REPOSITORY > RESULTS

Abstract

Neutron activation analysis (NAA) is an analytical method used for determination of the elemental mass fractions in different samples. It is based on the measurement of characteristic radiation, usually γ-rays, emitted from radionuclides formed through the process of neutron irradiation of the target element. The k0-NAA method is an NAA variant, in which the nuclear constants of an analyte are grouped with those of a comparator into a combined nuclear parameter k0 that is used in calculations of the measurement result.
The mass fraction of an element in the sample should always be reported with its corresponding uncertainty. The uncertainty is the measure of the quality of the result and the key for establishing the traceability of the measurement result.
In this dissertation, uncertainty propagation factors of the k0-NAA method were calculated, according to metrological principles for all the parameters directly involved in the measurement results. To facilitate the calculations, a computer program called ERON was developed. The behaviour of the uncertainty propagation factors under different measurements conditions was analysed and the uncertainty contributions of various quantities were evaluated. Critical parameters with associated high uncertainty propagation factors that strongly contributed to the compound uncertainty of the final result were identified.
To confirm the theoretical framework experimentally, samples with well defined mass fractions of an added analyte and known uncertainty were prepared. In order to study the effect of the nuclear parameters on the final uncertainty, thirteen elements with different nuclear constants were chosen. The prepared synthetic samples were irradiated, analysed and the mass fractions of the added elements were calculated using the same procedure as is used in real sample measurements. Using the developed computer program ERON, the combined uncertainty of the result for each measurement and for each reaction was calculated. The uncertaintes depended on the irradiation and measuring conditions, as well as on the reaction type and the observed γ-ray energy. A comparison of measurement results with known values from the sample preparation phase confirmed statistical control of the method.
A comparison was made between the combined uncertainty calculations of several different methods: the ERON program, a commercial program called GUM (GUM-PC 1999) and the spreadsheet calculator (Kragten 1994). All three programs returned similar results in accordance with their expected accuracy. The comparison confirmed the validity of the approach used in the ERON program. However, additionally developed models which were incorporated into the ERON program could not be tested with the remaining two programs.
The intrinsic uncertainty of all the elements from the IUPAC database (Kolotov and De Corte 2004) and the literature-based flux uncertainty of various irradiation conditions in different nuclear reactors were calculated. Further, the uncertainty contributions of different nuclear parameters to the combined uncertainty for different types of reactors were compared. For all nuclides, the intrinsic and flux uncertainty is larger in less thermalized reactors compared to well thermalized ones, where there are several nuclides for which both uncertainties are negligible.
A statistical analysis of measured and reported results for the certified reference material NIST 2782 revealed strong correlations between measurements of different elements at different times. A hypothesis to explain such correlations was developed. Using the described approach it is possible to detect oscillations of certain parameters during the measurements and correct the results accordingly.