An intercomparison of γ-spectrometry on two samples of biological origin by eight laboratories in four countries

Abstract

This report gives details of the first inter-laboratory comparison of γ-spectrometry to be run within SPERA, the South Pacific Environmental Radioactivity Association since its inauguration in 1991. Laboratories in Australia, Chile, French Polynesia and New Zealand participated in the exercise. Two ‘unknown’ samples of biological origin were analysed. The first was a sample of milk powder derived from IAEA reference material. This sample provided an assessment of overall accuracy of 134Cs, 137Cs and 40K determinations. The second sample consisted of dried fish flesh including natural 40K and spiked with a mixed nuclide solution containing 210Pb, 109Cd, 54Mn, 60Co and trace 133Ba. Together the samples gave information on analytical precision over a range of energies and activities. When the results were compared with the recommended values and confidence intervals of the IAEA reference material, the overall accuracy of the γ-spectrometry analytical procedures was found to be good. The average mean values for combined laboratory data fell within the recommended value ranges for each isotope. Ninety percent of the individual laboratory isotope mean values were within two standard errors of the 95% confidence interval of the standard, 75% were within 1 s.e., and 33% of the analyses fell within the confidence interval. Technical precision was also adequate with the overall errors being of the same magnitude as that of the reference material values for each isotope with relative standard deviations of 5–10%. There was a tendency for standard deviations of the combined results to be larger than those reported or derived from individual laboratory results by a factor between 1.2–5.6. This result suggested an under-estimate of systematic errors within individual laboratories. The largest sources of error were derived from reporting and calculation of results which gave a 16% gross error rate.