The use of associated particle timing based on the D+D reaction for imaging a solid object

https://doi.org/10.1016/S0969-8043(01)00280-9Get rights and content

Abstract

Associated particle timing based on the D+D reaction has been applied for imaging a bulk sample, namely an aluminium box. The relatively low neutron energy, 2.8 MeV, allows a better spatial resolution from time-of-flight measurements. A combination of a Si detector for charged particles and an NaI(Tl) scintillator for inelastic-scatter gamma rays yielded an overall time resolution of 0.4 ns, giving a spatial resolution of better than 1 cm. A new reconstruction program was developed, yielding an image free from major artefacts.

Introduction

An important application of neutron activation analysis is in the non-destructive analysis of specimens (Csikai, 1987) or, in the medical field, the in vivo analysis of body composition (Ryde et al., 1987; for a recent review see Sutcliffe, 1996). Excellent sensitivity is obtainable for some elements that have large cross-sections for neutron capture followed by delayed-γ emission, especially when the samples are non-living and can be subjected to a very large neutron flux. Other elements have a less favourable cross-section, and some are difficult to measure because of unavoidable interference; e.g. in vivo determination of aluminium through the 27Al(n,γ)28Al reaction with a radionuclide source requires heavy filtering (Lewis et al., 1997), since otherwise, a much higher interfering γ yield is obtained from a phosphorus reaction that also decays via 28Al. An alternative in such cases is the process of neutron inelastic scattering leading to the emission of prompt γ rays. When the neutron is generated in an accelerator source, this technique can make use of the timing signal that is produced by an associated charged particle, hence leading to a significant reduction in the background contribution.

In several previous studies, the associated particle method has been described (Gordon and Peters, 1990; Kehayias and Zhuang, 1993; Vourvopoulos, 1994; Mitra et al., 1995) in which γ rays are detected in delayed coincidence with a recoil particle from the neutron-production target. In these experiments, neutrons have been generated in the D+T reaction, for which the associated particle is a nucleus of 4He. The attractive features of the D+T reaction are that the neutron yield is large, the recoil particle has a high energy allowing it to be detected easily, and the 14 MeV neutrons penetrate several centimetres into most condensed materials. The present work, however, has been carried out using the D+D reaction, which may not have these advantages but is preferable for other reasons. The low neutron energy enables only a few of the nuclear excited states to be attained for most materials, and hence simplifies the structure of the γ-ray spectrum (Jiggins and Habbani, 1976; Kacperek et al., 1990; Evans et al (1997), Evans et al (1998)). Relatively high cross-sections are found for several elements. The velocity of the neutrons is less, so giving a better spatial resolution from time-of-flight (TOF) measurements. The absence of radioactive tritium may also be an advantage in some situations.

Section snippets

Apparatus

The accelerator was a SAMES model T with modified target chamber and ion optics, operated at 140 kV. The ion source was an r.f. glow discharge (100 W, 90 MHz), and the beam was accelerated onto a self-implanted water-cooled target. A silicon surface-barrier detector with an active surface of 150 mm2 was installed in a side-branch, so as to detect recoil particles at a fixed mean angle of 135° to the beam line, and the corresponding neutrons emerged at about 32° (this angle being a function of the

Modifications to the apparatus

Some modifications were made to the target and to the charged particle detector assembly in this work with the aim of enhancing the performance of the experimental set-up. In the earlier work by Al-Lehyani (1997) and Darko (1999), the neutron yield was relatively low, at 5.7×106 ns–1 for a beam current of 0.75 mA at an acceleration voltage of 140 kV. In order to obtain a higher neutron yield, the self-implanted copper target was replaced by a home-made TiD self-implanted target. This increased the

Observation of induced γ-rays

Measurements of the neutron beam geometry, and time-of-flight resolution, suggested that good results might be obtained for an object with dimensions of order 15 cm placed at 120 cm from the neutron source. A suitable collimator was constructed to reduce the number of scattered neutrons reaching the γ-ray detector, and also reduce the general background radiation in the laboratory. In medical applications, the collimator would also have the important function of shielding the parts of the patient

Simulation studies

To a first approximation, the amplitude of the TOF spectrum at a given instant of time will correspond to data acquired from a plane at the corresponding distance from the neutron source and perpendicular to the beam axis. If this approximation is valid, then the image can be reconstructed using conventional algorithms, such as Fourier transformation, or filtered back-projection (FBP). There is, however, considerable divergence of the neutron beam at the sample position, with the result that a

Experimental results

The aluminium test object was placed in the neutron beam at 72 different positions, 5° apart, and TOF data were recorded. The raw data were acquired in timing channels that corresponded to a spatial resolution of about 13 mm, but this exceeds the resolution that the detectors are capable of. Some smoothing therefore needed to be applied by combining groups of 12 channels into bins of more realistic width, giving about 70 pixels across the picture. A typical TOF spectrum is shown in Fig. 5, for

Discussion

The above results show that the present technique is able to produce an image of the distribution of a selected element, with a resolution better than 1 cm. The technique rejects signals from any of the selected element present outside the test object. In addition, all chemical forms of the selected element are imaged, and the element can be visualised even when surrounded by a matrix of other materials, although these features were not demonstrated in the present work.

The technique is capable

Acknowledgements

Q.B.M. was supported by a Commonwealth Scholarship. The earlier phases of the project were supported by EPSRC Grant No. GR/H69397. We also thank the workshop staff of the Department of Physics, University of Wales Swansea, for technical assistance.

References (17)

There are more references available in the full text version of this article.
1

Present address: Department of Physics, University of Zambia, PO Box 32379, Lusaka, Zambia.

View full text
Copyright © 2002 Elsevier Science Ltd. All rights reserved.