal. Nucl. Chem. 253, 47–52. UNSCEAR, 2000. United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), Sources and Effects of Ionising Radiation. United Nations, New York. Vajda, N., LaRosa, J., Zeisler, R., Danesi, P., Kis-Benedek, G., 1997. A novel technique for the simultaneous determination of 210Pb and 210Po using a crown ether. J. Environ. Radioact. 37, 355–372. WHO, 2011. World Health Organization, Guidelines for Drinking-water Quality, fourth ed. World Health Organization, Geneva. Zikovsky, L., 2000. Uranyl ions, nitrate ions and some major cations as causes of anomalous results when determining gross alpha radioactivity in ground water. Radioact. Radiochem 11, 26–28.
