Experimental and Computational Dose Rate Evaluation Using S_N and Monte Carlo Method for a Packaged ²⁴¹AmBe Neutron Source

Abstract

We present a systematic computational dose rate evaluation for a packaged 1.8-Ci ²⁴¹AmBe source using both Monte Carlo and deterministic approaches, with some experimental measurements for correlation. The ²⁴¹AmBe source is stored in an extended 55-gal-drum container. Computational dose rate analysis is performed using MCNP6 (Monte Carlo) and PENTRAN (S_N) on the Center for High Performance Computing system at the University of Utah. Limited information is available regarding internal drum shielding construction, and a reverse engineering approach is presented here to estimate the dose rate and compare with measured experimental values. Our analysis shows that a deterministic three-dimensional quadrature (S_N) and anisotropic scattering (P_N) order of S₂₀P₂ is sufficient for dose rate calculations of the ²⁴¹AmBe source with polyethylene surrounding the source as shielding material. Higher quadrature orders, i.e., at least S₇₀ for neutrons and S₄₀ for photons, are needed in the presence of air due to severe streaming effects, and this is dependent upon the distance between the source and measurement locations. With air surrounding the ²⁴¹AmBe source, the Monte Carlo method is considered to be better for neutron dose calculations while the S_N method is considered better for photon dose calculations. Good agreement from both computational verification and experimental validation are observed for the dose “hot spot” in the extended 55-gal drum. The differences noted between the MCNP6/PENTRAN calculations are within 6% for the neutron dose rate and 30% for the photon dose rate. It is observed that more than 95% of the dose is attributed to neutrons. Detailed studies including a literature data validation, PENTRAN S_N convergence study, buildup factor analysis, and dose rates with different shielding materials are presented in the narrative.

Keywords:

• ²⁴¹AmBe
• PENTRAN
• MCNP6
• dose rate

Acknowledgments

This work is supported by the Energy Solutions Presidential Endowment from the University of Utah. We also acknowledge the contributions of the Utah Center for High Performance Computing (UCHPC). All MCNP6 computations were performed on the REDWOOD cluster at the UCHPC, supported by the National Institutes of Health shared instrumentation grant 1S10OD021644-01A1. PENTRAN calculations were performed on the NOTCHPEAK cluster “Civil” partition at UCHPC. The authors would like to thank David Dolan and Micah Shelley from the Radiation Safety Office of the University of Utah; Steven Pappas and Michael Hartos from the Utah Nuclear Engineering Program; and Matthew L. Lund, currently working for the Idaho National Laboratory, for assistance in this effort.

Notes

^a ANS-6.1.1, “Neutron and Gamma-Ray Fluence-to-Dose Factors,” American Nuclear Society, La Grange Park, Illinois.