Source of correlated noise in experiments using a 252Cf source

The 2%2Cf source method to measure reactivities is under intensive study, both theoretically and experimentally, at the Oak Ridge National Lab. Potential applications to the measurement of very large subcriticalities require a careful analysis of the source of correlated noise. Still, 252Cf experiments are analyzed by applying the Schottky prescription of the source of noise according to Cohn, i.e., without corrections due to the neutrons removed by the detector; this is equivalent to assuming that detector fluctuations are exactly proportional to the fluctuation of the neutron field. If this is true, a nonzero correlation should be measured in a moderator driven by a one-neutron-per-event source, which certainly is not the case. Sheff and Albrecht emphasized later that the Schottky source of noise has to be divided into two components related to multiple neutrons per event (fission) and to a single neutron per event (absorption plus leakage), and that only the first component is the source of correlated noise for the detector. The differences between the two formalisms have very practical implications when the 252Cf method is applied to measure large subcriticalities. In this paper the authors analyze systems without multiplication in order to emphasize the differences; also additional sources of correlations due to finite resolution time are individualized and quantified. Whenever possible, a master equation approach is used to avoid a posteriori cosmetic to the Langevin/Schottky approach