An Experimental Validation of Spectrally Matched Neutron Detection Systems Using ³He and BF₃

Abstract

Here we perform the matching of neutron counts in two detector gasses through capture reactions and radiation transport–optimized moderating materials. One of our detectors uses helium-3 (³He) gas and has been widely used as a neutron detection material in proportional detector tube designs. This study examines boron trifluoride (BF₃) as a potential gas for neutron detection in place of ³He based on a previously studied “spectrally matched” design derived from deterministic adjoint analyses that closely mimic the spectral response of ³He. The integrated spectral response of each tube, i.e., the count rate, is calculated and measured at various distances from an isotropic neutron source where similar “total sources” are achieved in either detection system. Our results show the integrated spectral response of a dual BF₃ tube detector was within 10% of a single ³He tube when exposed to a similar source. GEANT4 Monte Carlo simulations were used to calculate the total source for each detector and showed count rates within 5% of those produced by MCNP, providing a strong confidence in its behavior in the thermal energy regime. We provide results in this study to partially validate the replacement based on the spectrally matched design, which will lead to further validation through the utilization of multiple neutron spectra via simulated and experimental studies.

Keywords:

• ³He
• BF₃
• GEANT4
• neutron detection
• spectral matching

Acknowledgments

The support and resources from the CHPC at the University of Utah are gratefully acknowledged. We also thank the University of Utah Radiation Safety Office and UNEF staff for their ongoing efforts in supporting work with the Pu-Be source.

Disclosure Statement

Author G. Sjoden owns a controlling financial interest in the PENTRAN 3-D Sn code through HSW Technologies LLC.

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/00295450.2023.2203291.

Notes

^a As a note, this does not consider any losses due to detector deadtimes or electronic losses. For this study, these are essentially negligible as deadtimes for all detectors were all well below 1% and discriminators were set to reduce noise contributions in the final count rate.

^b The bias is calculated via $Bias\left(\mathrm{\%}\right)=\frac{R-R_{ref}}{R_{ref}}\times 100$, where R and R_ref are the GEANT4 and MCNP reaction rates, respectively.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.