The boundary conditions in which Poisson statistics can be applied in nuclear spectrometry are investigated. Improved formulas for the uncertainty of nuclear counting with deadtime and pulse pileup are presented. A comparison is made between the expected statistical uncertainty for loss-free counting, fixed live-time and fixed real-time measurements.
Five copper alloys were prepared with modern powder metallurgical processes in the frame of the European project Improvement of Means of Measurement on Archaeological Copper-Alloys for Characterisation and Conservation (IMMACO) and certified for As, Pb, Sn and Zn mass fractions. Similar in their composition to archaeological bronze alloys, these Certified Reference Materials (CRMs) are to be used for calibration of XRF instruments for characterization of archaeological samples either in the laboratory or in the field. This paper presents and compares the successful contribution of our non-destructive analytical methods (k0-NAA and PIXE) to the IMMACO project and to the certification of the five reference materials.
The concepts of the Guide to the expression of Uncertainties in Measurements for chemical measurements (GUM) and the recommendations of the Eurachem document "Quantifying Uncertainty in Analytical Methods" are applied to set up the uncertainty budget for k0-NAA. The "universally applicable spreadsheet technique", described by Kragten, is applied to the k0-NAA basic equations for the computation of uncertainties. The variance components — individual standard uncertainties — highlight the contribution and the importance of the different parameters to be taken into account.