Tagging fast neutrons from an 241Am/9Be source

We report on an investigation of the fast-neutron spectrum emitted by 241Am/9Be. Well-understood shielding, coincidence, and time-of-flight measurement techniques are employed to produce a continuous, polychromatic, energy-tagged neutron beam.

Sources of fast neutrons for controlled irradiations include nuclear reactors, particle accelerators, and radioactive sources. Drawbacks associated with nuclear reactors and particle accelerators include their accessibility and availability, as well as the very high cost per neutron. In contrast, radioactive sources provide neutrons with a substantially lower cost per neutron. Drawbacks associated with radioactive sources include the complex mixed field of radioactive decay products which complicate the experimental situation. As a first step towards developing a source-based fast-neutron irradiation facility, we have employed well-understood shielding, coincidence, and time-of-flight (TOF) measurement techniques to attenuate and subsequently unfold the mixed decay-product radiation field provided by an 241 Am/ 9 Be (hereafter referred to as Am/Be) source, resulting in a polychromatic energy-tagged neutron beam.

Am/Be source
The heart of the irradiation facility consists of a (nominal) 18.5 GBq Am/Be radioactive source [15]. This source is a mixture of americium oxide and beryllium metal contained in an X.3 capsule 2 (see Fig. 1).
Radioactive 241 Am has a half-life of 432.2 years and decays via α emission (5 different energies averaging ∼5.5 MeV) to 237 Np. The dominant energy of the resulting background gamma-rays from the decay of the intermediate excited states in 237 Np is ∼60 keV. 237 Np has a half-life of over 2 million years. 9 Be is stable.
Fast neutrons are produced when the decay α particles interact with 9 Be.
Depending on the interaction and its kinematics, 12 C and a free neutron may be produced. The resulting free-neutron distribution has a maximum value of about 11 MeV and a sub-structure of peaks whose energies and relative intensities vary depending upon the properties of the Am/Be source containment capsule and the size of the 241 AmO 2 and Be particles in the powders employed -see the detailed discussion presented in Ref. [20]. In general, approximately ∼25% of the neutrons emitted have an energy of less than ∼1 MeV with a mean energy of ∼400 keV [15]. The average fast-neutron energy is ∼4.5 MeV.
Both the gamma-ray and neutron dose rates at a distance of 1 m from our unshielded source in the X.3 capsule were measured to be 11 µSv/h, for a total 2 An X.3 capsule is a tig-welded, double-layered, stainless-steel cylinder approximately 30 mm (height) × 22 mm (diameter). The kinematics and the reaction cross section for the 9 Be(α, n) interaction determine the state of the recoiling 12 C nucleus produced in the reaction. The calculations of Vijaya and Kumar [22] (for example) suggest that the relative populations of the ground/first/second excited states for the recoiling 12 C nucleus are ∼35%/∼55%/∼15%. If the recoiling 12 C nucleus is left in its first excited state, it will promptly decay to the ground state via the isotropic emission of a 4.44 MeV gamma-ray. Mowlavi and Koohi-Fayegh [23] as well as Liu et al. [24] have measured R, the 4.44 MeV γ-ray to neutron ratio for Am/Be, to be approximately 0.58. Again, this is seemingly dependent upon the Am/Be capsule in question. Regardless, almost 60% of the neutrons emitted by an Am/Be source are accompanied by a prompt, time-correlated 4.44 MeV γ-ray. We exploit this property of the source to determine neutron TOF and thus kinetic energy by measuring the elapsed time between the detection of the 4.44 MeV γ-rays and the detection of the fast neutrons. Note that by employing this technique, we necessarily restrict our available "tagged" neutron energies to a maximum value of ∼7 MeV as 4.44 MeV of the reaction Q-value are "lost" to the de-excitation gamma-ray.

YAP:Ce 4.44 MeV gamma-ray trigger detectors
The 2 YAP:Ce 3 fast (∼5 ns risetime) gamma-ray trigger detectors (hereafter referred to as YAPs) were provided by Scionix [25]. A detector (see Fig. 2) consisted of a cylindrical 1" (diameter) × 1" (height) YAP crystal [26] coupled to a 1" Hamamatsu Type R1924 photomultiplier tube (PMT) [27] operated at about −800 V. Gains for the YAP detectors were set using a YAP event trigger and standard gamma-ray sources. Typical energy resolution obtained for the 662 keV peak of 137 Cs using such a detector was about 10%. YAP:Ce is  radiation hard and quite insensitive to neutrons of all energies, which makes it ideal for detecting gamma-rays within the large fast-neutron field of the Am/Be source. We stress that because of their small volume, the YAP detectors were not used for spectroscopy, but simply to trigger on any portion of the energy deposited by the 4.44 MeV gamma-rays emitted by the source. A 3 mm thick Pb sleeve placed around the source (see Sec. 2.4) to attenuate the high intensity 60 keV gamma-ray field and a 350 keV ee discriminator threshold proved to be an effective combination for the YAP detection of these 4.44 MeV gamma-rays.

NE-213 fast-neutron and gamma-ray liquid-scintillator detector
The NE-213 [28] fast-neutron and gamma-ray detector employed in this work is shown in Fig. 3

Configuration
A block diagram of the experiment configuration is shown in Fig. 4. The Am/Be source was placed so that its cylindrical-symmetry axis corresponded to the vertical direction in the lab at the center of a 3 mm thick cylindrical Pb sleeve (with the same orientation) to attenuate the 60 keV gamma-rays associated with the decay of 241 Am 4 . A YAP detector was placed with its crystal approximately 5 cm from the Am/Be source at source height. The crystal orientation was such that its cylindrical symmetry axis also corresponded to the vertical direction in the lab. This detector triggered overwhelmingly on the 4.44 MeV gamma-rays radiating from the source which came from the decay of the first excited state  cosmic rays and room background 5 .

Electronics and data acquisition
The analog signals from the YAP trigger detector and the NE-213 detector were passed to LRS 2249A and 2249W CAMAC charge-to-digital converters (QDCs) and PS 715 NIM constant-fraction (timing) discriminators. The resulting logic signals from the discriminators were passed to LRS 2228A CAMAC time-to-digital converters (TDCs) and LRS 4434 scalers. These signals were recorded on an event-by-event basis for offline processing using a LINUX PC-  framework. Connections to VME and CAMAC crates were respectively facilitated by a SBS 616 PCI-VME bus adapter and a CES 8210 CAMAC branch driver. In YAP calibration mode, signals from a YAP detector were periodically employed to trigger the DAQ and thus monitor the gains of the YAP detectors.
In TOF mode, signals from the NE-213 detector were used to trigger the DAQ so  Figure 5 shows a contour plot of the energy deposited in the NE-213 detector as a function of "pulse shape" (PS, see below) versus "L" (the energy deposited in the LG QDC). PS was calculated using the "tail-to-total" method [17][18][19]; namely, the difference in the energies registered by the LG and SG QDCs was normalized to the energy registered by the LG QDC. As the NE-213 scintillator responded differently 6 to gamma-ray and fast-neutron events, the two distinct distributions appeared in the PS versus L contour plot. Particle identification 6 In the liquid scintillator NE-213, gamma-ray scintillations are fast while neutronassociated scintillations have pronounced slow components. Analysis of the time structure of the scintillation components leads to particle identification (PID) and is known as pulseshape discrimination (PSD).  The sharp (blue) unshaded peak centered at about 2 ns is known as the "γflash" 7 . The gamma-flash corresponds to a pair of prompt, time-correlated gamma-rays produced in the source which triggered both the NE-213 detector and the YAP detector. The ∼1.8 ns FWHM of the gamma-flash is consistent with the timing jitter on our PMT signals. The tail of events to the right of the gamma-flash corresponds to non-prompt gamma-rays 8 and randoms (see below). The broad (red) shaded peak centered at about 25 ns corresponds to 7 The instant of the production in the source of the correlated pair of events which produce the time-of-flight data is known as "T 0 " and is located at a time-of-flight of 0 ns. 8 A non-prompt gamma-ray can result from inelastic neutron scattering.  The method of tagging the 4.44 MeV de-excitation gamma-ray and a comparable Am/Be source 10 were employed by Geiger and Hargrove [37] in obtaining the results shown in the middle panel. Both the neutrons and the gamma-rays from their source were detected in Naton 136 plastic scintillators. Agreement 9 While we employ the reference spectrum in our discussion of results, the interested reader may prefer Refs. [40] and [41]. 10 Their source capsule was slightly smaller and emitted about 50% more neutrons per second.

Results
12 with our results is very good. We attribute the small difference in the strengths observed in the two measurements to neutron-detection efficiency and acceptance effects which we do not consider. We attribute the relative broadening of their measured neutron distribution with respect to ours to their quoted poorer than 12% energy resolution for neutron detection, which based on the numbers Kumar [22], Van der Zwan [38], and De Guarrini and Malaroda [39]. The details of these calculations are beyond the scope of this paper, but clearly all three are in reasonable agreement both with each other as well as our results. We conclude we are tagging neutrons.

Summary
We have employed shielding, coincidence, and time-of-flight measurement techniques to tag fast neutrons emitted from an Am/Be source as a first step towards developing a source-based fast-neutron irradiation facility. The resulting continuous polychromatic energy-tagged neutron beam has a measured energy structure that agrees qualitatively with both previous measurements and theoretical calculations. We conclude that our approach works as expected, and anticipate that it can provide a cost-effective means for detector characterization and tests of shielding. We note that this technique will work equally well for all Be-compound neutron sources.