A New Method for the Determination of the Neutron Multiplicity Counter Dead Time Parameter

Correlated neutron counting using shift register analysis is an important and long established tool for the detection and quantification of fission events in waste and safeguards. A crucial step in the manipulation of the resultant multiplicity histograms to correlated event rate is the application of correction factors for dead-time losses. Dead time correction can limit the accuracy especially when highly efficient counters are used to measure items of high neutron output where a high proportion of the primary neutrons are random (e.g. (a, n) events). There is presently no complete treatment for this problem and so advances in instrumentation design to lessen the effects are an important active area of interest. But these approaches will themselves require similar correction albeit of lesser magnitude until a higher event rate is reached. Thus dead time corrections will always remain important. At present the most popular approach to making allowance for dead time losses is based on the approximation developed by Dytlewski. In this scheme the losses in the signal triggered rates can be compensated for by using histogram multipliers which are functions of a single free parameter - the dead-time parameter, t. In practice however some additional empirical adjustments are often included. The issue then becomes of how to best find the optimum value of the dead-time parameter. In this paper we present a new method based on measuring the Singles, S, Doubles, D, and Triples, T, rates for a set of Cf-252 sources spanning a wide dynamic range. The plot of the uncorrected T/D rate ratio against the observed S-rate is shown empirically to be linear with a slope of 4.t. We demonstrate this behavior for a wide variation of counter types spanning a wide range of efficiency and a wide number of preamplifiers in the system. The new approach is quick and simple to implement and is visually striking. It appears to offer a robust and accurate approach to what is other wise a somewhat involved none-linear optimization process.