Multilaboratory Intercomparisons of Neutron Dosimetry Systems

During 1965 two intercomparison studies of dosimetry systems and techniques were conducted at the Oak Ridge National Laboratory (ORNL). Personnel from several large United States nuclear installations participated. By utilizing the Health Physics Research Reactor (HPRR), it was possible to compare the various moderator-thermal detector systems, threshold and activation detector systems, sodium-activation analyses and techniques, and various solid-state detector systems. Because the reactor could be operated in either a steady-state or pulse mode, it was possible also to compare these results with measurements made with the absolute proportional counter and other laboratory systems. The various systems used compared to within ±25% for the unmoderated fission spectrum of the HPRR, but discrepancies became large, i.e. ±50% or greater, when a moderator shield was interposed between reactor and detectors. The response per unit dose of the moderator-foil system tends to increase with decreasing neutron energy, and some of the solid-state detectors indicate too low a response in this same energy region. The ORNL threshold detector system was developed in conjunction with the proportional-counter system, and, as a result, the two systems agree. However, modifications of the threshold detector system, including the use of 235U foils in Cd rather than Pu in 10B, also showed agreement to within±25% for the unmoderated and moderated spectra. Other modifications of the ORNL threshold detector system included the use of glass disks in conjunction with thin fissile foils to form the detector units. In this case the evaluation was based on the etching of the fission-fragment tracks in the surface of the glass to render them visible by use of an optical microscope. Although 235U was used instead of 239Pu in the boron ball, the comparison with the unmodified threshold detector system was within ± 10%. It is apparent that there are several systems which are suitable for nuclear accident dosimetry with an accuracy of approximately ±25%. Combinations of certain currently used systems could be used to improve the reliability and accuracy of each system. (author)