Determination of the 240Pu/ 239Pu atom ratio in low activity environmental samples by alpha spectrometry and spectral deconvolution

Abstract

Accurate measurement of the 240Pu/ 239Pu ratio in environmental samples is of importance as it may provide evidence for the definitive identification of a particular source-term. Usually, the measurement is performed by mass spectrometry, unless the activities involved are relatively high. In such cases, as the published literature shows, it is feasible to determine this ratio using high resolution alpha spectrometry and appropriate algorithms to deconvolute the partially resolved 239,240Pu multiplet. In this paper, a simple technique, based on commercially-available software developed for gamma spectra analysis (MicroSAMPO®), is described by which this complex multiplet can be resolved at the much lower activities typical of many environmental samples. In our approach, it is not necessary to make any alterations to the normal alpha spectrometric set-up (including energy dispersion), other than to improve collimation. The instrumental function is defined for each spectrum by fitting a modified gaussian with exponential tails to the comparatively well-resolved 242Pu “doublet” (used as tracer) and, if present, the 238Pu “doublet”. The fitted peaks are used to create an energy calibration file with which, using published energy data, the positions (in channels) of the component peaks of the multiplet are predicted. These positions are not altered subsequently when MicroSAMPO's interactive multiplet analysis facility is used to quantify the relative spectral intensities of the components. Before calculating the 240Pu/ 239Pu ratio, it is advisable to correct for coincidence summing of alpha particles and conversion electrons. The technique has been applied to the determination of the 240Pu/ 239Pu ratio in a set of environmental samples, most of which were supplied by IAEA-MEL under their laboratory intercomparison programme. Subsequently, replicate samples were analysed independently using thermal ionisation mass spectrometry. The agreement between the two sets of data was most satisfactory. Further validation of this deconvolution technique was provided by the good agreement between the measured alpha-emission probabilities for the component peaks in the 239, 240Pu multiplet and published values.