A review of the prompt neutron nu-bar value for 252Cf spontaneous fission

Abstract Passive neutron multiplicity counting can be used to determine the absolute emission rate of 252Cf spontaneous fission sources. Such sources can then be used to characterize and calibrate instruments used to assay Pu for materials control and accountancy, nuclear safeguards, and nuclear security applications. An important contribution determining the ultimate accuracy of this technique is how well the neutron multiplicity distribution of 252Cf is known (actually the low order factorial moments are typically used). This knowledge governs how well the emission rate of the 252Cf sources can be determined. Additionally, because 252Cf is used as a reference standard against which other spontaneously fissile systems are compared, it also governs the scaling of the 252Cf characterization and calibration data to the corresponding performance estimates for the spontaneously fissile Pu-nuclides. The main scale setting parameter is nu-bar, the average number of neutrons released following fission, of 252Cf. This is a fundamental nuclear constant. In this work we briefly review the 14 high quality absolute determinations of nu-bar and subject them to a full covariance evaluation. This is compared to several variants of weighted mean, including an extension where an allowance is made for the uncertainty in the estimated variance on each point. The prompt nu-bar value was found to be robust to the different statistical assessment methods applied, and moreover, the overall uncertainty estimates were of similar magnitude. Based on the full covariance treatment a prompt nu-bar value of (3.7573 ± 0.0056) prompt neutrons per fission was obtained, where the uncertainty is at the 1– σ level (68% confidence interval). This compares favorably with the value of Zucker and Holden, (3.757 ± 0.010) prompt neutrons per fission, long used in the nuclear safeguards community. The relative standard deviation of 0.15% obtained from the analysis reported here is sufficiently small that we advocate for a new high accuracy determination of the delayed neutron contribution, since this is needed to correct the Mn-bath class of nu-bar measurements. Collectively the measurements span the period 1963 to 1985. Of the 14 measurements six carry very little weight (individually 2% or less), five carry intermediate weight (individually 3%–7%), while three values are highly weighted (individually 15%–32%). These have reported relative standard deviations of between 0.2% and 0.3%. Given the small number of measurements with small reported uncertainties, and the fact that the youngest such determination is over 30 years old, we also strongly suggest that new measurements of the highest currently achievable quality are needed. In this regard a high efficiency 3He based counter would remove the residual concern over γ -sensitivity of doped liquid scintillator tank data.