Radioactivity measurements in the aquatic environment using in-situand laboratory gamma-ray spectrometry

G. Eleftheriou a,b, C. Tsabaris a,n, E.G. Androulakaki a,b, D.L. Patiris a, M. Kokkoris b,C.A. Kalfas c, R. Vlastou b

a Hellenic Centre for Marine Research, Institute of Oceanography, P.O. Box 712, GR-19013 Anavyssos, Greeceb National Technical University of Athens, Department of Physics, Zografou Campus, GR-15780 Athens, Greecec National Centre for Scientific Research “Demokritos”, Institute of Nuclear Physics, GR-15310 Agia Paraskevi, Greece

H I G H L I G H T S

� In-situ underwater gamma-ray spectrometry method is validated by the lab method.� MC simulations using MCNP5 reproduced experimental energy spectra and efficiency.� MDA of the in-situ method was an order of magnitude lower than the lab method.� The in-situ method was sensitive and cost-effective compared to the lab method.

a r t i c l e i n f o

Article history:Received 8 March 2013Received in revised form25 July 2013Accepted 16 August 2013Available online 12 September 2013

Keywords:In-situ gamma-ray spectrometryHPGeInter-calibrationExtended sourcesMDAMCNP5

a b s t r a c t

The in-situ underwater gamma-ray spectrometry method is validated by inter-comparison withlaboratory method. Deployments of the spectrometer KATERINA on a submarine spring and laboratorymeasurements of water samples with HPGe detector were performed. Efficiency calibrations, MonteCarlo simulations and the Minimum Detectable Activity (MDA) estimations were realized. MDAs variedfrom 0.19 to 10.4 (lab) and 0.05 to 0.35 (in-situ) Bq/L, while activity concentrations differed from 7%(for radon progenies) up to 10% (for 40K), between the two methods.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Monitoring of radionuclides in the marine environment is ofsignificant importance, since human activities (e.g. mineralexploration, mining, production of electrical energy) may lead toaccidental or intentional dumping of nuclear radioactive materialsin the sea, resulting in enhanced levels of radioactivity (Matishovand Matishov, 2004). Furthermore, natural radionuclides (mainlyradium and radon isotopes) are used as tracers in hydrogeologicalapplications such as the investigation of submarine groundwaterdischarges (Burnett et al., 2006; Schubert et al., 2006, 2008) andgeophysical studies such as monitoring of radioactive gasses ingroundwater (Tsabaris et al., 2011) as well as in hydrocarbonbearing zones (Gadallah et al., 2010).

Marine radioactivity measurements are performed mainly bytwo approaches (a) sample collection/grabbing in combinationwith on-site or laboratory measurements and (b) in-situ instru-mentation in direct contact with the environment under study.The laboratory method is widely applied using the standardgamma-ray spectrometry for complex geometries betweensource and detector (Saegusa et al., 2000; Venturini andVanin, 1993). The in-situ approach, has reached a high level ofanalytical performance during the last decade, mainly using theunderwater gamma-ray spectrometry (Povinec et al., 2006,2008; Tsabaris et al., 2010, 2012). This approach is also com-bined with stationary floating platforms for real-time datatransmission (Aakenes, 1995; Wedekind et al., 1999; Tsabarisand Ballas, 2005; Tsabaris, 2008a).

Two kinds of detection systems have been mainly utilizedfor in-situ gamma-ray applications, those based on High PurityGermanium (HPGe) semiconductor crystals and those based on

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/apradiso

Applied Radiation and Isotopes

0969-8043/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.apradiso.2013.08.007

n Corresponding author. Tel.: þ30 22910 76410; fax: þ30 22910 76323.E-mail address: tsabaris@hcmr.gr (C. Tsabaris).

Applied Radiation and Isotopes 82 (2013) 268–278

NaI(Tl) scintillators (Povinec et al., 1996). Despite their superiorresolution, HPGe systems exhibit several limitations for autono-mous operation, concerning high power consumption (100–300 W) and short acquisition period (�4 h) of deployment dueto the cooling necessity (with liquid nitrogen). On the contrary, NaI(Tl) detection systems exhibit higher efficiency and much lowerconsumption (�2 W), which render them suitable for prolongedunderwater deployments. However, the disadvantages of theNaI(Tl) systems compared to HPGe ones are their poor energyresolution and the voltage drifts of the amplification signal,which can take place during the measurement, producingoverlapping peaks.

The most important task using an underwater spectrometeris the calculation of the system's detection efficiency. A lot ofeffort has been made during the last years for the experimentalcalibration of detection systems in water tanks, by dilutingcalibrated standard sources (Vojtyla, 2001; van Put et al., 2004;Tsabaris et al., 2008b). Recently, experimental gamma-ray spectrain water were compared with Monte Carlo simulations using theGEANT4 code (Vlastou et al., 2006). A methodology for thedetermination of the detection efficiency of the in-situ systemin water (marine efficiency εm), based on experimental mea-surements and simulation data, is described in Bagatelas et al.,2010.

The calculation of the detection efficiency for extended samplesusing HPGe detectors is still a complex task, since calibrationsources should be prepared with composition and density similarto the real samples', and should be measured in the samegeometry as well. The experimental determination of absolutephoto-peak efficiency can be performed only in specific energiesaccording to the reference calibration source. The detectionefficiency is calculated in a wide range of energies (not only atthe energies of the reference sources), by fitting the experimentaldata with standard mathematical functions from literature(Debertin and Helemer, 1988; Dias et al., 2004). Alternatively,appropriate simulation codes (e.g. MCNP5 (X-5 Monte Carlo Team,2003), GEANT4 (GEANT4 Collaboration, 2003), PENELOPE (Salvatet al., 2006) and FLUKA (Ferrari et al., 2005)), may provideabsolute photo-peak efficiency estimations at each requiredenergy and for any geometry. Among the simulation codes, theMCNP and GEANT codes have been applied to gamma-ray spectro-scopy yielding a satisfactory agreement between experimental andsimulated data (Rodenas et al., 2000; Karamanis et al., 2002;Saegusa et al., 2004; Lépy et al., 2010).

In this work, an inter-comparison exercise has been realized,comparing results of the two methods (in-situ and laboratory),aiming at testing the performance of the in-situ underwatergamma-ray spectrometer KATERINA (Tsabaris et al., 2008b) – theabbreviation comes from the Greek words Innovative Sensor forArtificial and Natural Radioactivity (Tsabaris et al., 2008c) – underrealistic conditions. For this purpose, a method was developed forthe direct measurement of water samples by an HPGe detectorunder two geometries, as described in detail below.

2. Materials and methods

2.1. Study area and field work

The inter-comparison exercise was performed using watermasses (in-situ) and water samples (lab) from the same marineenvironment, namely a submarine groundwater discharge locatedin the Bay of Stoupa, (southwestern Greece). The groundwater isemanated from fissures in the bedrock at a depth of 25 m withhigh flow rates (several decades cm/s) (Tsabaris et al., 2012)forming an aggressive environment. Several deployments of the

in-situ system and a number of sample collections were per-formed by divers during the period from July 2009 till May 2010,covering a full hydrological cycle of the submarine spring. Thesystem KATERINA was positioned and secured immediately abovethe spring, where the crystal position was approximately 2 mabove the seafloor and 23 m below the sea surface. The grabbing ofthe water samples took place exactly from the same point,whenever this was feasible. To avoid radon gas losses andground-seawater exchange, the sample containers should beclosed immediately after grabbing. The containers should be madefrom material durable to withstand depth pressures and with lowradiation absorption for the efficient detection of gamma-rays. Tomeet the requirements, cylindrical bottles with frustum ending(1/4 of the total height) made from heavy-duty borosilicate glasswere selected. Both bottles were 7 mm thick with round opening(20 mm radius) and had diameters of 86 and 126 mm, totalheights of 160 and 240 mm and volumes of 0.6 and 2.2 L,respectively. The samples were transported to the laboratory theday after each sampling (to reduce the 214Bi and 214Pb concentra-tion losses due to 222Rn decay (T1/2¼3.82 d)) and measureddirectly with an HPGe detector without any pre-treatment orpre-concentration. Decay corrections were performed for theradon progenies in order to calculate their activity concentrationsat the sampling date and time.

2.2. In-situ method

The deployments were performed using the autonomous in-situ underwater gamma-ray spectrometer KATERINA. Initially, themeasuring system was energy calibrated and tested for its stabilitywith respect to temperature variations and energy resolution. Thecalibrations were carried out inside a tank having a volume of5.5 m3

filled with freshwater (Tsabaris et al., 2008b).The specific activity r, in Bq/L, is given by the equation

rðBq=LÞ ¼ CPSεmIγ

ð1Þ

where CPS denotes counts per second recorded for each photo-peak, Iγ the gamma-ray emission intensity and εm the marinephoto-peak efficiency, namely the product (εV) of the detectorefficiency ε and the effective volume of water V that is measuredfor each gamma-ray.

The marine photo-peak efficiency εm and the Minimum Detect-able Activity (MDA) of the system have been calculated andpresented in detail elsewhere (Bagatelas et al., 2010). The MDAcalculations for five main radionuclides as derived from laboratorymeasurements in the water tank for a 24 h acquisition time arepresented in Table 2.

2.3. Laboratory method

2.3.1. Experimental set upAll laboratory measurements were carried out at the facilities

of Hellenic Centre for Marine Research (HCMR). The detectionsystem consists of a p-type coaxial HPGe detector (ORTEC GEM-FX8530P4) with 50% nominal relative efficiency and with aresolution of 2.15 keV at 1.33 MeV along with a computerizedMCA system (CANBERRA Model 8715) for the data acquisition.A lead shielding (58.5 mm thick and 21 mm height) surroundedthe detector in order to reduce the ambient gamma-raybackground.

The calibration of the HPGe system for two extended geome-tries was performed, following a similar methodology, asdescribed in Tsabaris et al., 2007. Two radioactive referencesources of Europium 152Eu/154Eu were produced with the same

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278 269

geometries as the sampling containers with volumes of 0.6 and2.2 L. The calibration procedure included the following steps:

(i) The energy calibration and absolute efficiency calculation of thedetector for point-source geometry were performed; using areference point source of Europium 152Eu/154Eu (97% and 3%),placed at a distance of 25 cm from the detector's window. After2 h of acquisition, the analysis of the spectrum led to the energyand efficiency calibration of the detector for point sources.

(ii) A small quantity ((1.0070.02) ml) of diluted Europium152Eu/154Eu (93% and 7%) reference source (Amersham Co.)was retained by a syringe. The syringe containing the liquidsource was measured similarly to a point source (placed at a

distance of 25 cm from the window of the HPGe detector) andwas found to be (35.170.7) Bq/L. Activity corrections due tothe syringe geometry (Gardner et al., 1980) and the radiationattenuation into the material (Gilmore, 2008) have beencalculated analytically and found to be less than 0.1%.

(iii) The syringe content was diluted again with distilled waterand specific quantities were added to the containers (priorfulfilled with distilled water). The residual activity in thesyringe was then measured in order to determine the activityactually injected-dropped into the inert material. In such away, two extended radioactive sources with water volumes of0.6 and 2.2 L were produced with the same activity concen-tration ((2.570.1) Bq/L).

Fig. 1. Geometrical models of the two measurement systems of 0.6 and 2.2 L container, as seen with the MCNP5 Visual Editor, and the cross-section of the detector's crystalwith the aluminum casing, provided by the manufacturer (reproduced under permission of ORTECs).

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278270

The solutions were homogenized mechanically and measuredfor 1 h in contact geometry with the detector's window (see Fig. 1).Background samples containing only distilled water were alsomeasured to subtract the contribution of the laboratory ambientradiation from the experimental spectra. Additional measure-ments of a specific proportion of KCl diluted in distilled waterwere performed for both geometries (40K reference sources with(143.770.1) Bq/L activity concentration), in order to obtain preciseefficiency values for the 40K photo-peak at 1460.8 keV.

A cross-check measurement of the 0.6 L container has also beenperformed at the laboratory of the Physics Department at theNational Technical University of Athens (NTUA). The measurementset up was exactly the same (detector and electronics), but thedetector was placed horizontally inside a cylindrical shieldingconsisted of an external lead layer (58.5 mm) and an internal Cdlayer (2.5 mm) covered inside with Cu (2.5 mm). The backgroundradiation was reduced than the corresponding one in the HCMRconfiguration by a factor of 2. However, due to dimensionalrestrictions of the shielding, the containers of the 2.2 L could notbe measured in the NTUA set up.

Once the efficiency was determined, the measured countsunder any photo-peak were converted to activity concentrationdata (in Bq/L) using the equation

AðBq=LÞ ¼ CPSεVIγ

ð2Þ

where CPS denotes the counts per second recorded for each photo-peak, ε is the efficiency of the specific photo-peak, V the volume(in L) of the water sample, and Iγ is the gamma-ray emissionprobability.

2.3.2. Simulation set upThe absolute efficiency of the HPGe detector for the extended

sources geometry has been experimentally determined for tengamma-ray energies emitted from the 152Eu radionuclide. Toextend the efficiency values in a wide range of energies, MonteCarlo simulations were performed using the MCNP5 code. TheMCNP5 code simulates the trajectory of elementary and secondaryproduced particles as they interact with matter till they are fullyabsorbed in a medium (X-5 Monte Carlo Team, 2003). In thepresent work, the code was implemented for: (a) the validation ofthe Monte Carlo model by reproducing the experimentally derivedspectra of 152Eu/154Eu reference sources, and (b) theoretical esti-mations of absolute photo-peak efficiency in a wide range ofenergies.

The experimental geometry set up was reproduced in theMCNP5 code input including detector's structural details, lead

shielding and one of the two extended sources. Geometrical andphysical characteristics of the materials used in the simulationmodel were set to the values provided by the manufacturer.Additional measurements were performed for the calculation ofthe containers' density. The detailed MCNP5 set up model for the0.6 L container along with the detector information provided bythe manufacturer are illustrated in Fig. 1.

The active volume of the detector crystal was determinedassuming an inactive Germanium dead layer, estimated from thecomparison between experimental and simulated spectra of thepoint-source geometry (25 cm distance from the crystal), aftersuccessive code modifications and tests. Good agreement (o5%)between experimental and simulated spectra was achieved fordead layer values of 12 μm and 600 μm, for the front and thelateral sides of the crystal, respectively.

The simple physics model provided by the code was chosen, inwhich the trajectories of secondary particles were recorded withthreshold energy of 1 keV for both electrons and photons. Thephoto-peaks of 152Eu and 154Eu with intensities above 1% wereincluded in the code input with emission probability normalizedto unity, taking into account the decay scheme and the isotopicratio of the 152Eu/154Eu reference source. The energy binning wasset to 4096 channels. The resolution of the detector system wasreproduced in the simulation spectra, using the Gaussian EnergyBroadening (GEB) option of the code, with respect to the valuesderived from the experimental calibration. The total number ofgenerated events was set to 107, yielding a relative error of lessthan 1% for each photo-peak centroid.

3. Results

3.1. Development of the laboratory method

3.1.1. Simulated spectraSimulated spectra have been produced for point source as well

as for two extended geometries of 152Eu/154Eu reference sources.A typical experimental spectrum, as obtained from the extendedsample of reference source (volume of 2.2 L), is depicted in Fig. 2along with the simulated one.

The quality of the simulated spectrum is satisfactory (seeFig. 2a, c and d), while the net area of the simulated photo-peaks deviates less than 10% from the experimental ones in themeasured energy range. The asymptotic form of the experimentalphoto-peaks at higher energies (see Fig. 2e and f) may beattributed to the influence of the electronics set up (pole-zerocancellation) that has not been included in the simulation model.

Fig. 2. Experimental (black line) and simulated (gray line) spectra of 152Eu/154Eu reference source measured in 2.2 L container geometry. Captions a–f correspond to variousenergy regions.

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278 271

Nevertheless, a number of photo-peaks in the experimentalspectrum do not appear in the simulated one (Fig. 2b and f):(a) the peak at 32 keV was found only in the experimentalspectrum since it is superimposed by X-rays of the La (which isused for the chemical separation of 152Eu) and its traces areencountered in the final reference source solution, (b) the 33 keVand at 48 keV photo-peaks of the experimental spectra are due tothe superposition of X-rays and low intensity (o1%) gamma-rays,and they were not included in the simulation spectra and (c) the1447.8 keV photo-peak derived from the coincidence summingeffect of photons of different energies emitted simultaneouslywithin the solid angle subtended by the detector could not besimulated, due to the inherent structure and limitation of theMCNP5 code.

Moreover, the Pb X-ray photo-peaks in the energy rangebetween 65 and 85 keV are slight underestimated at the simulatedspectrum. This can be attributed to the default values of MonteCarlo codes for specific modeling parameters that strongly influ-ence the behavior of simulations leading often to underestima-tions (Karfopoulos and Anagnostakis, 2010). Underestimations inthe case of low energies are even stronger with the use ofextended sources (Hardy et al., 2002; Vargas and Guerra, 2006).For these reasons, simulations only for energies above 90 keV upto 2200 keV have been carried out in order to estimate the specificabsolute photo-peak efficiency values, covering the energy rangeof the gamma-rays of the dominant radionuclides found in theaquatic environment. In addition, the energy of 90 keV can beconsidered also the feasible lower limit of spectrometry for NaIdetectors and therefore for the in-situ method described here.

3.1.2. Efficiency calibrationThe experimental absolute photo-peak efficiency values have

been compared with the Monte Carlo calculations, for ten discretegamma-ray energies of the most intensive photo-peaks of 152Eu.Coincidence summing effect corrections have been taken intoaccount in the analysis of the experimental spectra, using theEFFTRAN code (Vidmar et al., 2011). The coincidence summingcorrections were of the order of 10% for both extended geometries,varying from �1.7% (for 1085.84 KeV gamma-ray) up to 20%(for 867.39 keV gamma-ray), while for the point like source thecoincidence summing effect was negligible for the employedsource to detector distance (25 cm). The results for three differentgeometries (one point source and two extended geometries) arepresented in Table 1, along with the relative uncertainty of thecalculations and the variance between the two methods.

The results of the comparison exhibit a reasonable agreementbetween experimental measurements and theoretical estima-tions. The variance (absolute values) between experimental and

simulated absolute photo-peak efficiency ranged from 0.4 to 8.4%without any systematic positive or negative trend with respect tothe energy of the gamma-rays. The increased deviations for thelow-energy gamma-rays reveal a possible code failure in treatingself-absorption effects in extended sources. This trend has beenalready pointed out in the literature and could be probablyattributed to uncertainties in the parameterization of the photo-electric effect data in MCNP and the use of linear interpolation oflogarithmic photon cross section data even below the K-shellabsorption edge of 109.7 keV (Karamanis et al., 2002; Karamanis,2003).

The photo-peak detector efficiency curves using the twoextended geometries were produced for a broad range ofgamma-ray energies. Various runs of the MCNP5 code for photonswith energies ranging from 90 to 2200 keV have been performedand the simulated absolute photo-peak efficiency values, alongwith the experimental ones, were fitted by the empirical functionof the type (Debertin and Helemer, 1988):

ε¼ aEb

cþEdð3Þ

where, ε is the absolute photo-peak efficiency, E is the gamma-rayenergy and a, b, c, d are free parameters. The uncertainties of thefree parameter values were less than 3%, while the efficiencycurves χ2 values were more than 0.95 and the mean deviationbetween fitted and experimentally determined values were 6% forboth curves.

The efficiency curves for the two extended geometries, pre-sented in Fig. 3, indicate that the efficiency of the small geometryexhibits higher values mainly due to the difference of the solidangles. Moreover, in the small geometry the path of gamma-raysinside the sample is shorter, thus reducing the phenomenon ofself-absorption.

3.1.3. Minimum detectable activityThe capability of the measuring systems to detect marine

radioactivity depends on the detector efficiency for the specificconfiguration (and consequently on the energy of the detectedgamma-ray), the background contribution from the sample, aswell as, from cosmic radiation, water sample volume and acquisi-tion time. Based on the Currie formula (Currie, 1968), the mini-mum detectable activity (MDA) of volumetric water samples in Bqper unit of volume is given by the equation

MDAðE; tÞ ¼ LDεðEÞVIγt

ð4:1Þ

where ε(E) is the absolute efficiency at the specific gamma-rayenergy E, V the volume of the sample (volume units), Iγ theemission probability of the detected gamma-ray, t is the

Table 1The variation between experimental and simulated (MCNP5) absolute efficiency values of 152Eu (for the ten most intensive gamma-rays), along with their relativeuncertainties, for a point (25 cm distance from the detector) and two extended (0.6 and 2.2 L containers) reference sources.

152Eu Absolute efficiency

Energy (keV) MCNP5 point source Exper. point source Var. (%) MCNP5 0.6 L Exper. 0.6 L Var. (%) MCNP5 2.2 L Exper. 2.2 L Var. (%)

121.78 4.59E�05 (1.4%) 4.23E�05 (3.2%) 7.9 1.62E�02 (0.4%) 1.77E�02 (4.0%) 8.4 0.99E�02 (0.4%) 1.01E�02 (2.0%) 0.9244.7 2.71E�05 (1.6%) 2.65E�05 (4.6%) 2.0 1.17E�02 (0.9%) 1.16E�02 (4.2%) �1.1 7.25E�03 (0.4%) 6.72E�03 (2.2%) �7.9344.28 1.93E�05 (1.5%) 1.86E�05 (3.6%) 3.9 8.96E�03 (0.6%) 9.35E�03 (4.1%) 4.2 5.13E�03 (0.5%) 5.16E�03 (2.0%) 0.6443.96 1.49E�05 (2.1%) 1.47E�05 (1.8%) 1.3 6.59E�03 (0.9%) 6.64E�03 (5.4%) 0.8 4.50E�03 (0.5%) 4.49E�03 (2.9%) �0.3778.92 8.59E�06 (1.7%) 8.45E�06 (1.8%) 1.6 4.78E�03 (1.2%) 4.88E�03 (4.2%) 1.8 2.99E�03 (0.7%) 2.92E�03 (2.2%) �2.3867.39 7.63E�06 (2.5%) 7.87E�06 (2.5%) �3.2 4.42E�03 (2.2%) 4.45E�03 (3.1%) 0.7 2.87E�03 (0.7%) 2.76E�03 (3.1%) �4.2964.05 6.97E�06 (1.6%) 7.06E�06 (1.5%) �1.3 4.13E�03 (1.3%) 4.10E�03 (4.2%) �0.6 2.59E�03 (0.7%) 2.60E�03 (2.2%) 0.3

1085.84 6.25E�06 (1.7%) 6.45E�06 (2.1%) �3.2 3.82E�03 (1.7%) 3.84E�03 (4.4%) 0.5 2.41E�03 (0.7%) 2.50E�03 (2.3%) 3.31112.09 5.88E�06 (1.7%) 6.23E�06 (1.7%) �6.0 3.76E�03 (1.5%) 3.73E�03 (4.3%) �0.8 2.36E�03 (0.7%) 2.34E�03 (2.2%) �1.01408.02 4.83E�06 (1.6%) 5.07E�06 (1.6%) �4.9 3.19E�03 (1.4%) 3.16E�03 (4.1%) �0.9 1.98E�03 (0.8%) 2.00E�03 (2.1%) 1.1

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278272

acquisition time (in sec) and LD is the detection limit (in counts)given by the equation (Gilmore, 2008)

LD ¼ k2þ2ks0 ¼ 2:71þ4:65B0:5 ðfor k¼ 1:645; 95% confidence levelÞð4:2Þ

where k is the single-sided interval factor corresponding to thechosen confidence level, s0 is the standard deviation of thedistribution (s02¼2Β) and B is the area of background continuumunder the photo-peak (in counts).

The mean MDA values for the two sample geometries werecalculated for several characteristic radionuclides found in themarine environment (214Pb, 214Bi, 208Tl, 228Ac, 40K and 137Cs). Thesystematic analysis of the background spectra of distilled waterprovided data for the most intensive photo-peaks of these radio-nuclides that were elaborated with Eq. (4.1). The results for theextended geometries from both laboratories (HCMR and NTUA)are presented in Table 2 along with the MDA values of the in-situsystem (for 24 h acquisition time). As concerns the laboratorymethod, the lower values of MDA were deduced for 208Tl and 137Csfor both HPGe systems, due to their negligible concentrations inthe surrounding ambient (atmosphere, laboratory materials). TheMDA values between the two systems for the sample of 0.6 L differby a factor of �2 for all radionuclides, due to the more effectiveshielding at the NTUA lab. Also, the comparison between the twogeometries measured at the same lab (HCMR) suggests that thelarger geometry (2.2 L) provides an MDA improvement of about 7%for 208Tl, 13–15% for 214Pb, 214Bi, 137Cs, 228Ac and 130% for 40K.

Nevertheless, this geometry leads to higher MDA values (from 31to 53%) than the initial measurement at the NTUA lab, except for40K where the MDA value is lower by 46%. Although the detectionlimit of low energy gamma-ray emitters is significantly higherusing the big geometry, the measurement at 1460.8 keV exhibitslower MDA value making this geometry more sensitive to detect40K variations. Finally, the MDA values of the in-situ methodobtained in freshwater environment (with minimum radionu-clides presence) were found to be one order of magnitude lowerthan the corresponding ones using lab methods.

The MDA dependence on time was also examined for the largegeometry (2.2 L) by setting the acquisition time from a few hoursto a few days (see Fig. 4). The results reveal an exponentialreduction of MDA with respect to the period of measurement.This trend can be attributed to the proportional relation of thelimit of detection (LD) with the square root of number of back-ground counts (B), as presented in Eq. (4.2). As expected, since thebackground counts (B) are proportional to the acquisition time t,the relation of MDA with respect to the acquisition time iscalculated using the following fitting function: MDAðE; tÞ ¼ΑðEÞt�0:5. According to this study, one day of measurement seemsto be a sufficient acquisition period, since any further increase ofthe measuring time provides a negligible improvement in theMDA values.

3.1.4. Laboratory measurementsLaboratory measurements were performed for the two

extended sample geometries (0.6 and 2.2 L). In total, four cam-paigns took place for water sampling, while three more(at November 2009, February and April 2010) did not meet theinter-comparison requirements. The measurements concernedthe natural radionuclide 40K, which is dominant in the marineenvironment with high concentrations (�12 kBq/L), and variousprogenies of 238U and 232Th radioactive series. More specifically,228Ra was measured via 228Ac (at 911.2 keV), 224Ra via 212Pb (at238.6 keV) and 208Tl (at 583.2 keV); and 222Rn gas via its progenies214Pb (at 351.9 keV) and 214Bi (at 609.3 keV). The spectral analysiswas performed with the SPECTRW software package (Kalfas andTsoulou, 2003; Kalfas, 2011) taking into account the calculateddetector efficiency. The activity concentration results for thesuccessful campaigns are given in absolute units (Bq/L) in Table 3.

Table 2Minimum detectable activities for five radionuclides, as calculated from their mostintense gamma-rays (in keV), for the lab (0.6 and 2.2 L container) and the in-situconfiguration (acquisition time of 24 h).

Method MDA (Bq/L)

214Pb 208Tl 137Cs 214Bi 228Ac 40K(351.9) (583.2) (661.7) (609.3) (911.2) (1460.8)

Lab 0.6 L (NTUA) 0.63 0.30 0.19 0.68 1.11 6.610.6 L (HCMR) 1.42 0.62 0.31 1.59 2.16 10.402.2 L (HCMR) 1.33 0.54 0.27 1.41 1.90 4.64

In-situ KATERINA 0.11 0.05a 0.05 0.09 0.14 0.35

a As calculated from the 2614.5 keV gamma-ray.

Fig. 4. Minimum detectable activities calculations for specific radionuclides (208Tl,214Pb, 214Bi, 228Ac and 40K) as a function of time, using volume samples of 2.2 Lcontainer. All data series are fitted by the function y¼ axb (adj. R

240:99).

Fig. 3. Absolute efficiency calibration curves of the lab detection system forextended volumetric sources (0.6 and 2.2 L) in contact with the detector's window.

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278 273

In Fig. 5 a typical net spectrum of a 0.6 L sample container isshown, measured at the NTUA laboratory. The net spectrum isderived after the background subtraction, as measured using abackground sample of distilled water. In the spectrum the con-tribution of the most intense gamma-rays of 214Pb (at 351.9 keV)and of 214Bi (at 609.3 keV) is evident. The activity concentrationwas measured (1.0970.06) and (1.1770.10) Bq/L for 214Pb and214Bi, respectively. The energy peaks of 228Ac (911.2 keV) and 208Tl(583.2 keV) were not shaped as Gaussians and thus they were notincluded in the analysis due to poor statistics. As concerns 40K(measured at 1460.8 keV), the activity concentration was below

the detection limit (o6.61) Bq/L. The same sample of 0.6 L wasalso measured at HCMR but the activity concentrations of theaforementioned radionuclides were lower than the MDA of thedetection system.

However, the containers of the 2.2 L exhibits lower MDA valuein comparison with the smaller geometry 0.6 L for 40K. Since, theprecise measurement of 40K concentration is critical even in lowsalinity waters (e.g. in SGD studies providing information for themixing process between the emanating fresh/groundwater andthe seawater), the greater geometry of 2.2 L was finally adopted forthe subsequent measurements.

Table 3Inter-comparison results from the submarine groundwater spring at Stoupa using the lab (0.6 and 2.2 L container) and the in-situ method.

Date Nuclide Activity (Bq/L) Variation (%) Comments

HPGe KATERINA

10/10/2010 214Pb 1.1070.14 1.0370.08 6.7 0.6 L container measured at NTUA lab(the same measurement at HCMR set up was belowthe detection limit)

208Tl o0.30 0.2170.02214Bi 1.0270.08 1.0170.04 1.2228Ac o1.11 0.1770.0940K o6.61 6.2770.31

18/01/2011 214Pb 1.9570.11 2.1070.29 �7.2 2.2 L container measured at HCMR lab (sampling in �2 mdistance from the position of the KATERINA detector)208Tl o0.54 0.1670.02

214Bi 2.0870.23 2.0070.28 4.1228Ac o1.90 1.2270.1340K 6.9970.42 7.8070.43 �10.4

23/03/2011 214Pb 3.2670.18 3.2770.30 �0.5 2.2 L container measured at HCMR lab208Tl o0.54 0.3070.04214Bi 4.1770.14 4.2070.20 �0.8228Ac 3.4170.26 3.370.32 3.540K 5.8770.43 6.4070.40 �8.3

12/05/2011 214Pb 1.3870.07 1.3570.12 2.0 2.2 L container measured at HCMR lab208Tl o0.54 0.2170.02214Bi 1.5570.12 1.4570.15 6.9228Ac o1.90 o0.1440K 7.8370.45 7.470.52 5.8

Fig. 5. Measurement of extended sample using 2.2 L container. The net spectrum is zoomed at the photo-peaks of 214Pb (350. 93 keV), 214Bi (609.13 keV) and 40K(1460.83 keV).

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278274

3.2. In-situ measurements

The in-situ NaI detector KATERINA was deployed for severalmonths at the submarine groundwater discharge site (Tsabariset al., 2012), where variations of activity concentration wereprovided for key natural radionulcides (i.e. 40K, 214Bi, 208Tl).Gamma-ray spectra were acquired every 12 h and buffered intoan internal memory. External battery packs provided autonomousoperation for approximately 1 month. In every experimentalsurvey the system was recovered, the data were obtained, thebattery was replaced, and then the system was re-deployed bydivers as close as possible to the groundwater source. The choiceof the inter-comparison measurements was accomplished accord-ing to the following criteria and constraints: (a) by selecting,whenever a prediction was possible, calm and hydrologicallyuniform periods with constant meteorological conditions, in orderto reliably compare the integrated results from the in-situ methodwith the instant results from the laboratory measurements, (b) byselecting a proper deployment location, easily accessible by thedivers and (c) by taking into account the detector autonomyspecifications.

For the inter-comparison purposes integrated spectra of 24 h,obtained by the in-situ system, were analyzed using the SPECTRWsoftware, yielding daily mean values of the characteristic radio-nuclides. A characteristic acquired spectrum from the deploymentof October 2010 is shown as solid line in Fig. 6. The photo-peaks ofradon progenies 214Bi at 609.3 keV and 214Pb at 351.9 keV areclearly observed and from their analysis the activity concentrationwas deduced: (1.2770.09) and (1.0370.08) Bq/L for 214Bi and214Pb, respectively. The three photo-peaks of 228Ac at 911.2 (29%),964.6 (5.45%) and 968.9 (17.4%) keV, overlap with each other, sothey were analyzed as one, using the sum of the intensities as thetotal intensity of the overlapping photo-peak. The activity con-centration of 228Ac was measured (0.1570.08) Bq/L. The max-imum energy of the detection system was prior the deploymentadjusted to 2800 keV to detect the 2614.5 keV gamma-ray of 208Tl,

a decay product of 228Ac with a transition probability of 35.4%.So, the measured activity of 208Tl was normalized to the parentnuclide (228Ac) assuming radioactivity equilibrium of all radio-nuclides between 228Ac and 208Tl. 208Tl activity was found(0.1270.05) Bq/L, which is in good agreement with the 228Acconcentration within the statistical uncertainty of the measure-ment. The activity of 40K was calculated as described in literature(Tsabaris et al., 2012) by subtracting the contribution of 214Bi in theenergy region of the 1460.8 keV gamma-ray and it was foundequal to (8.270.4) Bq/L. This value is rather low for the marineenvironment, indicating a high mixing rate of the emanatingfresh-groundwater with the surrounding seawater. All the afore-mentioned results, along with the results from the other inter-comparison deployments are summarized in Table 3.

4. Discussion

4.1. Monte Carlo simulations

The simulation of the calibration procedure with a certifiedMonte Carlo code is a useful auxiliary tool for environmentalradioactivity laboratories, since a number of different geometriesof environmental extended samples can be examined. The simu-lated models tested here were proven satisfactory for the needs ofroutine radioactivity measurements of environmental samples, asthe mean uncertainty of the variance between experimental andsimulated values for the absolute photo-peak efficiency is 5%. Formore accurate measurements at lower energies (o90 keV), amore detailed investigation of the parameters used in the simula-tion is required for the optimization of the geometrical character-istics of the system (e.g. distance between crystal and Al window,diameter and thickness of the crystal's inactive zone) with addi-tional experimental methods like radiography (Dryak and Kovar,2006) and scanning of the detector with collimated gamma-raysources (Keyser, 2004).

Fig. 6. In-situ measurement at 8/10/2009, for 24 h acquisition time. In the spectrum the photo-peak analysis is depicted for 214Pb (at 351.93 keV), 214Bi (at 609.31 keV) and40K (at 1460.83 keV) using second order polynomial fitting for the background and asymmetric (2%) Gaussian fitting for the photo-peak area.

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278 275

4.2. MDA and volumetric efficiency comparison

An important rating of the measuring methods is their sensi-tivity, which is expressed by the minimum detectable activity. TheMDA using the lab-based method (HPGe detector) and the in-situmethod (KATERINA system) are given in Table 2 for variousradionuclides (214Pb, 214Bi, 208Tl, 228Ac, 40K and 137Cs). The lab-based method exhibits higher MDAs, due to the smaller volume ofthe water samples, when no pre-treatment or pre-concentrationtook place. Depending on the gamma-ray energy and the measur-ing set up, the MDA values of the laboratory method were found tobe 4–30 times higher than those of the in-situ method. The MDAof radon and thoron progenies varied from 0.6 to 2.6 Bq/L for thethree laboratory configurations, while for the in-situ methodranged from 0.05 to 0.14 Bq/L. Smaller differences occur for the137Cs (�5 times lower MDA values for the in-situ method), whilethe greatest variances are observed for the 40K (13, 19 and 30 timeslower MDA values compared to the 2.2 L (HCMR), 0.6 L (NTUA) and0.6 L (HCMR) configurations, respectively).

The combination of effective shielding and increased samplevolumes can result to even more reduction of MDA values (seeTable 2). Alternatively, the laboratory method can be furtherimproved using different geometry for the sample container in amanner that more sample volume is placed adjacent to the crystal,surrounding it. Such container geometry may be utilized using aproperly modified Marinelli geometry beaker/container madeform a durable material (like borosilicate glass) so it will becapable to withstand the depth pressure during the procedure ofsampling.

The volumetric efficiency (εV) is defined as the ratio of detectedto total counts emitted in a specific volume of water (Tsabariset al., 2008b). This quantity is characteristic of the measuringmethod, depends on the gamma-ray energy and provides quanti-tative results (in Bq/m3 or Bq/L) of the gamma-ray emitter. Thevolumetric efficiency curves of the two extended geometries aredepicted in Fig. 7 together with the efficiency curve of the in-situmethod. The volumetric efficiency curves of the laboratory meth-ods are significantly lower than the efficiency curve of the in-situmethod. This could be attributed mainly to the fact that the in-situdetector acquires gamma-rays from a greater volume of water(effective volume) in comparison with the laboratory detector.This dependence of volumetric efficiency on the water volume isclearly observed in laboratory measurements, since the volumetricefficiency for the large geometry (2.2 L) is almost two-times

greater than the smaller one (0.6 L), even though the absolutephoto-peak efficiency exhibit the opposite behavior.

4.3. Inter-comparison of the results

The main scope of the study was the comparison of the resultsobtained by two different methods: one based on continuous in-situ measurements and the other based on a laboratory measure-ment of groundwater samples by means of an HPGe detector. Animportant problem for a consistent comparison was the fact thatany sample analysis reflects the characteristics of the spring at thespecific moment of sampling while the in-situ system providesintegrated values over time (in this study over an acquisitionduration of 24 h). Indeed, the characteristics of the submarineground water spring may significantly change day by day. Toovercome the problem, the sampling dates were appropriatelychosen during hydrological uniform periods, without significantmeteorological changes which may influence the behavior of thespring. The sampling dates that have been chosen for the inter-comparison purposes were based on the time-series data listed in(Tsabaris et al., 2012), in such a way to meet the requirement ofsteady behavior of the spring within at least the first 24 h.

The results of the two methods are shown in Table 3. Thedifferences vary from 0.5 to 10.4% (in absolute values). Comparingthe activity concentration of radon progenies (214Pb and 214Bi), asatisfactory agreement was found between the results of the twomethods as their variations (from �7 to 7%) lie to be within theuncertainties of the measurements (5–14% both for the lab and thein-situ method). Due to the low 232Th progeny concentrations,proper comparison for 228Ac could be only achieved for onemeasurement and good agreement (variance of 3.5%) wasobserved between the two methods. Concerning 40K, the variancesbetween the measured activities were slightly higher from themeasurement uncertainties (�6%) for the January and Marchsampling periods (�8 and �10%, respectively). This could bemainly attributed to the fact that even small changes of thespring's behavior (e.g. emanation flow rate) can notably alternatethe salinity of the groundwater, resulting in considerable change ofthe 40K concentration. Especially for the January sampling period,the greater difference is due to the high flux of the emanatinggroundwater, obliging the divers to collect the samples at adistance of �2 m far from the position of the in-situ system.

Taking into account the experimental conditions and theresults of the inter-comparison, the main advantages and draw-backs of the lab and the in-situ method are summarized in Table 4.According to the MDA values, especially those of 137Cs, the in-situmethod is considered sufficient to detect directly increased activ-ity concentrations of potential accidental releases. Taking intoaccount that the activity concentration of 137Cs in the water variedfrom 0.03 to 0.5 Bq/L at the Black Sea for the time period of 1986–1991 (Valukovski et al., 1994), both methods turn out to beefficient for the determination of radioactive releases connectedto contamination in aquatic environments. However, for monitor-ing the evolution of pollutants concentrations for a long period oftime only the in-situ method provides sufficient sensitivity.

5. Conclusions

In this work an inter-comparison exercise has been performedusing different methods of measuring radionuclides in the marineenvironment (in-situ and lab-based gamma-ray spectrometrywithout pre-concentration of water samples). The procedureimplementation includes measurements in the field (in-situ) andin the laboratory (lab-based method) along with Monte Carlo

Fig. 7. Volumetric efficiency (εV) curves of the laboratory set up for 0.6 and 2.2 Lcontainers; and the in-situ KATERINA detector.

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278276

simulations. The results of this exercise are considered comparableand satisfactory, within their uncertainties.

The MDA of the in-situ method using KATERINA system (an7.5�7.5 cm NaI(Tl) scintillator crystal mounted along with theappropriate electronics inside a 85�500 mm cylindered housingfrom acetal) was found to be 4–30 times lower than that of thelaboratory method, while the laboratory volumetric efficiencycurves were more than one order of magnitude lower comparedto the efficiency curve of the KATERINA detector. The resultsrevealed that the in-situ method using the KATERINA detectionsystem is an effective detection method, if the observed photo-peaks are well separated in the energy spectrum. Moreover, thedeveloped simulation model is valuable for time and materialsaving and its computational-theoretical investigation offers reli-able absolute photo-peak efficiency estimations for differentgeometries and materials of extended samples. Particularly, inthe framework of this study, it would be useful to apply theaforementioned methodology in accordance with other samplingprocedures, like the use of Marinelli beaker in combination withappropriate in-situ underwater sampling.

As a final conclusion the in-situ method was proved moresensitive, direct and cost-effective in comparison with the labora-tory based method. Even with the disadvantage of lower resolu-tion of the NaI(Tl) crystal, the in-situ spectrometer can beefficiently utilized for precise measurement of radionuclides inthe marine environment.

Acknowledgments

This work was partially supported by the Local Authorities ofMessinia Prefecture. The authors would like to thank the diversof the HCMR underwater team K. Katsaros, V. Stasinos andTh. Photopoulos and the captain of the vessel D. Eksarchouleasfor their kind assistance.

References

Aakenes, U.R., 1995. Radioactivity monitored from moored oceanographic buoys.Chem. Ecol. 10, 61–69.

Bagatelas, C., Tsabaris, C., Kokkoris, M., Papadopoulos, C.T., Vlastou, R., 2010.Determination of marine gamma activity and study of the minimum detectableactivity (MDA) in 4pi geometry based on Monte Carlo simulation. Environ.Monit. Assessment 165, 159–168.

Burnett, W.C., Aggarwal, P.K., Aureli, A., Bokuniewicz, H., Cable, J.E., Charette, M.A.,Kontar, E., Krupa, S., Kulkarni, K.M., Loveless, A., Moore, W.S., Oberdorfer, J.A.,Oliveira, J., Ozyurt, N., Povinec, P., Privitera, A.M.G., Rajar, R., Ramessur, R.T.,Scholten, J., Stieglitz, T., Taniguchi, M., Turner, J.V., 2006. Quantifying submarinegroundwater discharge in the coastal zone via multiple methods. Sci. TotalEnviron. 367, 498–543.

Currie, L.A., 1968. Limits for qualitative detection and quantitative determination.Anal. Chem. 40, 586–593.

Debertin, K., Helemer, R.G., 1988. Gamma and X-ray spectrometry with semicon-ductors detectors. North Holland, Amsterdam, pp. 219–223.

Dias, M.S., Cardoso, V., Vanin, V.R., Koskinas, M.F., 2004. Combination of nonlinearfunction and mixing method for fitting HPGe efficiency curve in the 59–2754 keV energy range. Appl. Radiat. Isot. 60, 683–687.

Dryak, P., Kovar, P., 2006. Experimental and MC determination of HPGe detectorefficiency in the 40–2754 keV energy range for measuring point sourcegeometry with the source-to-detector distance of 25 cm. Appl. Radiat. Isot.64, 1346–1349.

Ferrari, A., Sala, P.R., Fasso, A., Ranft, J., 2005. FLUKA: a multi-particle transportcode. CERN-2005-10, INFN/TC_05/11, SLAC-R-773. Available at: ⟨http://www.fluka.org/fluka.php⟩.

Gadallah, M.M., El-Terb, R.A., El-Kattan, S.M., El-Alfy, I.M., 2010. Spectrometry andreservoir characteristics of rudeis formation in Belayim marine oil field, Gulf ofSuez, Egypt. J. KAU Earth Sci. 21, 171–199.

Gardner, R.P., Verghese, K., Lee, H.M., 1980. The average solid angle subtended by acircular detector coaxial to a circular isotropic source. Nucl. Instrum. Methods176, 615–617.

GEANT4 Collaboration, 2003. Geant4—a simulation toolkit. Nucl. Instrum. MethodsPhys. Res., Sect. A 506, 250–303. Available at: ⟨http://geant4.cern.ch/⟩.

Gilmore, G.R., 2008. Practical Gamma-ray Spectrometry, second ed. J. Wiley andSons, New York, pp. 117–158.

Hardy, J.C., Jacob, V.E., Sanchez-Vega, M., Effinger, R.T., Lipnik, P., Mayes, V.E., Willis,D.K., Helmer, R.G., 2002. Precise efficiency calibration of an HPGe detector:source measurements and Monte Carlo calculations with sub-percent precision.Appl. Radiat. Isot. 56, 65–69.

Kalfas, C.A., Tsoulou, E., 2003. SIMULATE program: a gamma ray spectroscopy tool.Nucl. Instrum. Methods Phys. Res., Sect. A 500, 386–390.

Kalfas, C.A., 2013. A modern nuclear spectroscopy software package. Adv. Nucl.Phys., Proc. 20th Hellenic Conf. on Nucl. Phys., Athens, Greece 159 ⟨http://nuclpart.phys.uoa.gr/HNPS/Files/ANP2011.pdf⟩.

Lépy, M.C., Altzitzoglou, T., Anagnostakis, M.J., Arnold, D., Capogni, M., Ceccatelli, A.,de Felice, P., Dersch, R., Dryak, P., Fazio, A., Ferreux, L., Guardati, M., Han, J.B.,Hurtado, S., Karfopoulos, K.L., Klemola, S., Kovar, P., Lee, K.B., Ocone, R., Ott, O.,Sima, O., Švec, Sudar S., Chau Van Tao, A., Thien Thanh, Tran, Vidmar, T., 2010.Intercomparison of methods for coincidence summing corrections in gamma-ray spectrometry. Appl. Radiat. Isot. 68, 1407–1412.

Karamanis, D., Lacoste, V., Andriamonje, S., Barreau, G., Petit, M., 2002. Experi-mental and simulated efficiency of a HPGe detector with point-like andextended sources. Nucl. Instrum. Methods Phys. Res., Sect. A 487, 477–487.

Karamanis, D., 2003. Efficiency simulation of HPGe and Si(Li) detectors in γ- and X-ray spectroscopy. Nucl. Instrum. Methods Phys. Res., Sect. A 505, 282–285.

Keyser, R.M., 2004. Resolution and sensitivity as a function of energy and incidentgeometry for germanium detectors. Nucl. Instrum. Methods Phys. Res., Sect B213, 236–240.

Karfopoulos, K.L., Anagnostakis, M.J., 2010. Parameters affecting full energy peakefficiency determination during Monte Carlo simulation. Appl. Radiat. Isot. 68,1435–1437.

Matishov, D.G., Matishov, G.G., 2004. Radioecology in Northern European Seas.Springer-Verlag, Berlin.

Povinec, P.P., Osvath, I., Baxter, M.S., 1996. Underwater gamma-spectrometry withHPGe and NaI(Tl) detectors. Appl. Radiat. Isot. 47, 1127–1133.

Povinec, P.P., Aggarwal, P., Aureli, A., Burnett, W.C., Kontar, E.A., Kulkarni, K.M.,Moore, W.S., Rajar, R., Taniguchi, M., Comanducci, J.-F., Cusimano, G., Dulaiova, H.,Gatto, L., Hauser, S., Levy-Palomo, I., Ozorovich, Y.R., Privitera, A.M.G., Schiavo, M.A.,2006. Characterisation of submarine ground water discharge offshore South-eastern Sicily—SGD collaboration. J. Environ. Radioact. 89, 81–101.

Povinec, P.P., Bokuniewicz, H., Burnett, W.C., Cable, J., Charette, M., Comanducci, J.-F.,Kontar, E.A., Moore, W.S., Oberdorfer, J.A., de Oliveira, J., Peterson, R., Stieglitz, T.,Taniguchi, M., 2008. Isotope tracing of submarine groundwater discharge off-shore Ubatuba, Brazil: results of the IAEA–UNESCO SGD project. J. Environ.Radioact. 99, 1596–1610.

van Put, P., Debauche, A., De Lellis, C., Adam, V., 2004. Performance level of anautonomous system of continuous monitoring of radioactivity in seawater. J.Environ. Radioact. 72, 177–186.

Rodenas, J., Martinavarro, A., Rius, V., 2000. Validation of the MCNP code for thesimulation of Ge-detector calibration. Nucl. Instrum. Methods Phys. Res., Sect. A450, 88–97.

Table 4Advantages and drawbacks of the lab and the in-situ method for gamma-radiation measurements in the aquatic environment.

Method Advantages Drawbacks

Lab – Space-localized and time-specific measurements– Identification of multiple radionuclides due to excellent resolution

– Time consuming– Unsuitable for short lived radionuclides– High MDA values

In-situ – Direct measurement in multiple aquatic environments– Deployment in great depths unreachable by divers– Surveillance of activity concentrations variation (monitoring)– Alarm capability in case of violation of specific radiation thresholds– Low cost

– Integrated activity concentration values over space and time– Spectroscopic limitation in low energy gamma-rays (o90 keV)– Pure energy resolution

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278 277

Saegusa, J., Osihi, T., Kawasaki, K., Yoshizawa, M., Yoshida, M., Sawahata, T., Honda, T.,2000. Determination of gamma-ray efficiency curves for volume samples by thecombination of Monte Carlo simulations and point source calibration. J. Nucl. Sci.Technol. 37, 1075–1081.

Saegusa, J., Kawasaki, K., Mihara, A., Mitsuo, I., Yoshida, M., 2004. Determination ofdetection efficiency curves of HPGe detectors on radioactivity measurement ofvolume samples. Appl. Radiat. Isot. 61, 1383–1390.

Salvat, F., Fernández-Varea, J.M., Sempau, J., 2006. PENELOPE-2006: a code systemfor Monte Carlo simulation of electron and photon transport. OECD NuclearEnergy Agency, Issy-les-Moulineaux, France. Available upon request at: ⟨http://www.oecd-nea.org/tools/abstract/detail/nea-1525⟩.

Schubert, M., Buerkin, W., Peña, P., Lopez, A.E., Balcázar, M., 2006. On-sitedetermination of the radon concentration in water samples: methodical back-ground and results from laboratory studies and a field-scale test. Radiat. Meas.41, 492–497.

Schubert, M., Schmidt, A., Paschke, A., Lopez, A., Balcázar, M., 2008. In situdetermination of radon in surface water bodies by means of a hydrophobicmembrane tubing. Radiat. Meas. 43, 111–120.

Tsabaris, C., Ballas, D., 2005. On line gamma-ray spectrometry at open sea. Appl.Radiat. Isot. 62, 83–89.

Tsabaris, C., Eleftheriou, G., Kapsimalis, V., Anagnostou, C., Vlastou, R., Durmishi, C.,Kedhi, M., Kalfas, C.A., 2007. Radioactivity levels of recent sediments in theButrint Lagoon and the adjacent coast of Albania. Appl. Radiat. Isot. 65,445–453.

Tsabaris, C., 2008a. Monitoring natural and artificial radioactivity enhancement inthe Aegean Sea using floating measuring systems. Appl. Radiat. Isot. 66, 83–89.

Tsabaris, C., Bagatelas, C., Dakladas, Th., Papadopoulos, C.T., Vlastou, R., Chronis, G.T.,2008b. An autonomous in-situ detection system for radioactivity measure-ments in the marine environment. Appl. Radiat. Isot. 66, 1419–1426.

Tsabaris, C., Mallios, A., Papathanassiou, E., 2008c. Instrumentation for underwaterin-situ radon analysis. Sea Technol. 49, 21–26.

Tsabaris, C., Scholten, J., Karageorgis, A.P., Commanducci, J-F., Georgopoulos, D.,Kwong, L.-L.W., Papathanassiou, E., 2010. Application of an in-situ underwater

gamma spectrometer as a marine radon progeny monitor: continuous mon-itoring of groundwater discharges into the coastal zone. Radiat. Prot. Dosimetry142, 273–281.

Tsabaris, C., Patiris, D.L., Lykousis, V., 2011. KATERINA: an in-situ spectrometer forcontinuous monitoring of radon daughters in aquatic environment. MethodsPhys. Res., Sect. A 626–627, S142–S144.

Tsabaris, C., Patiris, D.L., Karageorgis, A.P., Eleftheriou, G., Papadopoulos, V.P.,Georgopoulos, D., Papathanassiou, E., Povinec, P.P., 2012. In-situ radionuclidecharacterization of a submarine groundwater discharge site at Kalogria Bay,Stoupa, Greece. J. Environ. Radioact. 108, 50–59.

Valukovski, S.M., Nikitin, A.I., Chumichev, V.B., Katrich, I.Y., Voitsekhovich, O.A.,Medinets, V.I., Pisare, V.V., Bokvum, L.A., Klersonsky, E.S., 1994. Cesium-137 andstrontium-90 contamination of water bodies in the areas affected by releasesfrom the chernobyl nuclear power plant accident: an overview. J. Environ.Radioact. 23, 103–122.

Vargas, M.J., Guerra, A.L., 2006. Application of PENELOPE code to the efficiencycalibration of coaxial germanium detectors. Appl. Radiat. Isot. 64, 1319–1322.

Venturini, L., Vanin, V.R., 1993. HPGe detector efficiency calibration for extendedsources in the 50–1400 keV energy range. Appl. Radiat. Isot. 44, 999–1002.

Vidmar, T., Kanisch, G., Vidmar, G., 2011. Calculation of true coincidence summingcorrections for extended sources with EFFTRAN. Appl. Radiat. Isot. 69, 908–911.

Vlastou, R., Nitziou, I.T., Kokkoris, M., Papadopoulos, C.T., Tsabaris, C., 2006. MonteCarlo simulation of γ-ray spectra from natural radionuclides recorded by a NaIdetector in the marine environment. Appl. Radiat. Isot. 64, 116–123.

Vojtyla, P., 2001. Calibration of monitors used for surveillance of radioactivity ineffluent water from CERN's accelerator installations. Appl. Radiat. Isot. 55,81–88.

Wedekind, Ch., Schilling, G., Gruttmuller, M., Becker, K., 1999. Gamma-radiationmonitoring network at sea. Appl. Radiat. Isot. 50, 733–741.

X-5 Monte Carlo Team, 2003. MCNP5—a general Monte Carlo n-particle transportcode. Version 5. LA-UR-03-198, LA-CP-03-0245, Los Alamos National Labora-tory, 24. Available upon request at: ⟨https://laws.lanl.gov/vhosts/mcnp.lanl.gov/mcnp5.shtml⟩.

G. Eleftheriou et al. / Applied Radiation and Isotopes 82 (2013) 268–278278