Improving the reliability of peak-evaluation results in gamma-ray spectrometry

In gamma-ray spectrometry the activities present in the samples are calculated from the areas of the peaks occurring in the spectra. Therefore, the utmost reliability of these input data must be maintained. To avoid systematic manual checks of the peak areas, the spectrum is analyzed consecutively by three peak-analysis programs. A comparative study was performed in which these peak-evaluation programs were tested to determine whether they yielded unbiased results. Only these programs can be used in a combined application. A consistency test of peak areas calculated by use of the three programs, for the same peak, was performed. Outlying results were discarded and the uncertainty of the best estimate for the area of a peak was calculated taking into account the dispersion of the peak-area results for this peak. The reliability of the computerized procedure has made manual checks of the peak areas unnecessary.