ARTICLE IN PRESS. Applied Radiation and Isotopes



Similar documents
Antoine Henri Becquerel was born in Paris on December 15, 1852

COMPENDIUM OF EPA-APPROVED ANALYTICAL METHODS FOR MEASURING RADIONUCLIDES IN DRINKING WATER

Analytical Grade Ion Exchange Resins

Comparison of natural radioactivity removal methods for drinking water supplies: A review

Proficiency testing schemes on determination of radioactivity in food and environmental samples organized by the NAEA, Poland

Experiment 10. Radioactive Decay of 220 Rn and 232 Th Physics 2150 Experiment No. 10 University of Colorado

The Radon Calibration Laboratory at the Federal Office for Radiation Protection (BfS)

Perspectives of a method for measuring soil-gas radon by an opened counting vial

WM2012 Conference, February 26 March 1, 2012, Phoenix, Arizona, USA

SHORT-TERM STABILITY TEST FOR URANIUM SOIL CANDIDATE A REFERENCE MATERIAL

Radioanalytical Data Validation Issues and Solutions

Waste Management 04 Conference, February 29 - March 4, 2004, Tucson, AZ Copyright WM Symposia, Inc. All Rights Reserved. Reprinted with permission.

COUNCIL DIRECTIVE 2013/51/EURATOM

Azeri, Chirag & Gunashli Full Field Development Phase 3 Environmental & Socio-economic Impact Assessment. A10.1 Introduction...

EPA Radionuclides Rule and the RadNet Program

A Comparison of an HPGe-based and NaI-based Radionuclide Identifier (RID) for Radioactive Materials

P O L Y T E C R.M. S.R.L.

RESULTS OBTAINED IN MEASUREMENTS OF Rn-222 WITH THE ROMANIAN STANDARD SYSTEM*

General and statistical principles for certification of RM ISO Guide 35 and Guide 34

Automation and Methodology Development for Environmental and Biological Determination of Pu, Np, U and Tc

RADON Although radon is agas, its decay products are not, and they occur either as unattached

An Innovative Method for Dead Time Correction in Nuclear Spectroscopy

Dissolved Mineral Radioactivity in Drinking Water

4. CHEMICAL, PHYSICAL, AND RADIOLOGICAL INFORMATION

Methods verification. Transfer of validated methods into laboratories working routine. Dr. Manuela Schulze 1

INFO-0545 RADIOISOTOPE SAFETY MONITORING FOR RADIOACTIVE CONTAMINATION

CRIIRAD report N Analyses of atmospheric radon 222 / canisters exposed by Greenpeace in Niger (Arlit/Akokan sector)

Performance Standard for Organisations Undertaking Radioanalytical Testing of Environmental and Waste Waters. Environment Agency July 2015 Version 2

Radium-223 Therapy: Handling & Radiation Safety Issues

Radium-226, dissolved Radon emanation method (R ) Parameter and code: Radium-226, dissolved (pci/l): 09511

UNITED STATES CONSUMER PRODUCT SAFETY COMMISSION DIRECTORATE FOR LABORATORY SCIENCES DIVISION OF CHEMISTRY 5 RESEARCH PLACE ROCKVILLE, MD 20850

RADON AND HEALTH. INFORMATION SHEET October What is radon and where does it come from?

Ionizing Radiation, Czech Republic, CMI (Czech Metrology Institute)

Certificate of Analysis

Radiation Protection s Benefits in the Production of Iodine-131

1.3 Radioactivity and the age of the solar system

The objective of this chapter is to provide criteria with which to assess the safety of

Eckert & Ziegler Reference & Calibration Sources

EUROPEAN COMMISSION DIRECTORATE-GENERAL TAXATION AND CUSTOMS UNION TAX POLICY Excise duties and transport, environment and energy taxes

Performance and advantages of qnmr measurements

AN INVESTIGATION INTO THE USEFULNESS OF THE ISOCS MATHEMATICAL EFFICIENCY CALIBRATION FOR LARGE RECTANGULAR 3 x5 x16 NAI DETECTORS

HS 1003 Part 2 HS 1003 Heavy Metals Test

Purification of reaction mixtures using flash chromatography.

Individual Monitoring Services (IMS) for External and Internal Exposures

Reference Materials for Environmental Performance Testing Dr Steve Wood Head of Regulatory and Legislative Services. ISPRA 25 June 2009

APPENDIX N. Data Validation Using Data Descriptors

DEPARTMENT OF ENVIRONMENTAL REGULATION. Technical Document DETERMINING REPRESENTATIVE GROUND WATER SAMPLES, FILTERED OR UNFILTERED

Radiation Protection from NORM and TENORM in the Oil and Gas Industry: Regulatory and Non-Regulatory Approaches

Modern approaches to determination of toxic metals in marine environmental objects. Atomic absorption and inductively coupled plasma, advantages and

Radiological Protection Principles concerning the Natural Radioactivity of Building Materials

[2] At the time of purchase of a Strontium-90 source, the activity is Bq.

RN-222 RELEASE TO THE ENVIRONMENT: COMPARISON BETWEEN DIFFERENT GRANITE SOURCES

Dissolved and precipitated oxalate

For convenience, we may consider an atom in two parts: the nucleus and the electrons.

RADIATION MONITORING NETWORK OF THE CZECH REPUBLIC

Lead Testing and On Site Calibration for Water Testing Detection Range: 2 100ppb

OLIVÉR BÁNHIDI 1. Introduction

Validation and Calibration. Definitions and Terminology

A Review of Emerging Gamma Detector Technologies for Airborne. Radiation Monitoring

Natural radioactivity in waste from former non-uranium mining areas in Norway

Questionnaire for NORM service providers

Start of introducing ICP MS method for analyzing water samples for Strontium. November 27, 2014 Tokyo Electric Power Company

4.2 Bias, Standards and Standardization

Nuclear Physics. Nuclear Physics comprises the study of:

Tutorial 4 SOLUTION STOICHIOMETRY. Solution stoichiometry calculations involve chemical reactions taking place in solution.

Environmental Radiation Monitoring in Taiwan

2016 No WATER, ENGLAND AND WALES. Water Supply (Water Quality) Regulations 2016

A Digital Spectrometer Approach to Obtaining Multiple Time-Resolved Gamma-Ray. Spectra for Pulsed Spectroscopy. W. K. Warburton a

Basics of Nuclear Physics and Fission

Automated Laboratory Quality Assurance Program: Using the ORTEC GammaVision -32 Software

Potassium-Argon (K-Ar) Dating

Ecological Aspects of Oil Shale Processing

How To Measure Radionuclide Composition

Unit 1 Practice Test. Matching

Solubility Product Constant

Treatment Centers for Radioactive Waste

Assessment of environmental radiation monitoring data in Hungary following the Fukushima accident

Which analytical tools/instruments/techniques are used for POPs. A general walk through the analysis

Guide to Method Validation for Quantitative Analysis in Chemical Testing Laboratories

The International Monitoring System noble gas network Matthias Auer, CTBTO, Vienna

MOLES AND MOLE CALCULATIONS

105 Adopted:

USE OF REFERENCE MATERIALS IN THE LABORATORY

Chapter 1 The Atomic Nature of Matter

CHEMISTRY STANDARDS BASED RUBRIC ATOMIC STRUCTURE AND BONDING

Proficiency Testing In A Global Analytical Environment Benefits & Challenges. AOAC September 2007

Elemental Analyses by ICP-AES

Chemistry 1000 Lecture 2: Nuclear reactions and radiation. Marc R. Roussel

CONCENTRATION OF U-238, U-235, RA-226, TH-232 AND K- 40 FOR SOME GRANITE SAMPLES IN EASTERN DESERT OF EGYPT.

Study of the selective determination of 90 Sr in environmental. and nuclear power plant samples

Chemistry 321, Experiment 8: Quantitation of caffeine from a beverage using gas chromatography

Transcription:

Applied Radiation and Isotopes 68 (1) 1 16 Contents lists available at ScienceDirect Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso EC comparison on the determination of 226 Ra, 2 Ra, 2 U and 8 U in water among European monitoring laboratories U. Wätjen a,, L. Benedik a,1, Y. Spasova a, M. Vasile a, T. Altzitzoglou a, M. Beyermann b a European Commission, Joint Research Centre, Institute for Reference Materials and Measurements (IRMM), Retieseweg 111, B-24 Geel, Belgium b Bundesamt für Strahlenschutz (BfS), Köpenicker Allee 1-, D-1318 Berlin, Germany abstract In anticipation of new European requirements for monitoring radioactivity concentration in drinking water, IRMM organized an interlaboratory comparison on the determination of low levels of activity concentrations (about 1 1 mbq L 1 ) of the naturally occurring radionuclides 226 Ra, 2 Ra, 2 U and 8 U in three commercially available mineral waters. Using two or three different methods with traceability to the International System of Reference (SIR), the reference values of the water samples were determined prior to the proficiency test within combined standard uncertainties of the order of 3% 1%. An overview of radiochemical separation and measurement methods used by the 45 participating laboratories are given. The results of the participants are evaluated versus the reference values. Several of the participants results deviate by more than a factor of two from the reference values, in particular for the radium isotopes. Such erroneous analysis results may lead to a crucial omission of remedial actions on drinking water supplies or to economic loss by an unjustified action. & 9 Elsevier Ltd. All rights reserved. 1. Introduction The activity concentration of natural radionuclides in drinking water is quite variable, depending on the properties of the aquifer rock and the prevailing lithology. Whereas the radionuclide activity concentration in waters usually poses no health concern, there are regions in which the geological situation can render radioactivity levels which need to be monitored in order to reduce the potential health risk of the public. Concerns about the radioactivity concentration of water intended for human consumption have been described (Benedik et al., 8, 9), and a variety of regulations implemented (WHO, 1993, 6; European Communities, 1998, 1) to limit public exposure to radioactivity from drinking water. The European Atomic Energy Community (EURATOM) Treaty obliges the member states of the European Union to monitor and report the levels of environmental radioactivity on their territory. Some details of sample taking and measurement requirements (sample types, sampling intervals, radionuclides, etc.) are regulated at European level. In order to verify the quality and in particular comparability of the values reported by the member states, comparison exercises were introduced by the European Corresponding author. Tel.: + 14 571 882; fax: + 14 584 273. E-mail address: uwe.waetjen@ec.europa.eu (U. Wätjen). 1 Present address: Jožef Stefan Institute, Jamova 39, SI-11 Ljubljana, Slovenia. Commission after the reactor accident of Chernobyl. Since 3, the Institute for Reference Materials and Measurements (IRMM) had responsibility for their organization. The metrological approach of IRMM in conducting such comparisons using samples carrying reference values traceable to SI units and the International Reference System for g-ray emitting radionuclides (SIR) was presented at the most recent ICRM conference (Wätjen et al., 8). In anticipation of new requirements (European Communities, 1) for monitoring radioactivity in drinking water, IRMM has organized a comparison among member state laboratories requesting the determination of low levels of activity concentrations (around the detection limits prescribed in future legislation) of the naturally occurring radionuclides 226 Ra, 2 Ra, 2 U and 8 U in three commercially available mineral waters (W1, W2 and W3, the latter as two subbatches W3-5 and -51). 2. Determination of the reference values The reference values for these comparison samples (activity concentration) were determined by consensus between two independent expert laboratories according to clause 5.5 of the standard ISO 1 (ISO, 5) with the additional constraint that the analyses had to be performed in such a way that the determined values are traceable to SI units and the SIR (Wätjen et al., 8). IRMM and the Bundesamt für Strahlenschutz (BfS), Department for Radioprotection and the Environment, Berlin, as 969-843/$ - see front matter & 9 Elsevier Ltd. All rights reserved. doi:1.116/j.apradiso.9.11.7

U. Wätjen et al. / Applied Radiation and Isotopes 68 (1) 1 16 11 expert laboratories applied two largely different, independent methods (three in the case of 2 Ra). Since the methods used to determine the reference values in these waters have only partly been published earlier ((Benedik et al., 8) for 226 Ra and uranium at IRMM; (Vasile et al., 9) for two methods of 2 Ra at IRMM), some complementary information is summarised here. A detailed account of the reference value determination is given in the comparison report (Spasova et al., 1). 2.1. Determination of 2 U and 8 U At BfS, the determination of uranium isotopes was performed with proven methods of radiochemical separation and activity measurement (Bundesminister, 6). Whereas the radiochemical separation by extraction chromatography and the use of 2 U as a tracer is very similar to the method used at IRMM, the preconcentration consisted of evaporating water samples of 1 L to dryness and wet ashing of the residues. The thin sources for a- particle spectrometric measurement were prepared by electrodeposition on stainless steel discs (Spasova et al., 9). 2.2. Determination of 226 Ra At BfS, the standardized radon emanation method H-Ra-226- TWASS-1-1 (Bundesminister, 6) was applied. Radium was preconcentrated by coprecipitation with BaSO 4 from water samples of 1 L, using the natural Ba carrier addition as a tracer by determination with atomic absorption spectroscopy. The precipitate was dissolved in EDTA/NH 4 OH and transferred to a radon bubbler. The sample was de-emanated and stored for a minimum of 14 days to allow 222 Rn to grow in. The grown-in 222 Rn was then transferred to a Lucas-type scintillation chamber, where the a-emission of 222 Rn and its short-lived decay products were counted. The system was calibrated with a 226 Ra standard solution. 2.3. Determination of 2 Ra At IRMM, two independent methods were applied. The first one is based on the procedure described by Nour et al. (4), using coprecipitation with MnO 2 for preconcentration of 2 Ra from water volumes of 1.5 and 3 L, in combination with liquidscintillation counting of the daughter nuclide 2 Ac. For the second method, use was made of sources prepared a year earlier for the determination of 226 Ra by a-particle spectrometry. The basic idea is to separate and determine by a-particle spectrometry the 2 Th grown in as granddaughter nuclide of 2 Ra after more than a year. 229 Th is used as tracer for determination of the chemical yield. Both procedures are described in detail in Vasile et al. (9). The BfS laboratory followed the standard procedure H-Ra-2- TWASS-1-1 (Bundesminister, 6). Adapted from Burnett et al. (1995), radium and actinium were preconcentrated by coprecipitation with barium sulfate, the sulfate was converted to a more soluble carbonate, and after a waiting time of at least 3 h the grown-in 2 Ac was separated by extraction chromatography on RE Spec resin. Sources were prepared using micro-coprecipitation of actinium on cerium fluoride, immediately followed by b-particle measurement in a low-level proportional counter. The chemical yield of the radium separation was determined via atomic absorption spectrometry of the barium carrier. The proportional counter was calibrated with a standardised 89 Sr source. 2.4. Homogeneity and stability Since the reference values of the comparison material must be valid for each bottle of a mineral water batch, the homogeneity of each batch, i.e., the variation, s bb, of the nuclide activity concentration between bottles of the batch, are determined by random test. Any heterogeneity found contributes to the combined uncertainty of the corresponding reference value. A similar study is performed for the short term stability, u sts, of the samples. In the case of bottled mineral waters, adsorption of radionuclides to the container walls may limit its stability (i.e. the availability of radionuclides for analysis). Thus, the expanded uncertainty, U ref (k=2), of the reference value can be estimated as qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi U ref ¼ k u 2 char þs2 bb þu2 sts; ð1þ where s bb and u sts are defined above, and u char is the combined standard uncertainty of the mean of measurement results from the two laboratories such that: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi P n i ¼ 1 u char ¼ ðu c;iþ 2 ; ð2þ n where u c,i is the combined standard uncertainty of the laboratory s (or method s) result (Pauwels et al., 1998). In numerical terms, no wall adsorption could be found in the corresponding tests, hence u sts can be set to (Spasova et al., 9). In the homogeneity studies performed, values between 2% and 1% were found for the different waters and radionuclides (Benedik et al., 8; Spasova et al., 9). Table 1 gives all reference values with their expanded uncertainties, together with the separate laboratory means of activity concentration. 3. Methods used by the participating laboratories Participants were free to use separation procedures and measurement methods of their own choice. Of the 45 laboratories which reported results, not all determined all four radionuclides. Table 2 gives an overview of the measurement techniques used. Together with other details of this comparison, more complete information on the applied methods can be found in the comparison report (Spasova et al., 1). 3.1. 2 U and 8 U Thirtytwo laboratories used a-particle spectrometry for the determination of 2 U and, except in one case, of 8 U. All laboratories added isotopic tracer ( 2 U or, in one case, 6 U) in order to determine the radiochemical yield of the procedure. Except for one, all laboratories preconcentrated uranium from the water samples and carried out radiochemical separation from the matrix components. For preconcentration from water, nine laboratories applied evaporation and laboratories carried out co-precipitation, using phosphates (6 labs), Fe(OH) 3 (9 labs), and MnO 2 (2 labs), respectively. The radiochemical purification was performed with ion exchange chromatography, extraction chromatography and solvent extraction, respectively. Some laboratories used a combination of these techniques. Sources for a-particle spectroscopic measurement were prepared by electrodeposition ( labs), micro-coprecipitation with rare earth fluorides (9 labs), and micro-coprecipitation with cadmium chloride (1 lab), respectively. The four laboratories using g-ray spectrometry for 8 U evaporated the water and measured the dry residue directly. The activity concentration of 8 U was determined via 2 Th (at 63

U. Wätjen et al. / Applied Radiation and Isotopes 68 (1) 1 16 Table 1 Activity concentrations in mineral waters determined by the two laboratories, IRMM and BfS, and comparison reference values A ref calculated as mean of these results (reference date 1 May 6). Water Nuclide mean value with combined standard uncertainty u c Reference value A ref with expanded uncertainty U ref Activity concentration (mbq L 1 ) Activity concentration (mbq L 1 ) IRMM a IRMM second method b BfS W1 U-8 1.37.2..4 11.271.7 U-2 14..3 15.77.5 15.72.2 Ra-226 9876 97711 98715 Ra-2.72.3 29.1.6 75 76 W2 U-8.9.2.857.6 (.887.29) U-2 3.9.9 4.7.2 (4.7.5) Ra-226.72.6 45 43711 Ra-2 6677 69.3.7 877 7 W3-5 U-8 22..5 21.27.5 21.671.9 U-2 44.771..37.9 43.573. Ra-226 3.7.2 4.37.5 (3.67.8) Ra-2 o 9 6.57.6 6.971.4 (6.771.2) W3-51 U-8.57.8.571.6 U-2.971.2.972.4 Ra-226 3.7.2 4.37.5 (3.67.8) Ra-2 o9 6.57.6 6.971.4 (6.771.2) Values A ref given in brackets were not required to be determined by comparison participants. Uncertainties of laboratory results given as combined standard uncertainty u c, uncertainty of reference values as expanded uncertainty U ref =u k with a coverage factor k=2, corresponding to a level of confidence of about 95%. a 2 Ra via liquid-scintillation counting of 2 Ac. b via 2 Th ingrowth and a-spectrometry. Table 2 Number of participating laboratories reporting results and measurement techniques used for determining the four radionuclides. Radionuclide U-2 U-8 U tot Ra-226 Ra-2 Reported results all methods 2 Reported results by method: a-particle spectr. 31 a 3 a 11 1 (via 224 Ra) Gross-a counting 3 ICP-MS 1 1 g-ray spectr. 4 6 24 Radiochemical NAA 1 LSC 1 b 2 b 3 Fluorimetry 2 2 2 Gross-b counting 4 Sorption emanation technique 9 a Plus one additional laboratory (Water-1 only). b Plus one additional laboratory (Water-3 only). and 93 kev), assuming equilibrium between 8 U and its progeny 2 Th. One laboratory determined 2 U bya-particle spectrometry and 8 U by radiochemical neutron activation analysis via its induced 9 U nuclide. One laboratory determined the uranium isotopes with inductively coupled plasma mass spectrometry (ICP-MS) from sample aliquots of 1 ml without additional sample preparation. Three laboratories used liquid-scintillation counting (LSC) in a/b-discrimination mode. Two laboratories determined total uranium using fluorimetry after evaporating the water samples and fusing the residues with a mixture of NaF and NaCO 3 at 9 1C; the activity concentration of 8 U was calculated assuming the average natural isotopic composition and that of 2 U based on the assumption of equilibrium. 3.2. 226 Ra a-particle spectrometry was applied in eleven laboratories, all of which included the use of tracers. Laboratories used a large variety of coprecipitation techniques for preconcentration with MnO 2, Pb(NO 3 ) 2, BaSO 4, Fe(OH) 3, and Pb(CrO 4 ) 4, respectively. One laboratory determined 226 Ra directly from a Ra-adsorbing disc (MnO 2 ) which had been immersed into the sample for 6 h. Sources for a-spectrometric measurement were prepared by coprecipitation with BaSO 4 or BaCl 2, or by electrodeposition on stainless steel discs. The recoveries of the reported radiochemical procedures were determined with 3 Ba, 2 Ra (via its progeny 2 At), and 224 Ra, respectively. Six laboratories applied g-ray spectrometry; five evaporated the water samples (between 3 and L) to dryness and one laboratory preconcentrated radium by coprecipitation with BaSO 4. The residues after evaporation or the Ba(Ra)SO 4 precipitates were sealed, and 226 Ra was determined by the 1 and 69 kev g-ray lines of its 214 Pb and 214 Bi progenies after an ingrowth period of at least 3 weeks. Nine laboratories applied the sorption emanation technique similar to the procedure described above for BfS (see Section 2.2). Twelve laboratories measured samples, after different preconcentration procedures had been applied, by liquid-scintillation counting (LSC). Six laboratories coprecipitated radium with MnO 2, or as Ba(Ra)SO 4. The precipitate was purified with a variety of techniques to remove interfering radionuclides, dissolved and mixed with scintillation cocktail. In five of the six laboratories, the scintillation vial was tightly closed and stored for up to 1 month to allow ingrowth of 222 Rn and its daughters. One of these 6 laboratories measured radium immediately after sample preparation. Two laboratories of the remaining 6 freeze dried the original samples to preconcentrate, whereas two other laboratories thermally preconcentrated the samples, desorbing at the same time all dissolved radon, followed by the usual waiting time in scintillation vials. One laboratory mixed the sample with a

U. Wätjen et al. / Applied Radiation and Isotopes 68 (1) 1 16 Preconcentration Calcination (1 lab) Evaporation (~ 15 labs) - to dryness - to reduced volume Separation/purification Empore Radium Rad disk (2 labs) (1 lab) (2 labs) Extraction chromatography (3 labs) Coprecipitation with (2 labs) - BaSO 4 (9 labs) - lead sulphate (1) - Ca phosphomolybdate (1) - MnO 2 (1) Solvent extraction (TBP) (1 lab) Ion exchange resin (2 labs) direct measurement of water (1 lab) source preparation measurement (1 lab) Fig. 1. Summary of preconcentration and radiochemical separation procedures used for the determination of 2 Ra by the comparison participants. 18 16 1 1 1 8 6 8 226 Ra - W1 27 16 2 26 46 18 15 29 19 49 11 31 43 2 9 37 538 7 45 24 --> (76) Fig. 2. Mean laboratory results for 226 Ra compared to reference value of activity concentration in Water-1. Error bars, if applicable, indicate the standard deviation s of the laboratory mean, solid and dashed lines are reference value A ref 7U ref (k=2). Encoded laboratory numbers are indicated. Values in brackets indicate measurement results falling outside the scale of the figure. mineral oil scintillator and measured directly by LSC. Another laboratory used the Empore Radium Rad disc and.1 M EDTA, which selectively extracts radium from an acidic solution. Three laboratories preconcentrated radium with lead sulphate and/or BaSO 4, and the determination of 226 Ra was carried out by measuring the gross-a activity of the Ba(Ra)SO 4 precipitate on a filter using a low-background gas proportional counter. 3.3. 2 Ra About half of the laboratories evaporated the samples, to reduce volume or to dryness, as a preconcentration step. A bit fewer than half performed coprecipitation, with Ba(Ra)SO 4 as the most often formed precipitate. Some of the laboratories applied a combination of preconcentration procedures or added purification steps; Fig. 1 gives an overview of all methods used for 2 Ra. Not all laboratories determined the recovery of their concentration and purification procedures. Some used 3 Ba as a tracer, or stable Ba (in the form of added carrier) that was then determined by photometry or ICP-MS. The laboratory which determined 2 Ra via a-particle spectrometry of 224 Ra used 2 Ra in equilibrium with 229 Th as a tracer. A few laboratories used external standards ( 226 Ra, 2 Ra) or gravimetry to determine chemical yield. 4. Comparison results As far as the determination of radium is concerned, the results of the comparison are far from satisfactory. Fig. 2 shows that the values determined by many laboratories for 226 Ra in Water-1 were too low (up to a factor of 14), while two laboratories determined values too high (as much as up to 8 times the reference value). By contrast, the reference value determined by IRMM and BfS has a relative expanded uncertainty, U ref, of 15%. One should note that a 226 Ra activity concentration of 1 mbq L 1 in water is more than twice the limit of detection

14 U. Wätjen et al. / Applied Radiation and Isotopes 68 (1) 1 16 required by the EC draft directive, and measurement results five times too high or too low would result in unnecessary or omitted remedial action, respectively (cf. Table 3). It is not satisfactory that only 29 laboratories (out of ) are able to determine the 226 Ra activity concentration within % from the reference value. The results for 226 Ra in Water-2, although at lower activity concentration, reveal a similar, although slightly better performance (Fig. 3). Again, laboratories show results spread around Table 3 Reference concentrations a and required limits of detection proposed in the draft European Council Directive (European Communities, 1). Nuclide Reference concentration (Bq L 1 ) Detection limit (Bq L 1 ) U-8 3..2 U-2 2.8.2 Ra-226.5.4 Ra-2.2.8 a Reference concentrations refers to activity concentrations triggering under certain conditions and after taking account of summing effects remedial actions such as restrictions of use when surpassed. the reference value to within7%. The extreme deviations are, in this case, 6 times too low and a factor of 3 too high. Comparing laboratory numbers in Figs. 2 and 3, it appears that there is a group of laboratories with one (lab ) as an exception which incorrectly implement the chosen 226 Ra analysis procedures systematically. Even more problematic is laboratory performance in the determination of 2 Ra. In Water-2, there are only 14 laboratories (of 29 submitting results) that are within7% of the reference value without forming a distinct plateau of measurement results around the reference value (Fig. 4). Five laboratories determine values too low, and 1 submit results too high (by 4 %). The situation for 2 Ra in Water-1 (not shown) is similar; 16 laboratories, i.e., more than half, obtain results that are too high by 4%. One of the laboratories (no. 24) determines an excessively high activity concentration of 885 mbq L 1 (A ref = mbq L 1 ; U ref =6 mbq L 1 ; k=2) in Water-1, whereas its result for Water-2 is a factor of about 3 too low. This bad performance is unacceptable in the view of unnecessary or omitted remedial action resulting from such erratic analytical results. 1 1 226 Ra - W2 1 8 6 27 3 22 2 39 19 43 38 7 37 11 5 24 45 8 1621 15 31 26 18 46 49 4 29 Fig. 3. Mean laboratory results for 226 Ra compared to reference value of activity concentration in Water-2. Error bars and indication of reference value as in Fig. 2. 16 2 Ra - W2 1 1 1 8 6 5 19 8 21 29 15 2 39 22 38 11 4 24 Fig. 4. Mean laboratory results for 2 Ra compared to reference value of activity concentration in Water-2. Error bars and indication of reference value as in Fig. 2.

U. Wätjen et al. / Applied Radiation and Isotopes 68 (1) 1 16 15 6 5 3 1 8 U - W1 49 22 31 18 7 27 4 814 51 37 5 43 38 26 5 16 21 --> (92) 11 --> (1) 46 45 Fig. 5. Mean laboratory results for 8 U compared to reference value of activity concentration in Water-1. Error bars and indication of reference value as in Fig. 2. In general, the analysis of uranium activity concentration in mineral water is much better controlled by the vast majority of participating laboratories. For example, Fig. 5 depicts the results for the determination of 8 U in Water-1. In this case, of laboratories obtain results within7% of the reference value; only two laboratories obtain results that are too low (maximum by 45%), and six laboratories report values 4% too high (in the two worst cases, by a factor of about 8 and 11, respectively). The comparison results for the determination of 2 U in both waters and 8 U in Water-3 are similarly acceptable. 5. Conclusions Three commercially available mineral waters were provided as comparison samples. The reference values for the natural radioactivity concentration of 226 Ra, 2 Ra, 2 Uand 8 Uinthesewaters were established using independent determination methods at IRMM and Bundesamt für Strahlenschutz (BfS). Various radiochemical methods were applied by the 45 laboratories participating in the comparison. The comparison results show that there are many discrepant measurement results for the radium isotopes; 19 results, corresponding to 14% of all, are off by a factor of two or more. For uranium, this proportion is much more favourable, with only 6% (9 results out of 15) off by a factor of two or more. As these samples had rather low activity concentrations around the detection limits required by the draft EC directive and not all laboratories are routinely analysing water for these radionuclides yet, unsatisfactory comparison results for 226 Ra and 2 Ra may not be unexpected. The comparison clearly demonstrates, however, that a number of monitoring laboratories need to improve their analysis procedures for radium in order to correctly identify the drinking water sources for which remedial action (with respect to their natural radioactivity concentration) needs to be taken. References Bundesminister für Umwelt, Naturschutz und Reaktorsicherheit, 6. Messanleitungen für die Überwachung der Radioaktivität in der Umwelt und zur Erfassung radioaktiver Emissionen aus kerntechnischen Anlagen, Lieferungen 1 7, Stand 6. Chapter : Bestimmung von Uran, Plutonium und Americium mit extraktionschromatographischen Verfahren, H-U/Pu/Am- AWASS-1-1. Chapter 8: Verfahren zur Bestimmung von Radium-226 in Trinkwasser und Grundwasser, H-Ra-226-TWASS-1-1. Chapter 11: Verfahren zur Bestimmung der Aktivitätskonzentration von Radium-2 in Trinkwasser und Grundwasser, H-Ra-2-TWASS-1-1. Elsevier Urban & Fischer Verlag, München. ISBN 3-437-11456-5 and ISBN 3-437-21599-X. Burnett, W.C., Cable, P.H., Moser, R., 1995. Determination of radium-2 in natural waters using extraction chromatographic resins. Radioact. Radiochem. 6, 44. European Communities, 1998. Council Directive of 3 November 1998 on the quality of water intended for human consumption (98/83/EC). Official Journal L (5..1998) 54. European Communities, 1. Commission Recommendation of December 1 on the protection of the public against exposure to radon in drinking water supplies (1/9/Euratom). Official Journal L4 (..1) 85 88. European Communities, 1. Council Directive, in preparation. ISO, 5. International Standard ISO 1:5(E), Statistical methods for use in proficiency testing by interlaboratory comparisons. International Standardization Organization, Geneva, Switzerland. Nour, S., El-Sharkawy, A., Burnett, W.C., Horwitz, E.P., 4. Radium-2 determination of natural waters via concentration on manganese dioxide and separation using Diphonix ion exchange resin. Appl. Radiat. Isot. 61, 13 18. Pauwels, J., Lamberty, A., Schimmel, H., 1998. The determination of the uncertainty of reference materials certified by laboratory intercomparison. Accred. Qual. Assur. 3, 18 184. Spasova, Y., Benedik, L., Vasile, M., Beyermann, M., Wätjen, U., Pommé, S., 9. 2 U and 8 U in mineral water: reference value and uncertainty evaluation in the frame of an interlaboratory comparison. J. Radioanal. Nucl. Chem. 1, 1 1. Spasova, Y., Wätjen, U., Benedik, L., Vasile, M., Altzitzoglou, T., Beyermann, M., 1. Evaluation of EC comparison for 226 Ra, 2 Ra, 2 U and 8 U in mineral waters. Report EUR xxxxx EN, ISBN 978-92-79-xxxxx-x. Vasile, M., Altzitzoglou, T., Benedik, L., Spasova, Y., Wätjen, U., 9. Radiochemical separation and determination of 2 Ra in mineral waters by low-level liquid scintillation counting. In: Eikenberg, J., Jäggi, M., Beer, H., Baehrle, H. (Eds.), LSC 8, Advances in liquid scintillation spectrometry. Radiocarbon, Tucson, AZ, USA, pp. 375 38. Wätjen, U., Spasova, Y., Altzitzoglou, T., 8. Measurement comparisons of radioactivity among European monitoring laboratories for the environment and food stuff. Appl. Radiat. Isot. 66, 7 749. WHO, 1993. Guidelines for drinking water quality, Recommendation, second ed. vol. 1. World Health Organization, Geneva, Switzerland. WHO, 6. Guidelines for drinking water quality, Recommendation, third (current) ed. including the first addendum, vol. 1. World Health Organization, Geneva, Switzerland. Benedik, L., Spasova, Y., Vasile, M., Wätjen, U., 8. Determination of 2 U, 8 U and 226 Ra in bottled drinking water by alpha spectrometry. In: CD Proceedings of the 7th International Conference on Nuclear and Radiochemistry (NRC-7), Budapest, Hungary, 24 29 August 8. Benedik, L., Vasile, M., Spasova, Y., Wätjen, U., 9. Sequential determination of 21 Po and uranium radioisotopes in drinking water by alpha-particle spectrometry. Appl. Radiat. Isot. 67, 77 775. Discussion: Q(Christian Hurtgen): In the intercomparison you have asked for people to give you a result for values that were above the EC

16 U. Wätjen et al. / Applied Radiation and Isotopes 68 (1) 1 16 detection limits, although some of them can measure lower than that. It is also a good idea to ask the laboratories to give results even if there is nothing, as it is a way to look at their blank. A(Uwe Wätjen): That is correct. We had discussed this, but - seen the short reaction time in which the laboratories had to report - we had decided not to ask for values much below the detection limits required by future legislation. Q(Guy Ratel): Do you have an action plan to resolve the problems that were hidden in the measurements? We can see that only a small number of laboratories could reach this level of quality. A(Uwe Wätjen): Together with EU member state representatives some action will certainly take place, because the European recommendations will have to be adopted by national legislation if they are not in place already. There are some countries that already have clear legislation that radium and uranium have to be measured. Concerning the analysis results, we discuss these values with participants of comparisons we had organized. In general, all of our comparisons are followed by a workshop for participants, sometimes combined for several comparisons. Q(Arvic Harms): When you measure Radium 2, you separate Actinium 2. How do you correct for your losses of actinium? Do you use yield rates for the actinium? If you go back to the separation scheme for Radium 2 I think you separate the Actinium 2 from the Radium 2 and you separate it by using a column. A(Uwe Wätjen): Yes, you are right, this has been calibrated. Mirela can you help with details? A(Mirela Vasile): We did a separate experiment using a natural thorium solution and we separated Actinium 2 which we measured with gamma-ray spectrometry using the 911 kev gamma line. We obtained for our separate experiment a 83% recovery which we then used in the activity calculation as a constant to correct for the actinium recovery from the second column. Q(Arvic Harms): So you did validate the method? A(Mirela Vasile): Yes, that s what we did. Q(Pierino de Felice): Are all the participants from European member states and did they pay a participation fee? A(Uwe Wätjen): No, this comparison is free of participation fees, but it is limited. Not everybody can participate; one needs a nomination by the national representatives. Q: 5% of laboratories with % uncertainty is not really a good result if you think that these laboratories will produce scientific results from let s say unknown samples and publish their results in their papers. So we have to take this into account. Do you have any follow up action plan as to how to improve the performance of these laboratories? A(Uwe Wätjen): The action plan for us consists in organizing discussions with the labs that participated, we cannot do much more. It is the responsibility of the lab to have its analysis procedures under control. With a demonstration of reference values and documentation of how these are established, they should at least know that they measured wrong and that the reference values are reliable. It is up to them to change the procedures or their implementation, we cannot oblige them to do so. We can only point it out to them and to their national representatives or authorities who nominated them to participate in the comparison. We hope that the national representatives are then also putting pressure on the laboratories to improve where necessary. Of course, in such a workshop there is opportunity to meet those people who participated in the comparison, and hopefully they will learn from each other.