Determination of environmental radioactivity at two different concentration levels

In two recent intercomparisons, the performance of laboratories in the determination of radionuclides in environmental materials and in food was again investigated. The first dealt with strontium isotopes and 137Cs in liquid milk at levels of nanocuries per litre, the second with the same radionuclides and others in fresh water, at levels of picocuries per litre. In both intercomparisons the results could be compared with known input values. Thirty-two laboratories from 14 countries participated in the milk intercomparison. Except for 137Cs, the number of outlying results was small (6%). The standard deviation was highest for 89Sr (54%). The differences of the overall means from the input values were not significant. In the second intercomparison, 32 laboratories from 19 countries returned 108 results. The lower concentration level led to a wider scatter of the data, 20% had to be rejected, the average standard deviation was 32% and the average difference of the overall means from the input values 28%. The positive bias found for all the radionuclides was probably due to incomplete separations. In conclusion 90Sr and 137Cs were adequately determined in the first intercomparison (higher concentration), 89Sr was not. At the lower concentration level used in the second intercomparison, all the radionuclides investigated, except tritium, were found to present analytical problems. (orig.)