Evaluation of different parameters affecting the liquid scintillation spectrometry measurement of gross alpha and beta index in water samples

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ Esta es la versión aceptada del artículo publicado en: This is an accepted manuscript of a paper published in: Applied Radiation and Isotopes (2011): 26/09/2025 DOI: https://doi.org/10.1016/j.apradiso.2011.04.020 Copyright: © 2011 Elsevier Ltd. All rights reserved. El acceso a la versión publicada del artículo puede requerir la suscripción de la revista. Access to the published version may require subscription. “This is an Accepted Manuscript of an article published by Elsevier in [Applied Radiation and Isotopes] on [2011], available at https://doi.org/10.1016/j.apradiso.2011.04.020” 1 EVALUATION OF DIFFERENT PARAMETERS AFFECTING THE LIQUID SCINTILLATION SPECTROMETRY MEASUREMENT OF GROSS ALPHA AND BETA INDEX IN WATER SAMPLES M.Palomo1, M.Villa2, N.Casacuberta3, A.Peñalver1, F.Borrull1, C.Aguilar1* 1Unitat de Radioquímica Ambiental i Sanitària, Universitat Rovira i Virgili, Tarragona, Spain 2Centro de Investigación, Tecnología e Innovación. Servicio Radioisótopos. Universidad de Sevilla, Spain 3Institut de Ciència i Tecnologia Ambientals - Departament de Física, Universitat Autònoma de Barcelona. Spain *Corresponding author: [email protected] ABSTRACT Liquid scintillation spectrometry is a fast competitive technique for the simultaneous evaluation of gross alpha and beta indexes. However, the implementation of this technique should not be considered as straightforward and the pre-concentration methods to decrease the detection limit, together with quenching and alpha and beta crossover corrections, should be carefully chosen according to the needs of the laboratory. Both aspects are being approached in this work as to find an easy and robust method for alpha/beta measurement in water samples, taking into account quenching and alpha/beta crossover interferences effects. Results showed that most of the pre-concentration methods increased the quenching in the measurement, although HNO3 0.05M points to be the best solution for pre-concentration and re-dissolution of the sample as converges into low quenching and maximum recovery. Subsequently, in the measurement of water samples with different conductivities, the analysis of the raw counts to obtain gross alpha and beta indexes was carried out, using different approaches to implement quenching and interference corrections. If quenching and salt content in the sample are relatively low, interference and quenching efficiency corrections do not improve the accuracy of the results within the usual precision assumed for a result of gross alpha and beta index (25%). Special attention must be paid when corrections are applied to high quenched or saline samples and when alpha and beta activities values are different in several orders of magnitude. 2 Keywords: gross alpha, gross beta, salinity, liquid scintillation counting, quenching effects, interference corrections. 1. INTRODUCTION Gross alpha and gross beta determinations are indexes, which can provide a global radioactive content characterization in waters. However, these parameters only give a qualitative result but through their value we can know if further analyses of specific radionuclides are required [Rusconi et al. 2006]. As it is known, those indexes are not an absolute determination of the radioactivity content in samples, because alpha and/or beta efficiencies are energy dependent (since the radioactivity content of the sample is unknown, it is not possible to make a proper efficiency calculation). This is the reason why relative determinations are obtained referred to a specific alpha or beta emitter, which constitutes the standard calibration source. The radionuclides selected correspond to radioisotopes with energy of emissions coincident with the average energy of the radionuclides expected in the sample. Therefore, what we finally obtain is the average estimation of the alpha and/or beta content in the sample and the precision of the results is always low, regardless the method used for their determination. Up to date, the activity concentrations recommended by the Spanish normative for gross alpha and gross beta indexes in drinking water are 0.1 Bq·L-1 and 1 Bq·L-1, respectively [RD. 140/2003]. It is necessary to get fast, reliable methods for their determination in order to comply with the limits established by law. The most widespread screening methods for gross alpha and gross beta index determination employ gas-flow proportional detector counting (GPC), and/or ZnS scintillator counter techniques [ISO9696, 1991 and ISO9697, 1991]. Self-attenuation problems are the main disadvantage of evaluation using ZnS or/and GPC methods and will lead to inaccuracies in the counting efficiency. Those inaccuracies increase in precipitates from high dissolved solid waters, due to inhomogeneous thicknesses 3 and differences of composition of the sample and the standards; affecting mainly to the alpha counting [Dickesen et al. 2008]. Simultaneous measurement of gross alpha and gross beta activities in water samples by liquid scintillation counter (LSC) has proven to be an attractive alternative to GPC or ZnS scintillation. In the last years, LSC has played an important role in this field since it improves some of the performances of GPC and ZnS methods: it provides adequate sensitivity, minimum sample preparation time and operator intervention, lower counting times and lower costs. Furthermore it avoids alpha spectrometry self-attenuation problems. One of the main drawbacks of the LSC measurements is the quenching [Salonen and Hukkanen 1997]. Quenching in LSC is defined as the decrease of photons that reach the photomultiplier in a LS Counter. This decrease is energy dependent and affects both counting efficiency and alpha/beta separation [Pates et al. 2002] and its effect in quantification has to be evaluated. In literature several expressions to calculate gross alpha ad beta indexes are found [see for instance, Sánchez-Cabeza and Pujol 1998, Salonen 2006, Rusconi et al. 2006 and DeVol et al. 2007]. However these formulae are subsequently applied using a variety of approaches that include, or not, efficiency calculations due to quenching effects and/or interference corrections [Forte et al. 2007, Happel et al. 2004, Kleinschmidt 2004, Ruberu et al. 2008, Salonen 2006, Sánchez-Cabeza and Pujol 1995, Wisser et al. 2006]; but, in some cases, it is not clearly specified how these corrections are applied. To help the determination of gross alpha and beta environmental levels through LSC, a preconcentration method is necessary. For example, in Spain, gross alpha index is around 20 mBq·L-1 in tap water and 250 mBq·L-1 for beta index [PVRA 2007]. Different preconcentration methods are described in the bibliography, such as slow evaporation of the sample [Sánchez-Cabeza and Pujol 1995, Martín et al. 1999, Dávila et al. 2002, and Forte et 4 al. 2007], or freeze-drying of the samples [Salonen and Hukkanen 1997 and Wisser et al. 2006]. High conductivity and high pH values in samples before or after pre-treatment might increase quenching degree in the sample depending on the procedure applied. For example, Sánchez-Cabeza and Pujol (1995) proposed direct evaporation 1:10, as pre-concentration method; however, the implications of this evaporation in the quenching degree of the sample were not considered. To avoid this problem Rusconi et al. (2006) suggested an evaporation process under controlled pH conditions, using nitric acid to get a constant quenching degree. In this work, several procedures for evaporation were tested and the performances of each pre-treatment technique were evaluated. In addition, the calculation of alpha and beta indexes was done following three different calculation procedures: either considering or not interferences and/or quenching. The differences between the three calculation methods, in terms of accuracy and precision, for samples with different salinity were studied. The aim of this paper is to present a robust method, which includes minimum sample preparation, for the measurement of gross alpha/beta index in aqueous samples of different salinities. We developed a pre-treatment procedure and also we evaluated which is the most convenient calculation method for gross alpha and beta indexes, to obtain similar information to the one provided by classical methods, such as the based on GPC and ZnS counting. The limitations in the application of the developed method in terms of quenching levels were also evaluated. 2. MATERIALS AND METHODS 2.1. Instrumentation and calibration An ultra-low-background Quantulus 1220TM liquid scintillation spectrometer from Perkin Elmer (Turku, Finland) has been used for the measurements. A Pulse Shape Analyzer was used (PSA) to perform a simultaneous alpha/beta gross counting, which allows the 5 distinction between alpha and beta pulses. Besides, Quantulus 1220 uses the External Standard Quench Parameter (SQP(E)) to quantify quenching; when quenching increases in the sample, the SQP(E) parameter decreases. The calibration of the counter was performed through the evaluation of the optimum PSA threshold level and counting efficiency. Standard solutions of 241Am and 40K as alpha and beta emitters were used in the calibrations. 241Am had a 4.7  0.4 kBq·g-1 certified nominal activity concentration, provided by CIEMAT (Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas, Spain). 40K was prepared from a standard solution of 1 g·L-1 of KCl supplied by Merck (Darmstatd, Germany). The ratio between 40K and stable potassium is fixed using a factor of 27.6 Bq of beta activity per gram of total potassium [WHO 2004]. The liquid scintillation cocktail used was Optiphase Hifase 3, supplied by Perkin Elmer, polyethylene scintillation vials were supplied by Sarstedt (Nümbrecht, Germany). The amount of sample was 8.5 (mL) in all the standards and samples prepared. The vial was filled to the top with scintillation cocktail (16 mL), in order to control 222Rn losses. When possible, the sample was measured immediately after sample preparation, so that the contribution of 222Rn was considered negligible. When this was not possible, the sample was measured after 24 days, when radon has reached secular equilibrium, and using the counts from 214Po peak the contribution of 222Rn was calculated and subtracted from the gross alpha spectrum (Villa et al, 2005). Results of the calibration are displayed in Table 1. The counting windows (set to channels 400-800 and 500-1000 for alpha and beta windows, respectively) were chosen so that all alpha events (in the range 4-8MeV) and beta events (with energy values higher than 200 keV) were included [Rusconi et al. 2006]. Alpha (241Am) and beta (40K) emitters were independently measured, and the interferences of their spectra in beta () and alpha () windows respectively were determined for different PSA values. The chosen optimum PSA 6 (95) is the one that minimizes total interference ( =  + ) [Villa et al. 2003]. Alpha and beta counting efficiencies were calculated for those radionuclides with no PSA discrimination. The mathematical expressions used to calculate     and  values are well known and can be found elsewhere (e.g. Rusconi et al. (2006), Sánchez-Cabeza et al. (1995), and Villa et al. (2003)). 2.2. Quenching evaluation Quenching leads to two direct and independent effects that had to be evaluated: efficiency diminution and changes in the interferences i.e. decrease of beta interference and increase of alpha interference [Villa et al. 2003]. The relationship between quenching degree (SQP (E)) and efficiency decrease was determined by using acidified standards solutions spiked with known amounts of alpha (241Am) and beta (40K) emitters. The five quenching degrees were prepared for each radionuclide studied, using increasing amounts of 65% HNO3 (0, 5, 10, 15 and 30L). Figure 1 (a and b) shows the measured quenching level SQP (E) versus counting efficiency for both 241Am and 40K, obtained directly from the measurement of the quenched standards with no PSA discrimination. No PSA separation is used here because quenching, efficiency and alpha and beta interferences are interrelated parameters (Villa et al. 2003) and to overcome those interrelationships, we have independently established the dependence of efficiency and interference with quenching (i.e. no PSA discrimination is done in efficiency calibration, and  and  are established regardless  and  efficiency) (Villa et al., 2003). Thus, first it was established the counting efficiency with no PSA discrimination (Figure 1) and subsequently alpha and beta interferences were calculated using the PSA system (Figure 2). This way both corrections can be independently included in the activity calculation equations, as will be further presented. 7 The curve efficiency versus quenching is usually fitted to a second order curve [Villa et al. 2003]. However, since quenching levels here were low enough (beta counting efficiencies ranging from 96% to 57%); results can be fitted to a first order curve, which will lead to lower uncertainties, following the equations below: 241Am: () = (- 50 ± 11) + (0.20 ± 0.01) SQP(E); R2 = 0.94 (1) 40K: () = (-162 ± 10) + (0.33 ± 0.01) SQP(E); R2 = 0.90 (2) Results showed that as the chemical quenching increases, counting efficiency for 40K decreases. The same trend is observed for the alpha emitter (241Am), although in a lesser extent than that for 40K. Efficiencies are therefore clearly different. This is due to the fact that alpha is a mono-energetic emission, thus efficiency counting would only diminish with quenching if the measurement is affected by large levels of quenching (i.e. SQP (E) = 700). On the contrary, the beta emission shows a continuous spectrum and thus counting efficiency is more sensitive to the quenching effect, as it is observed in Figure 1. Therefore, beta efficiency decrease from (93 ± 1)% to (57 ± 1)%, for a sample without and with quenching (i.e. SQP(E) value of 651). Finally, the dependence of the total interferences with increasing SQP (E) values is obtained measuring the acidified standards solutions at PSA 95 and calculating  and . When quenching increases there is a steady increasing in  whereas  decreases more slowly, resulting in an increasing of the total interference [Pates et al. 1998, Villa et al. 2003]. This dependence is shown in Figure 2 and results are fitted to a first order curve: (%) = (71 ± 4) + (-0.088 ± 0.006) SQP(E) (3) 8 Total interferences will increase with quenching. Hence the optimum PSA that minimize interference in unquenched samples will change for quenched samples. In the following sections we will study how this affects the accuracy of the activity calculation. 2.3. Pre-concentration procedures The pre-concentration of the sample before its measurement is devoted to lower the MDA and hence allow the measurement of gross alpha and gross beta in environmental samples. The matrix used for this evaluation was Milli-Q water spiked with a known activity of 40K (4.8 Bq). 100 mL of samples were initially pre-concentrated by evaporation. The subsequent sample pre-treatments are described in Table 2. Three replicates of each solution were analysed in order to establish repeatability and this parameter was evaluated in terms of relative standard deviation (RSD). A single parameter that includes the recovery from dissolution of the sample (Rq) and the beta counting efficiency,  (quenching dependent) was used to evaluate the total recovery of the different pre-concentration procedures: the effective recovery (Rqeff), whereby both parameters are evaluated together. It is calculated as follows: 60 )( 40  − == K bkgsample qeff A cpmcpm SQPRRq   (4) Where cpmsample corresponds to the count rates per minute of the sample in the counting window, cpmbkg corresponds to the count rates per minute of the background in the counting window, and K A40 is the spiked 40K activity to the sample. 2.4. Samples description and validation The gross alpha and beta indexes obtained using our LSC procedure were validated with aqueous reference matrices of a proficiency test organized in 2006 by Institut de Radioprotection et de Sûreté Nucléaire (ISRN) from France (code 84SH300). This water 15 drinking water [Zapata-Garcia et al. 2009]. These samples had high alpha and beta activity levels and were measured without a pre-concentration procedure (8.5 mL sample: 16 mL cocktail) because the estuary presented enhanced natural radioactivity levels [Absi et al. 2004; Villa et al. 2009]. Figure 5 (a and b) shows the results for both LSC methods. High quenching degrees were measured (Figure 5 caption). Concerning gross beta activity, the two LSC procedures reported values which coincide well within 2 (k=). However, for gross alpha activity the results were more dispersed. The results obtained applying the LSC 2, reported very different values for samples OT1, compared to the results obtained by our procedure. These divergences in alpha results (higher than 40%) were due to the use of a constant optimum PSA, without taking into consideration the variation of alpha and beta interferences with quenching degree. In Figure 2 the dependence of total interference with increasing SQP(E) values was displayed, and it was shown that the optimum PSA for unquenched conditions is different to the optimum PSA for a high quenched sample. Hence if we do not consider this effect and we use a wrong PSA value, the total interferences will increase up to 50%. For this reason the applicability of a LSC method is always restrained by the quenching degree in the sample and the optimum PSA calculated for an unquenched case is no longer valid for high quenched conditions and its use in that situation will lead to incorrect results. 4. SUMMARY AND CONCLUSIONS Gross alpha and beta indexes obtained by the developed method based on LSC was as accurate and precise as the results obtained from methods based on gas proportional counting and ZnS alpha counting, and performed well for the determination of gross alpha and gross beta activities around the levels given in the current legislation. This LSC method 16 will be useful when applied to screening studies, where the large number of analysed samples requires fast procedures but do not require high precision results. The water pre-treatment method proposed here was the total evaporation of the sample and then the dissolution with HNO3 0.05M, which leads to low quenching degree in the sample and maximum recovery ((82 ± 4) %). When activity indexes, alpha and beta, were calculated neglecting quenching and misclassification effects, the procedure was validated in samples from a proficiency test. When alpha/beta crossover and/or quenching corrections were applied, the results obtained did not improve accuracy within 25% dispersion, a widespread acceptance limit for gross alpha and beta counting. Nevertheless, interferences corrections should not be neglected if alpha and beta count rates differ in one order of magnitude or more. The developed method was also proved to be robust and reliable for water samples with moderate quenching degree (i.e. low and middle conductivity). The limitation of the technique lies on high quenched samples; in that case the procedure was no longer reliable due to the interrelated influence of quenching on the counting efficiency and also on alpha and beta interferences. 5. ACKNOWLEDGMENTS The authors would like to thank Dr. P. Masqué and J.M. Bruach from Universitat Autónoma de Barcelona their support and collaboration. Also to Generalitat de Catalunya, Departament d’Innovació, Universitats i Empresa, project 2009 SGR 223. 6. REFERENCES Absi, A., Villa, M., Manjón, G., Moreno, H.P., Periañez, R. 2004. Self-cleaning in an estuarine area formerly affected by 226Ra anthropogenic enhancements. Science of the Total Environment 329, 183-195. 17 Dávila J.I., López del Río H., Mireles F., Quirino L.L., Villalba M.L., Colmenero L., Montero M.E. 2002. Radioactivity in bottled waters sold in Mexico. Applied Radiation and Isotopes 56, 931–936. DeVol T.A., Theisen C.D., DiPrete D.P. 2007. Effect of quench on Alpha/Beta Pulse Shape Discrimination of Liquid Scintillation Cocktails. The Radiation Safety Journal volume 92, 2, 108-111. Dickstein P., Brashi L., Haquin G. 2008. Empirical compensation for the precipitation effect in the measurement of gross alpha content in drinking water. Health Physics 95, 5, 162-167. Forte M., Rusconi R., Cazzaniga M.T., Sgorbati G. 2007.The measurement of radioactivity in Italian drinking waters. Microchemical Journal 85, 98-102. Happel S., Letessier P., Ensinger W., Eikenberg J. H., Thakkar A. H., Horwitz E. P. 2004. Gross alpha determination in drinking water using a highly specific resin and LSC. Applied Radiation and Isotopes 61, 339-344 ISO 9696, 1991. Water quality—measurement of gross alpha activity in non-saline water— thick source method. International Standard 9696. International Organization for Standardization, Geneva, Switzerland. ISO 9697, 1991. Water quality—measurement of gross beta activity in non-saline water. International Standard 9696. International Organization for Standardization, Geneva, Switzerland. ISO/IEC 17025:2005. General requirements for the competence of testing and calibration laboratories. Kleinschmidt R.I. 2004. Gross alpha and beta activity analysis in water, a routine laboratory method using liquid scintillation analysis. Applied Radiation and Isotopes 61, 333–338. Martín A., Rubio M.P., Gómez V., Jurado M. 1999. Radioactivity in bottled mineral waters. Applied Radiation and Isotopes 50, 1049-1055. 18 Palomo M., Peñalver A., Borrull F., Aguilar C. 2007. Measurement of radioactivity in bottled drinking water in Spain. App. Radiation and Isotopes 65, 1165–1172. Pates, J., Cook, G., MacKenzie A.B., Passo Ch J. 1998. Implications of beta energy and quench level for alpha/beta liquid scintillation spectrometry calibration, The Analyst 123, 2201-2207. Programa de Vigilancia Radiológica Ambiental (PVRA 2007). Resultados 2007. Colección de Informes Técnicos 21. 2008. Madrid, Spain. Real Decreto 140/2003, Boletín Oficial del Estado n.45, 21 de Febrero del 2003, 7228-7245. Ruberu R.S., Yun-Gang L., Kusum Perera S. 2008. An improved liquid scintillation counting method for the determination of gross alpha activity in ground water wells. Health Physics 95, 4, 397-406. 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. Applied Radiation and Isotopes 64, 10-11, 1124-1129. Salonen L., Hukkanen H. 1997. Advantages of low-background liquid scintillation alphaspectrometry and pulse shape analysis in measuring Rn-222, uranium and Ra-226 in groundwater samples. Journal of Radioaanalytical and Nuclear Chemistry 226, 67-74. Salonen L. 2006. Alpha/Beta Liquid Scintillation Spectrometry in Surveying Finnish Groundwater Samples. Radiochemistry, 48, 606-612. Sánchez-Cabeza J.A., Pujol Ll. 1995. A rapid method for the simultaneous deteremination of gross alpha and beta activities in water samples using a low background liquid scintillation counter. Health Physics 68, 5, 674-682. Villa M., Manjón G., García-León M. 2003. Study of colour quenching effects in the calibration of liquid scintillation counters: the case of 210Pb. Nuclear Instruments and Methods in Physics Research A 496, 413-424. 19 Villa, M., Moreno, H.P., Manjón, G. 2005. Determination of 226Ra and 224Ra in sediments samples by liquid scintillation counting. Radiation Measurements 39, 543-550. Villa M., Mosqueda F., Hurtado S., Mantero J., Manjón G., Periañez R., Vaca F., García- Tenorio R. 2009. Contamination and restoration of an estuary affected by phosphogypsum releases, Science of the Total Environment 408, 69-77. Wisser S., Frenzel E., Dittmer M. 2006. Innovative procedure for the determination of gross-alpha/gross-beta activities in drinking water. Applied Radiation and Isotopes 64, 368- 372. World Health Organisation (WHO) 2004, Guidelines for Drinking-water Quality 1, 197- 209. Zapata-García D., Llauradó M., Rauret G. 2009. Establishment of a method for the rapid measurement of gross alpha and gross beta activities in sea water. Applied Radiation and Isotopes 67, 978-981. FIGURES Figure 1:  and  efficiency as a function of SQP(E) value. Measured quenching level SQP (E) vs counting efficiency for both 241Am (Figure 1a) and 40K (Figure 1b). Results are 20 directly obtained from the measurement of the quenched standards with no PSA discrimination. Figure 2: Total interferences, expressed in %, vs SQP (E) values obtained for different quenching degree. PSA threshold level is set to 95. Figure 3: Gross alpha (Figure 3a) and gross beta (Figure 3b) activities obtained through the developed measurement procedure and the classical method. Samples are from Ter River and drinking water from North Catalonia, Spain. Quenching levels are negligible. Three aliquots were prepared for each sample evaluated. The uncertainty is calculated using a confidence level of 95% (K=2). Figure 4: Gross alpha activity (Figure 4a) and gross beta activity (Figure 4b) through our measurement procedure (Option A, B and C) and an alternative procedure (LSC2) and ZnS (alpha) and GPC (beta). Samples are from two Spanish rivers (Ebro and Llobregat). SQP(E) values were 710 ± 2 for Ebro River (EBO) and 678 ± 3 for Llobregat River (LLOB). Being 756 ± 3 the reference value for an unquenched sample. Three aliquots were prepared for each sample evaluated. Uncertainty bars were obtained from the standard deviation of three replicates. Figure 5: Gross alpha activity (Figure 5a) and gross beta activity (Figure 5b) through our measurement procedure (Option A, B and C) and an alternative procedure (LSC2). Water samples are from Odiel-Tinto estuary (south of Spain). SQP(E) = 702, 706, 705, 706 for OT1, T3, T4 and T5 respectively. 21 Figure 1 22 Figure 2 23 Figure 3 24 Figure 4