Transcript
Radiation Measurements 42 (2007) 68–73www.elsevier.com/locate/radmeas Tritiuminwaterelectrolyticenrichmentandliquidscintillationcounting Wolfango Plastino a , ∗ , Iosif Chereji b , Stela Cuna b , Lauri Kaihola c , Pierino De Felice d ,Nicolae Lupsa b , Gabriela Balas b ,Valentin Mirel b , Petre Berdea b , Calin Baciu e a Department of Physics, University of Roma Tre, Via della Vasca Navale, 84, I-00146 Roma, Italy b National Institute of Research & Development for Isotopic and Molecular Technologies, Cluj-Napoca, Romania c PerkinElmer Life and Analytical Sciences, Wallac Oy, Turku, Finland d ENEA, National Institute for Metrology of Ionizing Radiations, Roma, Italy e Faculty of Environmental Sciences, Babes-Bolyai University, Cluj-Napoca, Romania Received 3 November 2005; received in revised form 12 April 2006; accepted 20 July 2006 Abstract A batch of electrolysis cells was developed for tritium in water samples enrichment by at least a factor 10. The cell batch is controlled bya pre-programmable electronic system that interrupts the current through any cell when the planned electrolyte volume is attained. A startingand a final distillation of water samples are carried out before and after the electrolytic enrichment. Both distillations are made to dryness inorder to avoid isotopic fractionation.A second electrolysis step allowed the tritium enrichment factor (EFT) to be doubled. The EFT was calculated by means of the deuteriumenrichment factor (EFD) that was measured by mass spectrometry. The EFT was also measured by liquid scintillation counting. The calculatedand measured EFT values were found in good agreement, especially for samples with significant tritium content.© 2006 Elsevier Ltd. All rights reserved. 1. Introduction Tritium ( 3 H) has a great significance in isotope hydrology inaddition to the stable isotopes 2 H and 18 O (Cameron and Payne,1965).Tritium is a pure beta emitter (E max = 18 . 6keV ) with therecently re-evaluated half-life of 12.32 years (Lucas and Unter-weger, 2000; Unterweger and Lucas, 2000). Its concentration isdescribedintermsoftritiumunits(TU),where1TU = 1atomof tritium per 10 18 atoms of hydrogen = 0 . 11919 ± 0 . 00021Bq / kgof water (Gröning and Rozansky, 2003).In the northern hemisphere, currently, the mean tritium con-centration in precipitations and surface water is 5–10TU. Ingroundwater systems and oceans the tritium concentration isfrequently much lower, even below the detection level of themost sensitive systems (Theodorsson, 1999).It has been established that useful hydrological and me-teorological information can be obtained by measuring thenatural tritium content of precipitations, surface water and ∗ Corresponding author. Tel.: +390655177277; fax: +39065579303. E-mail address: plastino@fis.uniroma3.it (W. Plastino).1350-4487/$-see front matter © 2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.radmeas.2006.07.010 groundwater. Tritium is one of the most important global con-taminants, produced not only by nuclear bomb tests, but alsoalways in increasing amounts by fission reactors and nuclearfuel reprocessing as well.Tritium contamination of the environment may be usedas a tracer in hydrological investigations or in groundwa-ter dating, and routine tritium analyses in water samples areusually performed by liquid scintillation spectrometry. Atpresent, a certain degree of enrichment is essential to obtainadequate net tritium count rates for most of the water sam-ples. Electrolysis is generally used for the enrichment process(Sauzay and Schell, 1972; Groeneveld, 1977; Taylor, 1977;Cook et al., 1998).In this work, tests to estimate the electrolysis methodologyhave been performed by tritium enrichment factor (EFT). Then,special attention was paid to measure the EFT by mass spec-trometry using the deuterium enrichment factor (EFD), and byliquid scintillation counting (LSC) to estimate the stability andreliability of this procedure and showing a good agreement of the EFT values detected, especially for samples with significanttritium content. W. Plastino et al. / Radiation Measurements 42 (2007) 68–73 69 2. Materials and methods In order to increase the tritium concentration in the watersamples to an easily measurable level, a batch of 5 electrolysiscells was developed.The cell construction and the material used(mild steel–stainless steel) are similar to those of the IAEAsystem (Cameron and Payne, 1965; Chereji, 1987).At the top of the anode tube (Fig. 1) a brass ring serves bothas a mounting aid and an electrical connection. The cathode ismade up of a mild steel tube. Prior to the cell assemblage, thecathodes have undergone a special chemical treatment (Zutshiand Sas-Hubicky, 1966) in order to improve the cell perfor-mance. This treatment is not required to be repeated for a longtime, as a typical black surface layer gradually develops andthe separation factor steadily increases during the first 30–40enrichment runs (Groeneveld, 1977). The gas flows through a stainless steel tube welded to thecathode that also provides the other electrical connection withthe help of a brass ring. From this tube the gas is released tothe atmosphere through a long and narrow plastic tube and asilica-gel trap which absorbs the evaporated and sprayed water.The procedure for electrolytic tritium enrichment of waterhas to exclude both the isotopic fractionation during the sam-ple treatment and the possible contamination with another wa-ter of high 3 H concentration. Before the enrichment run, the Fig. 1. Longitudinal section through the electrolytic cell. procedure includes a starting distillation of the samples andafter the enrichment, a final distillation. During distillation thetemperature should be as high as possible to attain a waterrecovery close to the unit. The distillation and the electrolysiscells reservoir should be separated from the surrounding air asit usually contains water vapor with a higher 3 H concentrationthan that of the sample. This is the reason why a long andnarrow PVC tube must be connected as a diffusion barrier bothto the condensation flask of the distillation apparatus and to theelectrolysis cells.With a view to obtaining a planned final volume automat-ically, a pre-programmable electronic system was developed.By using an electrolyte level sensor, this system can switch-off the current through any cell in which the planned volume of electrolyte is attained.Each cell was filled with about 330mL of the sample waterin which 2.5g NaOH was dissolved for the single step run.The cells were connected in series and a voltage correspond-ing to 2.2–2.7V was applied across each of them from thebattery-charger. The current was stabilized at a maximum of 10A and was reduced to a half value on the final stage of theelectrolysis run.A cooling bath with running tap water ( 4 . 8 ) ◦ C was used inorder to minimize the lost quantity of evaporated and sprayedwater.A total charge of about 977Ah is theoretically needed for thevolume reduction of about 305mL of water. When the presetleveloftheremainingelectrolyteisattainedinacell,thecurrentisautomaticallyinterruptedthroughthiscellandtheelectrolysisis continued in the remaining cells until the preset levels areattained in each of them. The disconnection should be madein such a way so as to stop a reverse current to flow throughthe cells as this has been found to damage the cathode surface(CameronandPayne,1965).Afterwardsthecellswereremovedfrom the cooling bath and the enriched samples poured out of the anode vessels into glass flasks and distilled without addinga neutralizing agent. 3. Results and discussion 3.1. Tritium enrichment Three water samples with different contents of tritium wereselected for electrolysis enrichment:(1) tap water (presumably with a very low level of tritiumcontent);(2) rain water (with a low tritium content);(3) contaminated moisture collected from the controlled areaof a nuclear laboratory, where some sources of high tritiumactivity were manipulated and stored.Two replicates,A and B, were taken from each sample. Both of them were distilled and electrolyzed separately: in the formerelectrolysis for low enrichment (l.e.), and in the latter electroly-sis for high enrichment (h.e.). In the electrolysis for h.e., in thefirst step reducing the volume was reduced from 355 to about50mL, and in the second step of the run the cells were refilled 70 W. Plastino et al. / Radiation Measurements 42 (2007) 68–73 with 305mL of sample water without adding more NaOH, inorder to keep the final electrolyte concentration within the tol-erable limits (the initial quantity of NaOH was also 2.5g as inthe electrolysis for l.e.).The EFT for the tritium content was calculated by means of the EFD for the deuterium content. The EFD isEFD = D fin D in , (1)where D in and D fin are the deuterium contents measuredby mass spectrometry before and after the electrolysis,respectively.Taking into account the relation between the enrichment fac-tor EFD and the separation factor for the deuterium isotope(Östlund and Werner, 1962),EFD = V in V fin ( − 1 )/ ,where V in is the initial volume of the sample introduced in acell, and V fin is the final volume of the sample remaining afterelectrolysis, the separation factor for the deuterium isotopecan be calculated from the following relation (Cameron, 1967): = log (V in /V fin ) [ log (V in /V fin ) − log EFD ] . (2)The separation factor for the tritium isotope can be deducedby using the relation (Bigeleisen, 1962) log = log , (3)where is a constant depending on the electrode material.Finally, the enrichment factor EFT can be evaluated with theaid of a similar relation for EFD (Östlund and Dorsey, 1977):EFT = V in V fin ( − 1 )/ . (4)The value = 1 . 4 (theoretically deduced by Bigeleisen, 1962for the electrolysis of water) can be used with the predictablelimits of ± 5% variation (Chereji, 1987).For an iron–nickel system, Östlund and Werner (1962) ex-perimentally found a value of 1.369 for , which is practicallythe same as the mean value determined by Neary (1997) using4 tritium standards.The results presented in Tables 1 and 2 were calculated with the experimental value for .As it can be seen, the enrichment factors EFT determined byusing the deuterium concentrations (measured by mass spec-trometry) are in the range of 11.0–12.2 for l.e., and 21.4–22.9for h.e., respectively, and in very good agreement with the lit-erature (Cook et al., 1998). 3.2. Liquid scintillation measurements For the tritium measurements, an ultra low-level liquid scin-tillation spectrometer Quantulus TM was used in undergroundconditions at Gran Sasso National Laboratory, where a 3800mwater equivalent shielding almost fully removes the cosmiccontribution to the background (Arpesella, 1992; Plastino et al.,2001; Plastino and Kaihola, 2004).By using glass vials having some content of 40 K, the effectof Cherenkov and fluorescence radiation on glass is impossibleto remove fully with Quantulus Pulse Shape Analyzer (Fig. 2).Some loss of 3 H efficiency is always associated with such at-tempts (Kaihola, 1991).The results presented in Table 3 were obtained by usingpolyethylene (PE) vials and a mixture of 10mL Ultima GoldLLT cocktail with 10mL of water sample (Schönhofer andKralik, 1999).Calibration of the LSC system was performed with a setof liquid scintillation sources provided by the Italian NationalInstitute for Metrology of Ionizing Radiations by using non-radioactive PE vials and a mixture of 10mL Ultima Gold LLTcocktail with 10mL tritium water (about 1kBq).The mean count rates are given for 2 × 3 repeats of the samplemeasurements. Sample stayed in counting chamber during each3 repeat sequence. Counting time for each repeat is 1h and timedifference to the later 3 repeats was 60h, i.e. 6h total countingtime for each sample.Tap water samples 1A0, 1B0, 1A1, 1B1, 1A2 and 1B2 haveno observable tritium concentration, as enrichment does notincrease the count rates. The mean count rate of these sampleswas taken as the background figure (Fig. 3). A great part of the remaining background signals belowchannel 300 rising towards lower channels in the wide tritiumwindow is due to the presence of the radon gas in the labo-ratory air (Arpesella, 1992; Plastino and Kaihola, 2004). Thenitrogen luminescence by radon alpha particles in the countingchamber produces an identifiable spectral peak in the tritiumwindow. The cocktail itself contains a little 40 K, which showsas a rising background in channels 400–750, but very little inthe tritium window.The samples were at about the same quench level consider-ing the variation of the quench parameter SQP(E) (Gupta andPolach, 1985) and therefore, all tritium activity calculations aregiven for a fixed counting efficiency.From Table 3 the measured activity ratios for the moisturesamples (measured EFT) are very close to the EFT calculatedby using the deuterium measurements while there is a slightscatter in rain water results.The best figure of merit (FOM) is obtained for optimal chan-nels 20–160 window, with Efficiency ( Eff ) = 22 . 2% and meanbackground ( Bkg ) = 1 . 066countsmin − 1 :FOM = Eff 2 / Bkg = 460, while for wide tritium windowFOM = 384.The detection sensitivity for the tritium analysis method,given by the minimum detectable activity (MDA) or concen-tration (MDC), is enhanced by the order of magnitude of theenrichment factor EFT provided by the electrolysis, i.e.MDA or MDC/EFT = 0 . 95 / 11 0 . 08Bqkg − 1 or 0 . 7TU for l.e. and = 0 . 95 / 22 = 0 . 04Bqkg − 1 or 0 . 4TU for h.e., W. Plastino et al. / Radiation Measurements 42 (2007) 68–73 71Table 1Single step electrolysis for l.e.Sample Tap water Rainwater Moisture1A 1B 2A 2B 3A 3BCell no. 1 2 3 4 5 5 V in (mL) 330 330 330 330 330 330 V fin (mL) 25.2 24.9 25.9 25.6 25.8 25.2 D in (ppm) 143.2 143.2 130.7 130.7 140.8 140.8 D fin (ppm) 1334.7 1496.3 1462.0 1393.0 1348.7 1287.2EFT 11 . 3 ± 0 . 6 12 . 2 ± 0 . 6 11 . 4 ± 0 . 6 11 . 0 ± 0 . 6 11 . 2 ± 0 . 6 11 . 0 ± 0 . 6The same cell (e.g. no. 5) was used consecutively to electrolyse the sample 3A and its duplicate 3B.Table 2Two steps electrolysis for h.e.Sample Tap water Rain water Moisture1A 1B 2A 2B 3A 3BCell no. 1 2 3 4 5 5First step 1 V in (mL) 355 355 355 355 355 355 V fin (mL) 50 50 50 50 50 50Second step 2 V in (mL) 305 305 305 305 305 305 V fin (mL) 27.3 25.8 26.1 25.8 26.0 25.9 D in (ppm) 143.7 143.7 130.7 130.7 140.8 140.8 D fin (ppm) 2579.2 2633.1 2570.8 2443.0 2549.7 2525.3EFT 21 . 4 ± 1 . 1 22 . 2 ± 1 . 1 22 . 9 ± 1 . 1 22 . 4 ± 1 . 1 22 . 0 ± 1 . 1 21 . 8 ± 1 . 1The total initial volume used in the electrolysis is given by 1 V in + 2 V in = 355 + 305 = 660ml.Fig. 2. 3 H spectrum is the STD sample spectrum in a glass vial of 10:10 mixture of water and Ultima Gold LLT. Glass vial background spectrum is thecomposite spectrum of samples 1A0, 1B0, 1A1, 1B1, 1A2 and 1B2 from a similar mixture. Sample scales are normalized to maximum vertical displacement. where MDA = 0 . 95Bqkg − 1 is calculated for 10mL volume of water (0.01kg), in optimal window and with a counting timeof 24h, usually used (Neary, 1997). 4. Conclusions The usual improvement of the detection sensitivity by anorder (or more) of magnitude for the tritium content of watersamples can be performed by electrolysis.As the developed batch of cells was provided with a pre-programmable electronic system having an electrolyte levelsensor, the pre-set termination condition in the electrolysis canbe set in such a way as to ensure in each cell a desired fi-nal quantity of water, usually 20–30% higher than that usedfor LSC.Three water samples and their duplicates (a tap water,practically with unobservable content of tritium, a rain waterwith a low level of tritium concentration and a contaminated 72 W. Plastino et al. / Radiation Measurements 42 (2007) 68–73 Table 3Results of the measurements in wide tritium window (Channels 5–200)Sample in PE vials Count rate ( min − 1 ) Reference to mean BKG (Bq) Error reference to mean BKG (Bq) Measured EFT Calculated EFTSTD 3006 210.5001A0 1.418 − 0.004 0.0011B0 1.511 0.003 0.0011A1 1.573 0.007 0.001 − 1.8 11.21B1 1.404 − 0.005 0.001 − 1.8 12.01A2 1.418 − 0.004 0.001 1.0 21.41B2 1.517 0.003 0.001 1.2 22.22A0 4.651 0.223 0.0152B0 5.731 0.298 0.0172A1 37.79 2.545 0.041 11.4 11.42B1 32.74 2.190 0.038 7.3 11.02A2 55.48 3.783 0.051 17.0 22.92B2 78.58 5.402 0.064 18.1 22.43A0 42.57 2.880 0.0443B0 42.66 2.885 0.0443A1 458.1 31.98 0.24 11.1 11.23B1 453.9 31.70 0.24 11.0 11.03A2 896.8 62.73 0.43 21.8 22.03B2 878.5 61.45 0.42 21.3 21.8STD 3 H activity = 12 . 630decaysmin − 1 = 210 . 500 ± 1 . 100BqMean BKG (mean count rate of 1A1–1B2) = 1 . 473 ± 0 . 026min − 1 Eff = 23 . 8%Sample notations: 1A0, 1B0 = tap water before e.; 2A0, 2B0 = rain water before e.; 3A0, 3B0 = moisture before e.; 1A1, 1B1 = tap water after l.e.; 2A1,2B1 = rain water after l.e.; 3A1, 3B1 = moisture after l.e.; 1A2, 1B2 = tap water after h.e.; 2A2, 2B2 = rain water after h.e.; 3A2, 3B2 = moisture after h.e.;STD = standard sample with a known activity.Fig. 3. Composite background spectrum of the samples 1A0, 1B0, 1A1, 1B1, 1A2 and 1B2. moisture, collected from a nuclear laboratory) were subjectedto two enrichment processes:(1) an electrolysis for l.e. (with EFT 11);(2) an electrolysis for h.e. (with EFT 22).By comparing the EFT calculated with the aid of the deu-terium concentrations (measured by mass spectrometry be-fore and after the enrichment) with the ratios of tritiumcount rates determined by LSC there is an excellent agree-ment in the case of samples with a significant content of tritium only.The spectra accumulated with the samples in PE vials haverevealed some radon and 40 K contribution to the backgroundin the tritium window. Without these contributions the back-ground would be at least 50% lower than the mean backgrounddetected.The underground laboratory of Gran Sasso provides an ex-cellent environment for low background measurements, takinginto consideration that the cosmic contribution is almost en-tirely missing. Nevertheless, it seems that only a strict reductionand control of radon in the surrounding air will ensure the bestconditions which presume a residual background merely fromthe inherent radioactivity of the phototubes and of the bedrock. Acknowledgments We wish to thank an anonymous referee for constructivereviews and Prof. Eugenio Coccia, Director of Gran Sasso Na-tional Laboratory, for his kind collaboration.
