Contents lists available at ScienceDirect Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso The use of γ-rays analysis by HPGe detector to assess the gross alpha and beta activities in waters M.F.S. Casagrande, D.M. Bonotto⁎ Departamento de Petrologia e Metalogenia, Universidade Estadual Paulista (UNESP), Av. 24-A No. 1515, C.P. 178, CEP 13506-900 Rio Claro, São Paulo, Brazil H I G H L I G H T S • Screening of gross alpha and beta activities in waters. • Gamma rays analysis with HPGe detector. • Non-destructive essay. • Application to different groundwater samples. • Enhanced radioactivity levels in Brazilian aquifers. A R T I C L E I N F O Keywords: Gross alpha and beta γ-rays analysis HPGe detector Groundwater radioactivity Brazilian aquifer system A B S T R A C T This paper describes an alternative method for evaluating gross alpha and beta radioactivity in waters by using γ- rays analysis performed with hyper-pure germanium detector (HPGe). Several gamma emissions related to α and β- decays were used to provide the activity concentration data due to natural radionuclides commonly present in waters like 40K and those belonging to the 238U and 232Th decay series. The most suitable gamma emissions related to β- decays were 214Bi (1120.29 keV, 238U series) and 208Tl (583.19 keV, 232Th series) as the equation in activity concentration yielded values compatible to those generated by the formula taking into account the detection efficiency. The absence of isolated and intense γ-rays peaks associated to α decays limited the choice to 226Ra (186.21 keV, 238U series) and 224Ra (240.99 keV, 232Th series). In these cases, it was adopted appropriate correction factors involving the absolute intensities and specific activities for avoiding the interferences of other γ-rays energies. The critical level of detection across the 186–1461 keV energy region corresponded to 0.010, 0.023, 0.038, 0.086, and 0.042 Bq/L, respectively, for 226Ra, 224Ra, 208Tl, 214Bi and 40K. It is much lower than the WHO guideline reference value for gross alpha (0.5 Bq/L) and beta (1.0 Bq/L) in waters. The method ap- plicability was checked by the analysis of groundwater samples from different aquifer systems occurring in the Brazilian states of São Paulo, Minas Gerais and Mato Grosso do Sul. The waters exhibit very different chemical composition and the samples with the highest radioactivity levels were those associated with lithotypes pos- sessing enhanced uranium and thorium levels. The technique allowed directly discard the 40K contribution to the gross beta activity as potassium is an essential element for humans. 1. Introduction The radiological characterization of drinking waters has been a topic of concern of several organizations worldwide, mainly due to the hazard effects into human health caused by the exposure to natural radionuclides present in the hydrosphere (Cantor, 1997; Hopke et al., 2000). The chemical and radiological groundwater quality is associated to the geochemical context of the related aquifers. Under this perspective, it has been recognized the importance of the disintegration of 40K and a large number of natural radionuclides belonging to the 238U, 235U and 232Th decay series. Some rock-types contain relatively high concentra- tions of uranium and thorium, such as granites and alkaline rocks. Potassium is a major element in many common minerals like feldspars, despite the abundance of the radioactive 40K is only 0.012% (Heier and Billings, 1969; Rogers and Adams, 1969a, 1969b). Therefore, many countries have established guidance levels for in- gestion of radionuclides in drinking water. WHO (2011) proposed an effective dose of 0.1 mSv as an annual limiting value based on the in- gestion of 2 liters of water per day. WHO (2011) has also recommended https://doi.org/10.1016/j.apradiso.2018.02.027 Received 6 March 2017; Received in revised form 8 January 2018; Accepted 22 February 2018 ⁎ Corresponding author. E-mail address: dbonotto@rc.unesp.br (D.M. Bonotto). Applied Radiation and Isotopes 137 (2018) 1–11 Available online 23 February 2018 0969-8043/ © 2018 Elsevier Ltd. All rights reserved. T http://www.sciencedirect.com/science/journal/09698043 https://www.elsevier.com/locate/apradiso https://doi.org/10.1016/j.apradiso.2018.02.027 https://doi.org/10.1016/j.apradiso.2018.02.027 mailto:dbonotto@rc.unesp.br https://doi.org/10.1016/j.apradiso.2018.02.027 http://crossmark.crossref.org/dialog/?doi=10.1016/j.apradiso.2018.02.027&domain=pdf maximum activity concentration values for gross alpha (0.5 Bq/L) and beta (1.0 Bq/L) in waters. More sophisticated and time-consuming procedures for determining the radionuclides content should be adopted when the results of a screening are positive (WHO, 2011). Several techniques and detectors are available for determining the radioactivity in waters, including the gross alpha and beta activity measurements. The use of different approaches sometimes imply on debates and discussions due to uncertainties associated to such mea- surements. Precipitation and co-precipitation procedures relying on ISO (9696/9697) and EPA methods are examples of sample preparation processes that must be combined to detection systems, such as scintil- lation, flow proportional, and liquid scintillation counting. Each tech- nique has advantages and disadvantages in terms of efficiency, α/β spillover, self absorption, cost and time demand (Jobbágy et al., 2010). Bonotto et al. (2009) suggested a combined gamma and alpha spec- trometric technique for gross beta and alpha readings, respectively. A γ- rays NaI(Tl) well-type scintillaton detector was used to the indirect determination of the beta emissions as occur reduction of the absorp- tion effects caused by the natural salinity present in the water samples. Some disadvantages of the method are the non-simultaneous readings and the sample destruction due to evaporation until dryness and counting of the residue deposited on a planchet. Jobbágy et al. (2010) inserted this combined technique in their Table 2 together other methods available for determining gross alpha and beta activities. Jobbágy et al., (2014, 2015) also described the results of a European interlaboratory comparison (ILC) on gross alpha/beta activity de- termination that only included the most common standard methods, based on direct evaporation, co-precipitation and liquid scintillation counting. Therefore, the combined gamma/alpha spectrometric tech- nique described by Bonotto et al. (2009) and other methods were not included in the ILC held by Jobbágy et al., (2014, 2015), despite their usefulness in radiological studies focusing different aquifer systems, for instance, those performed by Bonotto and Bueno (2008) and Bonotto (2011). Semiconductor materials have been developed during the 60′s and used for gamma radiation and charged particles detection, such as high energy electrons, protons, alpha and beta particles, etc. Hyper-pure germanium gamma rays detector (HPGe) is an example of this tech- nology and its high spectral resolution is a very important aspect when compared with that of the NaI(Tl) scintillator (Hossain et al., 2012). This characteristic is directly associated with the capability of separa- tion of two adjacent energy peaks and unambiguous nuclide identifi- cation. This paper describes a method using a gamma spectrometric system associated with a coaxial HPGe detector for the characterization of gross alpha and beta activities in drinking waters. It is as a step forward of the analytical approach reported by Bonotto et al. (2009) as describes the usefulness of the high resolution γ-rays spectrometry with HPGe on providing gross alpha/beta activities in waters. The analysis of groundwater samples from different aquifer systems in the Brazilian states of São Paulo, Minas Gerais and Mato Grosso do Sul is also re- ported here. The experiments outline in this paper is not the same of that re- ported by Jobbágy et al., (2014, 2015), which included 71 registered European participant laboratories in the ILC. Despite its importance, such type of initiative involves costs, demanding time and several ar- rangements to its implementation, but the present conditions existing in South America don´t allow similar evaluation as only a few laboratories are able to perform these analyses in waters. Additionally, there is a large data scarcity for these parameters in Brazilian waters, justifying the use of faster and cheaper analytical methods like that suggested in this paper. 2. Theoretical approach Natural radioactivity is basically generated by the decay of cosmogenic radionuclides (14C, 26Al, 10Be, 3H, and 36Cl) and radio- nuclides associated directly with the geological environment (e.g. 238U, 235U, 232Th and 40K) (Bentley et al., 1986; Faure and Mensing, 2005). 238U, 235U, and 232Th originate decay series with a significant number of daughters and having Pb isotopes as stable end-members. U, Th and K are present in the lithosphere since the Earth´s formation and are still one of the major contributors to the environmental radioactivity and radiological characterization of the waters. 238U is the most abundant uranium isotope (99.3%). Its decay chain has 15 daughters, finishing at stable 206Pb (Chu et al., 1999). 235U has a lower (0.7%) abundance and a simpler decay chain finishing at stable 207Pb (Chu et al., 1999). 232Th practically represents the totality (99.98%) of thorium in nature and its series finishes at stable 208Pb (Chu et al., 1999). The radioactivity related to potassium is entirely based on the 40K (abundance =0.012%) disintegration. It decays to 40Ca through a β- emission (89.3%) and to 40Ar by electron capture (10.7%), in this case also occurring a 1460.83 keV energy gamma emission (Chu et al., 1999). The 235U and associated daughters contribution to the radiological quality of most waters is often neglected due to its low isotopic abun- dance relative to the one of 238U. But its shorter half-life implies on a higher specific activity value. In terms of solubility, uranium is of easy incorporation into waters when compared to thorium, even though solubility might be a very complex issue if pH, Eh, colloidal particles and concentration of other ions in water are considered. Despite the fact that uranium is more soluble, the concentration of thorium in crustal rocks, however, is approximately three to four times higher (Rogers and Adams, 1969a, 1969b; Thorton, 1983; Clark, 1989). The technique adopted here uses a gamma spectrometric system for the characterization of gross alpha and beta activities in an indirect way. Gamma rays associated with these particles emissions in the 238U, 232Th, and 235U decay series are shown in Tables 1, 2. Regarding to α decays, the most notable characteristic is the absence of peaks asso- ciated with intense and isolated gamma emissions as there are over- lapped signals that make the calibration more difficult. In this case, the most viable representatives of 238U and 232Th series are 226Ra (186.21 keV) and 224Ra (240.99 keV), respectively. On the other hand, there are several options for choosing gamma emissions related to β- decays, for instance: 238U series - 214Bi (609.31, 1120.29, and 1764.49 keV); 232Th series - 208Tl (583.19 and 2614.53 keV), 228Ac (911.20 and 968.97 keV), 212Pb (238.63 keV). 2.1. Calibration steps The γ-rays detector utilized in association with the spectrometric system was a coaxial EG&G ORTEC HPGe type whose resolution is 2.1 keV and efficiency 63% at the 1332.50 keV 60Co peak. The X-Cooler III unit provided a 77K temperature for the system, which also utilizes an EG&G ORTEC ASPEC-927 multichannel buffer and Gamma Vision software for the data acquisition. The calibration in energy was performed using the artificial radio- active sources 133Ba (356.02 keV), 137Cs (661.66 keV) and 60Co (1173.24, 1332.50 and 2505.74 keV – “sum peak”), as well the fol- lowing γ-rays emissions from a 1% uranium standard: 210Pb – 46.54 keV (4.2%), 226Ra – 186.21 keV (3.6%), 214Pb - 295.22 (19.3%) and 351.93 keV (37.6%), and 214Bi - 609.31 keV (46.1%). Two graphs of energy (keV) vs. channel were generated (Fig. 1): E =0.2009. Ch +4.4577 (lower energy interval of 40–600 keV) and E =0.2003. Ch +4.8454 (higher energy interval of 600–2600 keV), where E is the energy (keV) and Ch is the channel number in the multichannel ana- lyzer. The system calibration in activity concentration used radioactive sources containing uranium, thorium and potassium. New Brunswick Laboratory of U.S. Department of Energy, Argonne, Illinois, provided the U and Th sources, whereas mixture of KCl and silica supplied the 40K sources. Table 3 shows the main characteristics of the standards M.F.S. Casagrande, D.M. Bonotto Applied Radiation and Isotopes 137 (2018) 1–11 2 utilized, whilst Fig. 2 illustrates some of the acquired gamma spectra. Total alpha/beta measurements for these reference materials are not available. However, a scintillator crystal of Bi4Ge3O12 (BGO-bismuth germanate; 7.13 g/cm3 density) provided dose rate readings as shown in Table 3. The dose rate values increase according to the concentration raising of the certified values, adjusting to a straight line that exhibits a Table 1 Gamma emissions related to β--emitting radionuclides belonging to 238U, 235U and 232Th decay series with their respective absolute intensity (only values higher than 1% are listed). According to Chu et al. (1999). Decay series Radionuclide Energy (keV) Absolute intensity (%) 238U 234Th 63.29 4.8 92.38 2.8 92.80 2.8 214Pb 53.23 1.2 242.00 7.4 295.22 19.3 351.93 37.6 785.96 1.1 214Bi 609.31 46.1 665.45 1.5 768.35 4.9 806.17 1.2 934.06 3.0 1120.29 15.1 1155.19 1.6 1238.11 3.8 1280.96 1.4 1377.67 4.0 1401.50 1.3 1407.98 2.2 1509.23 2.1 1661.28 1.2 1729.60 2.9 1764.49 15.4 1847.42 2.1 2118.55 1.1 2204.21 5.1 2447.86 1.6 210Pb 46.54 4.2 232Th 228Ra 13.52 1.6 228Ac 99.51 1.3 129.07 2.4 209.25 3.9 270.25 3.5 328.00 3.0 338.32 1,3 409.46 1.9 463.00 4.4 772.29 1.5 794.95 4.2 835.71 1.6 911.20 25.8 964.77 5.0 968.97 15.8 1588.19 3.2 1630.63 1.5 212Pb 238.63 43.3 300.09 3.3 212Bi 39.86 1.1 727.33 6.6 785.37 1.1 1620.50 1,5 208Tl 277.35 6.3 510.77 22.6 583.19 84.5 763.13 1.8 860.56 12.4 2614.53 99.0 235U 231Th 25.64 14.5 84.22 6.6 211Pb 404.85 3.8 427.09 1.8 832.01 3.5 Table 2 Gamma emissions related to α-emitting radionuclides belonging to 238U, 235U and 232Th decay series with their respective absolute intensity (only values higher than 1% are listed). According to Chu et al. (1999). Decay series Radionuclide Energy (keV) Absolute intensity (%) 238U 226 Ra 186.21 3.6 232Th 228Th 84.37 1.2 224Ra 240.99 4.1 212Bi 39.86 1.1 235U 235U 109.16 1.5 163.36 5.1 185.71 57.2 202.11 1.1 205.30 5.0 231Pa 27.36 10.3 283.69 1.7 300.07 2.5 302.65 2.2 330.06 1.4 227Th 50.13 8.0 79.72 1.9 93.93 1.4 210.65 1.1 235.97 12.3 256.25 7.0 286.12 1.5 300.00 2.3 304.52 1.2 329.85 2.7 223Ra 122.32 1.2 144.23 3.2 154.21 5.6 269.46 13.7 323.87 3.9 338.28 2.8 445.03 1.3 219Rn 271.23 10.8 401.81 6.4 211Bi 351.06 12.9 Fig. 1. Calibration curves in energy of the γ-rays spectrometric system for the (top) lower energy interval of 40–600 keV and (bottom) higher energy interval of 600–2600 keV. M.F.S. Casagrande, D.M. Bonotto Applied Radiation and Isotopes 137 (2018) 1–11 3 strong significant Pearson correlation coefficient (r= 0.98). Gamma rays peaks representing β- and α disintegrations from 238U and 232Th decay series were selected for calibration purposes. The criteria adopted were the intensities of the gamma emissions, absence of interference from other γ-rays energies in the peak (overlap) and enough energy difference between two close peaks. The selected γ-rays energies were: 238U series - 226Ra (α-decay, 186.21 keV) and 214Bi (β-- decay, 1120.29 and 1764.49 keV); 232Th series - 224Ra (α-decay, 240.99 keV), 208Tl (β--decay, 583.19 and 2614.53 keV) and 228Ac (β-- decay, 911.20 and 968.97 keV). 40K data were acquired from γ-rays of 1460.83 keV energy. However, 226Ra peak overlaps 235U (185.71 keV) in the gamma spectrum. Its contribution in each peak corresponded to 5.9% of the total counting rate as estimated from both absolute intensities and specific activities. The 235U contribution was not considered in the gross alpha evaluation in this paper, but, if necessary, it could be easily in- cluded. 224Ra (240.99 keV) suffers a more complex interference that involves three disintegrations: 212Pb - 238.63 keV (232Th series); 227Th - 235.97 keV (235U series); 214Pb - 242 keV (238U series). In this case, the 224Ra contribution in the peak corresponds to 0.03% of the total counting rate. The activity concentration of each radioactive source was calculated from the concentration values (in ppm for U and Th and in % for K), adopting the following specific activity values: 238U - 12,437 Bq g−1; 232Th - 4057 Bq g−1; K - 31.3 Bq g−1 (Pearce, 2008). The effective in- tensity (counting rate by mass, cps.g−1) was determined in each stan- dard for all radionuclides of interest. Table 4 shows logarithmic equa- tions for the activity concentration of U, Th and K in function of the effective intensity. Two reference materials from New Brunswick La- boratory possessing U concentrations of 40 µg g−1 (CRM 107-A) and 400 µg g−1 (CRM 106-A) were subjected to the γ-rays analysis, allowing verify that the use of eqs. in Table 3 yielded results deviating up to 5% of the certified values, thus, evidencing absence of systematic errors (measurement bias). The detection efficiency (Ɛf) is the ratio of the number of detected events (d1) to the real number of decays (d0) from the radioactive source, i.e. Ɛf (%) = d1/d0 (Rodrigues et al., 2011). It may be also determined by the equation Ɛf (%) = (TS – TB) / (f. AS) where: TS is the counting rate (in cps) in the selected peak of the standard, TB is the background counting rate (in cps), f (in %) is the absolute intensity of the γ-rays peak and AS (in Bq) is the standard activity. The detection efficiency according to the γ-rays energy has been determined for the following radionuclides: 210Pb (46.54 keV), 214Pb (242.00, 295.22 and 351.93 keV), 214Bi (609.31, 1120.29, 1764.49 and 2204.21 keV), 208Tl (583.19 and 2614.53 keV), 228Ac (911.20 and 968.97 keV) and 40K (1460.83 keV). A detection efficiency curve was obtained with a fourth degree polynomial comprising lower gamma energies in the range of 46–968 keV and a quadratic function for higher values (up to 2614 keV) (Fig. 3). The detection efficiency tends to decrease according to the increase of the γ-rays energy. 3. Experimental Groundwater samples were analyzed for gross alpha and beta ac- tivities using the two described approaches. The first was based on the equations given in Table 4, whereas the second from the equation Asa = (Tsa – TB) / (f.Ɛf) where: Asa is the sample activity (in Bq), Tsa is the counting rate (in cps) in the selected peak of the sample and TB is the background counting rate there (in cps). In total, 108 groundwater samples from 33 municipalities comprising different geologic contexts and aquifer systems in the Brazilian states of São Paulo, Minas Gerais and Mato Grosso do Sul were analyzed (Fig. 4 and Table 5). They were collected from wells and public pipes and springs of easy access and often used by the local population and tourists, mainly in the localities classified as spa towns where people seek health treatment based on waters with specific chemical composition. After sampling, the waters were stored in properly sealed and identified polyethylene containers. The dominant geologic contexts are the Phanerozoic sedimentary rocks from Paraná Basin (mainly related to porous aquifer systems) and the crystalline terrains (associated with fractured aquifers). The first group of samples is chiefly located in the states of São Paulo and Mato Grosso do Sul (Fig. 4). Some tube wells are ~3000m deep, comprising almost the entire stratigraphic sequence in São Paulo State, except for the basal sandstones of Furnas Formation (Early Devonian). The other superimposed units are: Itararé Subgroup (Late Carboniferous to Early Permian) - diamictites and sandstones (Milani et al., 2007); Tatuí For- mation (Early Permian) - siltstones and sandstones (Soares, 1972), which, along with Itararé Subgroup, comprises the Tubarão Aquifer System; Passa Dois Group (Irati and Corumbataí formations, Late Per- mian in age) - fine grained sedimentary rocks, such as shales, mud- stones, siltstones, and layers of dolomites, often considered an aqui- clude due to the low permeability lithotypes (Iritani and Ezaki, 2012); Pirambóia and Botucatu formations - Mesozoic sandstones (Schneider et al., 1974), forming the Guarani Aquifer System (GAS), one of the largest aquifer units of the world; Serra Geral Formation (Late Jurassic to Early Cretaceous) - basalt flows and dikes generated by the breakup of Gondwana Supercontinent during Mesozoic Era (Milani et al., 2007), representing the unique fractured aquifer unit in Paraná Basin; Bauru Group (Late Cretaceous) - lithologically heterogeneous sandstones, siltstones, mudstones and locally conglomerates and limestones Table 3 Standards used for calibrating the HPGe γ-rays spectrometer used in this study. Standard Code Composition Weight (g) Concentrationa Dose rateb (nSv/h) K1 LII-KCL-1 61.5 g KCl 61.50 52±0.52wt% K 102.7 K2 LII-KCL-2 54.16 g SiO2 + 28.9 g KCl 83.06 25±0.25wt% K 76.1 K3 LII-KCL-3 80.12 g SiO2 + 5.80 g KCl 85.92 5±0.05 wt% K 78.9 K4 LII-KCL-4 82.11 g SiO2 + 2.90 g KCl 85.01 2.5± 0.025 wt% K 78.6 K5 LII-KCL-5 84.72 g SiO2 + 0.5 g KCl 85.22 0.5± 0.005 wt% K 85.4 U1 CRM 101-A Pitchblende ore – sílica mixture 50 1.007± 0.013 wt% U 879.2 U2 CRM 102-A Pitchblende ore – sílica mixture 50 0.1025±0.0019 wt% U 150.9 U3 CRM 103-A Pitchblende ore – sílica mixture 50 0.0499±0.0007 wt% U 166.4 U4 CRM 104-A Pitchblende ore – sílica mixture 50 0.00988±0.0002 wt% U 127.3 U5 CRM 105-A Pitchblende ore – sílica mixture 50 0.00102±0.00002 wt% U 128.4 Th1 CRM 106-A Monazite sand – sílica mixture 50 1.029± 0.003 wt% Th 423.0 Th2 CRM 107-A Monazite sand – sílica mixture 50 0.1028±0.0002 wt% Th 107.9 Th3 CRM 108-A Monazite sand – sílica mixture 50 0.0515±0.0002 wt% Th 87.5 Th4 CRM 109-A Monazite sand – sílica mixture 50 0.01052±0.00009 wt% Th 76.8 Th5 CRM 110-A Monazite sand – sílica mixture 50 0.00104±0.00001 wt% Th 75.3 a Reported by NBL (1999) for standards U1-U5 and Th1-Th5. b Reported by Barbosa (2016) for scintillations counting during 1800 s as performed with a BGO-bismuth germanate portable gamma rays spectrometer. M.F.S. Casagrande, D.M. Bonotto Applied Radiation and Isotopes 137 (2018) 1–11 4 (Fernandes, 2004), comprising the top unit of the sedimentary se- quence. In terms of fractured aquifer units, samples from Águas da Prata, Poços de Caldas and Pocinhos do Rio Verde municipalities can be grouped in the context of the Poços de Caldas Alkaline Massif (PCAM). The local geology can be defined as an alkaline intrusion into gneissic- granitic terrains during the Cretaceous Period. Thedeschi et al. (2015) describe it as a predominance of tinguaites and phonolites, with sub- ordinated nepheline syenites, lujaurites and piroclastic rocks. On the edge of the massif, near to Águas da Prata city, some of the springs are associated with diabase dikes (Serra Geral Formation) and silicified sandstones with similar structures observed in Botucatu Formation (Szikszay and Teissedre, 1977). Additionally, groundwater samples were taken from Alto do Rio Grande Strip, a Meso-proterozoic tectonic unit marginal to São Francisco Craton with Anchean to Paleo-proterozoic fragments affected by Brasiliano Orogeny (Hasui and Oliveira, 1984; Campos Neto, 1991). The lithotypes can be resumed by Archean gneissic-migmatite pre- dominant basement (Amparo Complex and equivalent units) covered by allochthonous Mesoproterozoic metavolcanic and metasedimentary sequences defined by Andrelândia/São João del Rey/Itapira Group, which consists of micaceous and pure quartzites, garnet schists, meta- graywackes, anfibolites, etc. (Campos Neto et al., 1990). Beato et al. (1999) describes the presence of alkaline veins with hydrothermal al- teration in the area of the water springs of Caxambu, which might be related to the syenite intrusive body identified in the region (Trouw et al., 2003). Marinelli-type beaker with 1 L capacity was used for the γ-rays analysis of the groundwater samples, which lasted about 8.5 h each. The gross beta activity was calculated by the sum of the activity con- centrations regarding to 40K and selected radionuclides from 238U (214Bi, 1120.29 keV) and 232Th (208Tl, 583.19 keV) decay series. The gross alpha activity was calculated by the sum of the activity con- centrations related to 226Ra (238U series) and 224Ra (232Th series). The statistical uncertainty in the gamma readings was ±5–10% in each Fig. 2. (top) γ-rays spectrum for uranium standard (1% U): peak 1 (226Ra – 186.21 keV), peaks 2 and 3 (214Bi – 1120.29 and 1764.49 keV, respectively); (middle) γ-rays spectrum for thorium standard (1% Th): peak 1 (224Ra – 240.99 keV), peaks 2 and 3 (208Tl - 583.19 and 2614.53 keV, respectively), peaks 4 and 5 (228Ac - 911.20 and 968.97 keV, respec- tively); (bottom) γ-rays spectrum for potassium standard (52% K): peak 1 (40K - 1460.83 keV). Table 4 Equations relating the activity concentration of each radionuclide of interest (Ai, in Bq.g−1) with their respective effective intensity (Ie, in cps.g−1). Radionuclide (energy) Decay mode Decay series Equation r2 214Bi (1120.29 keV) β- 238U log(AU) = 1.0409×log(Ie)+2.7229 0.9994 214Bi (1764.49 keV) β- 238U log(AU) = 1.0284×log(Ie)+2.8427 0.9989 226Ra (186.21 keV) α 238U log(AU) = 1.0117×log(Ie)+3.4256 0.9975 228Ac (911.20 keV) β- 232Th log(ATh) = 0.9998×log(Ie)+2.2109 0.9999 228Ac (968.97 keV) β- 232Th log(ATh) = 0.9977×log(Ie)+2.3261 0.9999 208Tl (583.19 keV) β- 232Th log(ATh) = 0.9966×log(Ie)+1.9734 0.9999 208Tl (2614.53 keV) β- 232Th log(ATh) = 0.9989×log(Ie)+2.5471 0.9998 224Ra (240.99 keV) α 232Th log(ATh) = 1.0569×log(Ie)+3.1067 0.9968 40K (1460.83 keV) e.c. – log(AK) = 1.2133×log(Ie)+3.1941 0.9895 Fig. 3. Efficiency detection curves generated for the spectrometric system and their re- spective correlation coefficients. Ɛf corresponds to the efficiency (in %) and e is the γ-rays energy (in keV) divided by 1000. M.F.S. Casagrande, D.M. Bonotto Applied Radiation and Isotopes 137 (2018) 1–11 5 peak, within 1σ standard deviation. The count rate in some samples was low during the gamma spectrometric readings. To determine whether the observed signal was ″true″ or ″false″, it was adopted the critical level of detection (Lc), expressed in number of counts as Lc = 2.33 (B)½ (Currie, 1968), where B is the number of background counts. The data on the background gamma spectra across the 186–1461 keV energy region allowed the estimate of Lc in activity concentration of 0.010, 0.023, 0.038, 0.086, and 0.042 Bq/L, respectively, for 226Ra (186.21 keV), 224Ra (240.99 keV), 208Tl (583.19 keV), 214Bi (1120.29 keV) and 40K (1460.83 keV). Therefore, such Lc values are much lower than the WHO (2011) guideline reference values for gross alpha and beta in waters, respectively, 0.5 and 1.0 Bq/L. 4. Results and discussion Table 6 reports the results of the activity concentration for 226Ra (186.21 keV), 224Ra (240.99 keV), 208Tl (583.19 keV), 214Bi (1120.29 keV) and 40K (1460.83 keV) as determined from the applica- tion of the equations in Table 4 to the acquired database. For 214Bi (1120.29 keV), the equation in Table 4 yielded activity concentration values compatible to those generated by the formula taking into ac- count the detection efficiency (Fig. 5). For 208Tl (583.19 keV), the equation in Table 3 yielded activity concentration values on average ~8% lower than those generated by the expression considering the detection efficiency (Fig. 5). However, the same did not happen for other selected γ-rays emissions associated to beta decay in the 238U series (214Bi - 1764.49 keV) and 232Th series (208Tl - 2614.53 keV; 228Ac - 911.20 and 968.97 keV) (Fig. 5). This is possibly related to the de- tection efficiency decrease when higher γ-rays energies are involved. Therefore, the most suitable γ-rays emissions for providing gross beta information in the 238U and 232Th decay series are 214Bi (1120.29 keV) and 208Tl (583.19 keV), respectively. Szikszay and Teissedre (1981), among others, reported high 222Rn levels for Villela spring (Águas da Prata city). However, such radio- nuclide was not monitored in this study. Comparing the data generated by Beato et al. (1999) and Bonotto (2014, 2015) with those reported in Table 6 it is possible verify that they differ significantly as a con- sequence of the use of different analytical techniques, different sam- pling periods and seasonal variations, among other possible factors. On the other hand, the acquired dataset in this paper for 10 water sources providing from Caxambu and Poços de Caldas spas was compared with that generated for gross beta readings from the gamma spectrometry through a NaI(Tl) scintillation detector as reported by Bonotto et al. (2009). Both databases are significantly related (Pearson correlation coefficient = 0.60). This confirms the usefulness of the new screening method described here as it helps on assessing radionuclides health risks from drinking water. Obviously, it would be advisable to have a comparison of results with other accepted methods as reported by Jobbágy et al. (2010), however, this is a task for further development due to the costs and time demand involved, among other factors. De- spite this, Oliveira et al. (2001), Godoy et al. (2001) and Bonotto (2015) used other conventional methods for measuring 228Ra in several of the waters analyzed in this paper and found values exceeding the WHO (2011) guidance level of 0.1 Bq/L. The effect of the storage time of the groundwaters in the Marinelli- type beaker was evaluated for 20 samples. Another γ-rays reading was done at least 30 days after the first measurement. The use of equations in Table 4 allowed generate a new dataset for emissions related to beta Fig. 4. Spatial distribution of municipalities where groundwater samples were collected and its relation with outcropping aquifer systems. Base map: ANA (2013). Municipalities: 1- Sidrolândia; 2-Amambaí; 3-Três Lagoas; 4- Presidente Epitácio; 5- Presidente Prudente; 6- Araçatuba; 7- Jales; 8- Fernandópolis; 9-Votuporanga; 10-Mirassol; 11-Monte Alto; 12- Taquaritinga; 13-Santa Ernestina; 14-Jaboticabal; 15-Sertãozinho; 16-Paraguaçu Paulista; 17-Sta. Cruz do Rio Pardo; 18-Bernardino de Campos; 19-Águas de Santa Bárbara; 20-Avaré; 21- Sarutaiá; 22-São Pedro; 23-São Carlos; 24-Águas da Prata; 25-Poços de Caldas; 26-Serra Negra; 27-Lindóia; 28-Águas de Lindóia; 29-São Lourenço; 30-Caxambu; 31- Cambuquira; 32- Lambari; 33- Pocinhos do Rio Verde. M.F.S. Casagrande, D.M. Bonotto Applied Radiation and Isotopes 137 (2018) 1–11 6 Table 5 Groundwater samples description and lithological context of the sampling points. Sample no. Spring/Well City/Statea Lithologic contextb 1 7 Bis São Lourenço/MG gbg 2 Sulfurosa São Lourenço/MG gbg 3 Poço 22 Três Lagoas/MS SGBP 4 Alcalina São Lourenço/MG gbg 5 Primavera São Lourenço/MG gbg 6 Ferrruginosa São Lourenço/MG gbg 7 Vichy São Lourenço/MG gbg 8 Roxo Rodrigues Cambuquira/MG gsg 9 Com. Augusto Ferreira Cambuquira/MG gsg 10 Fernandes Pinheiro Cambuquira/MG gsg 11 Souza Lima Cambuquira/MG gsg 12 Regina Werneck Cambuquira/MG gsg 13 Da. Leopoldina Caxambu/MG gqa 14 Número 6 São Lourenço/MG gbg 15 Da. Isabel/Conde Deu Caxambu/MG gqa 16 Ernestina Guedes Caxambu/MG gqa 17 D Pedro II Caxambu/MG gqa 18 Viotti Caxambu/MG gqa 19 Poço 21 Três Lagoas/MS BA 20 Mayrink Caxambu/MG gqa 21 Duque de Saxe Caxambu/MG gqa 22 Poço 4 (LA4) Lambari/MG bgm 23 Poço 2 (LA2) Lambari/MG bgm 24 Poço 1 (LA1) Lambari/MG bgm 25 Poço 3 (LA3) Lambari/MG bgm 26 Ademar de Barros Termas de Ibirá/SP BGSG 27 Carlos Gomes Termas de Ibirá/SP BGSG 28 Saracura Termas de Ibirá/SP BGSG 29 Seixas Termas de Ibirá/SP BGSG 30 São Jorge Serra Negra/SP ggm 31 Samaritana Pocinhos do Rio Verde/ MG nep 32 Amorosa Pocinhos do Rio Verde/ MG nep 33 Italianos Serra Negra/SP ggm 34 São Carlos Serra Negra/SP ggm 35 Santa Luzia Serra Negra/SP ggm 36 Santo Agostinho Serra Negra/SP ggm 37 Brunhara Serra Negra/SP ggm 38 Laudo Natel Serra Negra/SP ggm 39 Curie Águas de Lindóia/SP ggm 40 Sant'ana Serra Negra/SP ggm 41 Bioleve Lindóia/SP ggm 42 São Benedito Lindóia/SP ggm 43 Poço 15 Araçatuba/SP BA 44 Poço 40 Votuporanga/SP BGSG 45 Poço 74 Amambaí/SP SGBP 46 Poço60 Águas de Santa Bárbara/ SP SG 47 Poço 62 Águas de Santa Bárbara/ SP SGBP 48 Poço 61 Águas de Santa Bárbara/ SP SGBP 49 Poço 18 Araçatuba/SP BP 50 Poço 57 Avaré/SP BSBP 51 Poço 55 Avaré/SP BGSG 52 Poço 54 Avaré/SP BGSG 53 Poço 58 Avaré/SP SBPP 54 Poço 56 Avaré/SP BP 55 Poço 64 Bernardino de Campos/ SP SG 56 Poço 65 Bernardino de Campos/ SP SGBP 57 Beleza Águas de Lindoia/SP ggm 58 Poço 67 Santa Cruz do Rio Pardo/SP SG 59 Poço 28 Taquaritinga/SP BA 60 Poço 41 Votuporanga/SP BGSG 61 Poço 3 Jaboticabal/SP BGSG 62 Poço 4 Jaboticabal/SP BGSG 63 Poço 79 Presidente Epitácio/SP BA Table 5 (continued) Sample no. Spring/Well City/Statea Lithologic contextb 64 Frayha Poços de Caldas/MG nep 65 Poço 29 Taquaritinga/SP BSBO 66 Macacos Poços de Caldas/MG nep 67 Poço 11 Monte Alto/SP BGSG 68 Poço 13 Monte Alto/SP BSBO 69 Poço 12 Monte Alto/SP BGSG 70 Poço 38 Votuporanga/SP BA 71 Poço 23 Santa Ernestina/SP BGSG 72 Poço 39 Votuporanga/SP BA 73 Poço 24 Santa Ernestina/SP BGSG 74 Poço 25 Santa Ernestina/SP BSBO 75 Poço 26 Santa Ernestina/SP BSBO 76 Poço 74 Presidente Prudente/SP BA 77 Poço 75 Presidente Prudente/SP BGSG 78 Poço 76 Presidente Prudente/SP BGSG 79 Poço 80 Presidente Epitácio/SP BSBO 80 Poço 53 Jales/SP BSBP 81 Poço 73 Paraguaçu Paulista/SP ALL 82 Poço 77 Presidente Prudente/SP SBBPP 83 Poço 44 Fernandópolis/SP BGSG 84 Poço 45 Fernandópolis/SP BGSG 85 Poço 47 Fernandópolis/SP BGSG 86 Poço 46 Fernandópolis/SP BGSG 87 Poço 71 Paraguaçu Paulista/SP BA 88 Poço 50 Jales/SP BGSG 89 Poço 51 Jales/SP BGSG 90 XV de Novembro Poços de Caldas/MG nep 91 Quisisana Poços de Caldas/MG nep 92 Poço 78 Presidente Epitácio/SP BA 93 São Roque Águas de Lindóia/SP ggm 94 Poço 72 Paraguaçu Paulista/SP BA 95 Poço 27 Taquaritinga/SP BA 96 Poço 69 Santa Cruz do Rio Pardo/SP SGBO 97 Poço 66 Santa Cruz do Rio Pardo/SP SGBO 98 Poço 77 Sidrolândia/SP SGBO 99 Poço 16 São Carlos/SP SGBO 100 Poço 68 Santa Cruz do Rio Pardo/SP SG 101 Poço 33 São Pedro/SP BP 102 Poço 70 Sarutaiá/SP SGBO 103 Poço 2 Sertãozinho/SP SGBP 104 Villela Águas da Prata/SP alk 105 Vitória Águas da Prata/SP alk 106 Poço 36 Mirassol/SP BA 107 Poço 43 Votuporanga/SP BSPI 108 Poço 42 Votuporanga/SP SBBPP gbg = garnet-biotite-gneiss, migmatite, metabasite, pegmatitic veins, quaternary de- posits; gsg = garnet schist, gneiss, muscovite quartzite, amphibolite, pegmatitic veins, quaternary deposits; gqa = gneiss, quartzite, alkaline intrusions, quaternary deposits; bgm =biotite gneiss, metabasite, pegmatitic veins, quartzite, muscovite schist, qua- ternary sediments; ggm =granite, gneiss, migmatite, schist, quartzite, limestone, dolo- mite; nep = nepheline syenite, phonolites, pyroclastic rocks and volcanic tuffs; alk = alkaline rocks, silicified sandstones, phonolite, diabase; BA = rocks related to Bauru Group; SG = rocks related to Serra Geral Formation; BP = rocks related to Botucatu- Piramboia Formation; BGSG = rocks related to Bauru Group and Serra Geral Formation; BP = rocks related to Botucatu-Piramboia Formations; SGBP = rocks related to Serra Geral and Botucatu-Pirambóia Formations; BSBP = rocks related to Bauru Group, Serra Geral and Botucatu-Piramboia Formations; SBPP = rocks related to Serra Geral and Botucatu-Piramboia Formations and Passa Dois Group; SBBPP = rocks related to Serra Geral, Botucatu and Piramboia Formations and Bauru and Passa Dois Groups; SGBO = rocks related to Serra Geral and Botucatu Formations; BSBO = rocks related to Bauru Group, Serra Geral and Botucatu Formations; BSPI = rocks related to Bauru Group, Serra Geral and Piramboia Formations; ALL = rocks related to Serra Geral, Botucatu-Piramboia Fm, Bauru, Passa Dois, Tubarão and Paraná Groups. a SP=São Paulo, MG=Minas Gerais, MS=Mato Grosso do Sul. b According to Szikszay and Teissedre (1977), Beato et al. (1999), CPRM (2010), Trouw et al. (2003) and Thedeschi et al. (2015):. M.F.S. Casagrande, D.M. Bonotto Applied Radiation and Isotopes 137 (2018) 1–11 7 decays in the 238U and 232Th series. The data comparison of both readings indicated not coincident results. Possible explanation could be related to eventual radon leakage through the beaker wall (Zereshki, 1983). Also, deposition of radionuclides adsorbed to fine/colloidal particles in the container wall/bottom, yielding gamma emissions that not interact with the detector. “Radioactive Disequilibrium” against “Radioactive Equilibrium” readings showed, for 214Bi (1120.29 keV), higher values in the second case in practically all samples (17), as ex- pected from theoretical considerations (Ivanovich and Harmon, 1992). However, for 208Tl (583.19 keV), it happened for 11 samples, which may be a consequence of the more susceptibility of Th for adsorption into colloidal particles (Langmuir and Herman, 1980). Future experi- ments are planned in order to better investigate these and other pos- sible processes involved with the storage time of the water samples. The gross alpha activity estimated from γ-rays of 226Ra (186.21 keV) and 224Ra (240.99 keV) varied between 0.16 and 6.60 Bq/L (Table 5). The WHO (2011) guideline reference value of 0.5 Bq/L was not ex- ceeded only for 5 groundwater samples (4.6%). The gross beta activity estimated from γ-rays of 214Bi (1120.29 keV), 208Tl (583.19 keV) and 40K (1460.83 keV) varied between 0.10 and 6.51 Bq/L (Table 6). In this case, the WHO (2011) guideline reference value of 1 Bq/L was not ex- ceeded for 60 groundwater samples (~56%). In general, the gross beta measurements held elsewhere include a contribution from 40K. How- ever, potassium is an essential element for humans and is absorbed Table 6 Activity concentration for 40K and radionuclides belonging to the 238U and 232Th decay series in the groundwater samples analyzed in this paper. Sample No. 226Ra (Bq/L) 224Ra (Bq/L) 214Bia (Bq/L) 208Tlb (Bq/L) 40Kc (Bq/ L) Gross alphad (Bq/L) Gross betae (Bq/L) 1 0.64 0.67 0.12 1.18 0.12 1.31 1.42 2 0.94 0.70 0.51 1.14 0.50 1.64 2.15 3 0.14 0.78