Neutron multiplicity counting is a maturing technology. It has been implemented at many facilities to address the increasing need to rapidly measure impure plutonium bearing materials. At Hanford Site and Rocky Flats Environmental Technology Site, multiplicity counting has also been used with excellent results by the International Atomic Energy Agency to verify excess plutonium inventories now under their safeguards. Neutron multiplicity counting as currently implemented, however, will not address all forms of impure plutonium. Materials containing large concentrations of matrix elements like fluorine and beryllium cannot be assayed successfully without extremely long count times. Assays of compact plutonium metals and oxides having a large uranium concentration relative to their plutonium content tend to bias low because of a breakdown in the theoretical model now used to translate the measured multiplicity distributions to plutonium mass. In this paper, the authors will discuss the most recent efforts to extend the range of materials that can be measured successfully with thermal neutron multiplicity counting and a use of multiplicity counting to detect sample changes during long-term storage.

When samples with very low fission rates are assayed using thermal neutron multiplicity counting, shift register electronics, and the moments formalism for the singles, doubles, and triples count rates, large counting statistics errors can occur in the triples rates because of cosmic ray events. These counting statistics errors can be greatly reduced using gate multiplicity analysis when cosmic ray events produce high multiplicity neutron bursts. The gate multiplicity analysis method, measurement data, and results are presented.

The pyrochemical multiplicity counter designed and built at Los Alamos has been undergoing tests and evaluation at Lawrence Livermore National Laboratory (LLNL). Measurements have been performed using a variety of plutonium oxide and metal materials. The pyrochemical multiplicity counter uses the information contained in the higher moments of the neutron multiplicity distribution to deduce the three unknowns in the assay problem: 24°Pu-effective mass, ([alpha], n) neutron rate, and self-multiplication. This is an improvement over conventional neutron coincidence counting, which must rely on some estimate of the ([alpha], n) neutron rate or self-multiplication to deduce an assay result. Such conventional techniques are generally unsatisfactory for impure materials for which these quantities are unknown. We present the assay results obtained with the pyro-chemical multiplicity counter and discuss the procedures necessary to produce good assay results. Using these procedures, we have obtained assay accuracies of 1%--2% for oxide materials in 1/2 hour measurement times. We also compare these results to those that would have been obtained using conventional neutron assay techniques and discuss the correlations we have observed between assay results and the ratio of total neutron counts in the different rings of the pyrochemical counter.

In the past several years, several facilities have identified a need for a large multiplicity counter to support safeguards of excess weapons materials and the measurement control and accountability of large, unusual samples. The authors have designed and fabricated two large thermal neutron multiplicity counters to meet this need at two DOE facilities. The first of these counters was built for Rocky Flats Environmental Test Site for use in the initial inventory inspection of excess weapons plutonium offered to International Atomic Energy Agency safeguards. The second counter was built for the Lawrence Livermore National Laboratory (LLNL) to support their material control and accountability program. For the LLNL version of the counter, a removable, fast-neutron interrogation assembly was added for the measurement of large uranium samples. In the passive mode these counters can accommodate samples in containers as large as a 30-gal. drum. This paper reports on the measured performance of these two counters and the data obtained with them.

A weighted point model for thermal neutron multiplicity counting has been developed for the assay of impure plutonium metal samples. Weighting factors are introduced for the spontaneous fission and ({alpha}, n) contributions to the doubles and triples rates to account for the variations in neutron multiplication in these samples. The weighting factors are obtained from Monte Carlo simulations using the MCNPX code, which supports the simulation of spontaneous fission sources and can tally the source and detected neutron multiplicity distributions. Systematic behavior of the weighting factors was studied as a function of sample mass and geometry. Simulations were performed to evaluate the potential accuracy of assays performed with weighted point model analysis. Comparisons with experimental data are presented. The possible use of quads rates is explored.