Testing a Alternative Singles Rate Dead Time Correction Algorithm for use in Neutron Multiplicity Analysis

Neutron multiplicity analysis based on multiplicity shift register (MSR) logic applied to the pulse train in hardware or to list mode data is an established non-destructive assay technique for the quantification of spontaneously fissile materials. To obtain accurate results requires the data to be corrected for dead time losses. Dead time also influences the observed variance. Dead time correction algorithms are being evaluated at Los Alamos National Laboratory (LANL) as part of a strategic R&D program to enable correlated neutron measurements to be applied with confidence to ever more demanding scenarios, for example; reflective or highly multiplying items. One can treat dead time as part of the detector characteristics in a forward calculation. The most common assay solution or inversion involves first correcting the observed rates and adopting the simple algebraic point-model rather than running complex forward calculations iteratively. In examining the present day expressions for the dead time corrected singles, doubles and triples rates derived from MSR histograms according to the widely deployed dead time correction scheme developed by Dytlewski, we find that implicit within them is an alternative form of the singles dead time correction. The correction is assay item specific in that it derives from the dead time perturbed histograms themselves rather than being a function only of the observed singles rate. This means that the simplistic approximate correction for the trigger (singles) rate employed by Dytlewski may be replaced so that an internally self-consistent formalism covering all three multiplicity rates is obtained. A correction factor that depends on the concentration of correlated events present on the pulse train, as encoded in the histograms is to be expected and deals with the in-burst dead time losses. In this work we present the alternative singles correction factor. Dytlewski is not the only dead time correction scheme in common use and we examine other approximations and important results to shed light on the bigger problem.