Speaker
Description
Characterizing ambient neutron backgrounds is vital for mitigating background with shielding in deep underground rare-event physics searches. We present an experimental and analytical framework to reconstruct the ambient neutron flux up to 11 MeV at the SNOLAB J-Drift location using repurposed Helium-3 (He-3) Neutral Current Detectors (NCDs). First, a direct thermal neutron flux measurement was established using bare detectors. To cleanly isolate the thermal component from the epithermal tail, an upcoming publication introduces an experimental deployment utilizing the Cadmium difference method (Thermal flux = Bare detector flux - Cd setup flux). Integrating calibration and analytical wall-effect corrections, the direct ambient thermal flux was measured to be (4.37 +/- 0.12) x 10^-6 cm^-2 s^-1. Second, we address broad spectral deconvolution (0-11 MeV) by selecting specific moderator configurations (Bare, 1-inch, and 2-inch high-density polyethylene) calibrated at a corrected density of 948 kg/m^3 to minimize the condition number of the response matrix. In trials using standard flat energy binning, direct matrix methods yielded an integrated total flux of (1.11 +/- 0.05) x 10^-6 cm^-2 s^-1, demonstrating excellent numerical agreement with benchmarking iterative algorithms (MLEM and GRAVEL). Finally, to incorporate realistic physical priors, an enhanced unfolding framework was implemented by substituting flat bins with custom energy weights: a Maxwell-Boltzmann distribution for the thermal regime, a 1/E distribution for the epithermal regime, and a SOURCES-4C raw normalized spectrum for the fast regime. Under this physically motivated weighting scheme, Singular Value Decomposition (SVD) and LU decomposition smoothly processed the well-conditioned matrix, providing a total integrated flux of (1.01 +/- 0.05) x 10^-6 cm^-2 s^-1 with highly suppressed error propagation. Due to the high computational time demands of iterative techniques and the high stability of our optimized direct linear approach, future background monitoring will pivot exclusively to direct inversion frameworks.