Experimental characterization of FANT, a new thermal neutron source

Highlights

• FANT is a thermal neutron source based on ²⁴¹Am/⁹Be source inside HDPE cavity.

• Neutrons from a²⁴¹Am/⁹Be source are moderated and scattered in the system.

• Designed by Monte Carlo to increase the thermal neutron flux.

• Bonner Spheres Spectrometer (BSS) used to determine the neutron spectrum.

• BUNKIUT (UTA4) and NSDann Ver 4.0 spectral unfolding codes used.

Abstract

FANT is the acronym of Enhanced Thermal Neutron Source (Fuente Ampliada de Neutrones Térmicos, in Spanish). This is a parallelepiped box of high-density polyethylene moderator and an isotopic neutron source. The moderator has a cylindrical irradiation chamber where a rather uniform thermal neutron flux is obtained. The FANT design was previously optimized and the neutron spectra were estimated by Monte Carlo calculations with the MCNP6.1 code. To check the characteristics of the FANT thermal neutron field, measurements have been performed at the reference point inside the irradiation chamber with a Bonner sphere spectrometer holding a small ⁶LiI(Eu) thermal neutron detector. To unfold the neutron spectrum BUNKIUT with UTA4 response matrix and NSDann Ver 4.0 codes were used. Some issues have been found and recommendations are made about the use of large BSS inside narrow spaces, and about the capacity of NSDann code to unfold these kind of spectra. However, the results confirm that the moderation process in FANT is very effective and allows obtaining useful thermal neutron fluence rates.

Introduction

Thermal neutron fields can be obtained from isotopic sealed fast neutron sources, if inserted in large polyethylene moderator assemblies with a cavity in which neutrons thermalize by scattering with the walls. These kind of sources can be used for dosimetry, calibration and testing of neutron detectors. Some examples are the ETHERNES (Bedogni et al., 2016) and HOTNES (Bedogni et al., 2017) facilities, which were designed by using Monte Carlo methods, with the aim of increasing the thermal neutron fluence rates, reaching values from about 550 to 800 cm⁻² s⁻¹ in the first case and from about 700 to 1000 cm⁻² s⁻¹ in the second, depending on the irradiation plane chosen inside the cavity.

FANT is the acronym of Enhanced Thermal Neutron Source (Fuente Ampliada de Neutrones Térmicos, in Spanish) that has been recently built at the Neutron Measurement Laboratory of the Energy Engineering Department of Universidad Politécnica de Madrid (LMN-UPM) (Cevallos Robalino et al., 2020). The system is made of high-density polyethylene (HDPE) moderator material with a ${}_{95}{}^{241}\text{A}\text{m}/{}_4{}^9\text{B}\text{e}$ neutron source of 111 GBq nominal activity, with strength of (5.83 ± 0.14)x10⁶ s⁻¹ (Vega-Carrillo et al., 2009). Externally, FANT is built with several HDPE blocks that together have a shape of a parallelepiped with dimensions of 90 × 70 cm² in the base and 76 cm height. Internally, it has a horizontal cylindrical irradiation chamber of 32 cm-diameter and 70 cm-long. The ²⁴¹Am/Be neutron source is housed in the chamber, with a shadow cylinder placed between the irradiation area and the source to take advantage of moderated backscattered neutrons. The moderation process is effective allowing to obtain thermal fluence rates from about 1100 cm⁻² s⁻¹ to 1300 cm⁻² s⁻¹ inside the irradiation area (Cevallos Robalino et al., 2020). The device design was previously optimized by Monte Carlo calculations with MCNP6.1 code (Pelowitz et al., 2014).

In order to characterize the neutron field, the neutron spectrum was measured using a Bonner sphere neutron spectrometer (BSS) (Bramblett et al., 1960). The BSS is one of the most commonly used neutron spectrometers (Knoll, 2010), covering a wide energy range (thermal to hundreds of MeV) (Awschalom and Sanna, 1983). It consists of a thermal neutron detector placed in the center of a set of HDPE spheres whose diameters may range from 5.08 to 45.72 cm (2″–18″). As neutrons pass through each sphere, epithermal and fast neutrons are scattered in the polyethylene losing energy until they either reach thermal equilibrium or leave the moderator. Each combination of sphere and its interior detector is characterized by an absolute response function The relationship between the response functions (R_i(E)), the detector count rates (C_i) and the neutron fluence (Φ(E)) is described bythe Fredholm integral equation of the first kind (Awschalom and Sanna, 1983), shown in Equation (1).

$$C_i=\underset{Emin}{\overset{Emax}{\int }}{R}_i\left(E\right)\cdot \Phi \left(E\right)dE$$

Here, E_min and E_max are the energy values where the response functions are defined.

The discrete version of equation (1) becomes an ill-conditioned equation system, shown in Equation (2). Here the amount of energies are larger than the amount of detectors, therefore the neutron spectrum is obtained through the unfolding procedure.

$$C_{i,j}=\sum _{j=1}^n{R}_{i,j}·{\Phi }_j$$

Here, C_i is the count rate measured with the ith sphere and Φ_j is the neutron fluence rate in the jth energy group and R_i,j are the response functions.

To unfold the above equations system, there are several unfolding codes and most of them, except those based on Monte Carlo methods or on Artificial Intelligence, use iterative routines needing an initial spectrum to start the unfolding; the quality of the final solution can be affected by the initial guess spectrum (Vega-Carrillo et al., 2006a, 2006b).

In order to check the characteristics of the thermal neutron field obtained with the ²⁴¹Am/Be isotopic neutron source located in the FANT, the aim of this work was the characterization of the spectrum at the reference point inside the irradiation chamber by measuring it with the BSS and comparing it with the spectrum obtained by Monte Carlo calculations. Measurements were carried out with a BSS holding a small ⁶LiI(Eu) thermal neutron detector. To unfold the neutron spectrum the BUNKIUT (Lowry and Johnson, 1984) with the UTA4 response matrix (Hertel and Davidson, 1985) and the NSDann Ver 4.0 (Vega-Carrillo et al., 2006a; Ortiz-Rodriguez et al., 2014) codes were used.