The 31Si half-life

Half-life values are of paramount importance in analytical techniques based on radioactivity to rescale the measured activity to a reference time. An in depth analysis of the papers reporting the results used to calculate the recommended values discloses a broad lack of information on the performed experiments and, occasionally, an almost complete absence of uncertainty budgets [1]. This was the case for 31Si, the radionuclide produced via (n, γ) reaction adopted to quantify the 30Si mole fraction of the 28Si-enriched material manufactured to determine the Avogadro constant [2]. The now recommended 157.36(26) min 31Si half-life is the average of ten data accepted without a detailed discussion of the uncertainty budget [3]. Aiming at improving the reliability of the 31Si half-life value we carried out an experiment based on repeated observations of the 31Si decay via neutron activation and γ-counting. The resulting value, 157.16 (20) min, was recently published together with the evaluation of the uncertainty [4]. Details of the experiments, including the measurement model and the corrections applied to achieve the declared 0.13% relative standard uncertainty will be discussed during the presentation.