Calibration of Passive Neutron Counters Based on an Uncertain Neutron Source Term

In-situ passive neutron measurements of the K-25 gaseous diffusion cascade have been performed from the late-1970s to present day. The NDA program was transferred to Bechtel Jacobs Company in 1998, and for the period from 1999 through 2005, I was responsible for ensuring the material was properly accounted for by means of a measurement verification and validation program. This was accomplished by implementing a statistical methodology for selecting re-measurement locations, performing the measurements, and running the statistical model to demonstrate compliance with the DOE Order. The modus operandi for measurement engineers was: “don’t change anything. We are making precision measurements, not accurate ones.” This means that a significant amount of guesswork was required by the field engineer to ensure that the in-situ measurement was performed under identical conditions to the set of measurements prior. How else can you meet the repeatability requirements of the statistical methodology? If you measure too high, you’ve created material from nothingness. Measuring too low was not an acceptable option either. In 2005, DOE-EM asked us to examine the reliability and sustainability of the program. There were many technical, managerial, and project execution challenges to solve. In my mind (and to this day), the most difficult technical challenge was bounding the uncertainty of the neutron source term for “wet” uranyl fluoride: UO2F2 · nH2O. The more we examined the theoretical models in SOURCES 4C for slowing down and capture of alpha particles on fluorine, with the resulting emission of the “source” neutron, we realized that the uncertainties were greater than first thought, upwards of a factor of two. This paper will examine the radiation transport assumptions in SOURCES 4C, explain the origin of the various published values for (a,n) reaction rates, and provide an explanation for the compromises we made in determining the real but unknown reaction rates in the cascade for which the passive neutron counters were calibrated. Uncertainties in the neutron source term were propagated through both the calibration and measurement process, accordingly.