Working Report, 11/2006 – Radioactive isotopes in Danish drinking water

Radioactive isotopes in Danish drinking water

6 Analytical Methods

6.1 Total Alpha and Beta Radioactivity
6.2 Uranium
6.3 Radium
6.4 Radon

6.1 Total Alpha and Beta Radioactivity

For water samples collected during 2001, aliquots of 300-500 ml water were evaporated completely in 60 mm diameter aluminium trays, which were subsequently measured in a multi sample proportional gas-flow counter for 1000 minutes. This proportional counter distinguishes events originating from alpha and beta decays by analysing the pulse height. Self-absorption of the alpha particles in the dry solids of the evaporated salt was corrected for based on measurements using a standardized uranium solution. No correction for absorption of beta particles was considered. Efficiency calibration was made using natural uranium and ⁹⁰Sr (⁹⁰Y) standards of known activity. Detection limits were calculated as three times the standard deviation of blank values and gave values of 0.01 Bq/l for total alpha and 0.03 Bq/l for total beta radioactivity.

For water samples collected during 2002 and 2003, aliquots of 200 ml water was transferred to a glass beaker and acidified with HCl to about pH 1. The water was evaporated until 1-5 ml was remaining depending on the salt content. The samples were never allowed to run dry. The remaining sample was transferred with a pipette to a liquid scintillation cell and the beaker washed twice with weak (about 1%) HCl, which was added to the sample. Once transferred to the liquid scintillation cell, 10 ml of 'Ultima Gold LLT'-scintillator cocktail was added and mixed thoroughly with the sample. The samples were counted for 150-200 minutes using a Wallac Quantulus 1220 liquid scintillation spectrometer. This spectrometer distinguishes events originating from alpha and beta decays by analysing the pulse shape. By adjusting the pulse shape analyser (PSA), alpha and beta activities are presented separately for the same sample. Blank and background samples were made by evaporating 200 ml distilled water at the same time as the samples. Counting of the blanks was done together with the samples. Efficiency calibration was made using ^239+240Pu and ⁹⁰Sr (⁹⁰Y) standards of known activity. Detection limits were 0.01-0.03 Bq/l for total alpha and 0.03 Bq/l for total beta radioactivity. The detection limits were determined from blank samples (distilled water) evaporated simultaneously with the water samples and calculated as three times the standard deviation of blank values.

Different analytical equipment was used to determine total alpha and beta radioactivity during the project for what reason a comparison was made between results obtained from the gas-flow proportional counter and the liquid scintillation counter (LSC). Even though not ideal, samples previously measured on the gas-flow counter were dissolved and measured on the LSC. Furthermore, since water still remained from one of the stations (Feldbak) this was re-analysed and results compared with previous data.

The results of the test are presented in the following table.

Sample	Sample ID	Alpha (Bq/l)	Beta (Bq/l)
Pedersker 2001, proportional counter	1099	<0.01	0.58
Pedersker 2001, re-dissolved salt (LSC)	1099	<0.01	0.45
Feldbak 2001, proportional counter	1102	<0.01	0.83
Feldbak 2001, re-dissolved salt (LSC)	1102	0.03	0.48
Feldbak 2001, new evaporation (LSC)	1102	0.06	0.65

The variation between the results is acceptable considering the different methods used. For the Pedersker sample, the total beta results from the proportional counter and the LSC show a variation around the mean value corresponding to an 18% standard deviation. A significant difference was found between LSC and gas-flow measurements for the Feldbak total beta. This may partly be due to the delay in time between the two measurements. The two LSC measurements performed (re-dissolved salts and new evaporation) were done comparatively close in time and therefore may show less difference. One reason for the relatively small difference between the two techniques may be that the correction for self-absorption on the evaporated samples measured on the gas-flow counter was calibrated using natural uranium. Since uranium was the major contributor in samples from several water works this correction was probably unusually correct. For LSC no severe corrections for self absorption (apart from quenching) were necessary.

6.2 Uranium

For samples collected during 2001 and 2002, uranium was determined by alpha spectrometry on large (25-50 litres) samples following radiochemical separation. The ²³²U tracer used was calibrated gravimetrically against a uranyl sulphate salt.

For samples collected during 2003, uranium was analysed directly on selected samples using isotope dilution mass spectrometry (PlasmaTrace 2 HR-ICP-MS). As yield determinant ²³³U was used. Due to non-significant amounts of ²³⁴U present in the ²³³U tracer, two separate measurements were done for each sample, one with ²³³U tracer in order to obtain concentrations of ²³⁸U and one without tracer in order to obtain the ²³⁴U/²³⁸U ratio. The ²³³U tracer was calibrated against a Merck multi-element standard as well as an Aldrich uranium standard solution. The detection limit for the procedure used for ²³⁸U is lower than 1 nBq/l. The uncertainty given only considers the counting statistics. The detection limit was set as three times the standard deviation (based on counting statistics) of the blank value measured in 1% distilled nitric acid containing the same amount of yield determinant (²³³U) as the samples.

6.3 Radium

For drinking water samples collected during 2001, radium was determined on 5-l water samples by direct radon emanation into Lucas scintillation cells after allowing storage of the water in gas-tight bottles for three weeks to reach radioactive equilibrium between ²²²Rn and ²²⁶Ra. For samples collected during 2002 and 2003, radium was determined by liquid scintillation counting. Radium was co-precipitated onto MnO₂ from 2-10 l water using ¹³³Ba as tracer. The MnO₂ was dissolved and Ra(Ba) co-precipitated onto PbSO₄ which in turn was dissolved in 6-8 ml alkaline EDTA solution, transferred to a liquid scintillation cell and a mineral-based scintillator (Opti Fluor O, Packard BioScience) added in order to collect the radon. After a radon ingrowth period of three weeks, the samples were counted on a Wallac Quantulus 1220 liquid scintillation spectrometer. The detection limit for the procedure used for analysis of radium is lower than 1 mBq/l based on analysis of blank samples containing only the chemicals used for the procedure. Uncertainty is based on counting statistics only.

6.4 Radon

Radon concentrations were determined using a liquid scintillation counter and methods described previously (Sundhedsstyrelsen, 1986). The detection limit is approximately 1 Bq/l and analytical uncertainties are estimated to be less than 10%. Radon concentrations were corrected for decay for the time from sampling to measurement using the ²²²Rn half-life of 3.8 days.