J.I.W. Watterson

Factors affecting image formation in accelerator-based fast neutron radiography

Abstract Fast neutrons can be generated with accelerators via various reactions. The reactions 7 Li(p,n) 7 Be and D(d,n) 3 He were employed for this study to generate neutrons at various energies. Radiographs of different materials have been produced and analysed in terms of contrast and resolution. Monte Carlo methods have been used to evaluate the physical factors determining the image quality in these practical situations. In each case the theoretical and experimental values were compared. A series of gamma-ray radiographs generated using a Co-60 source provided an experimental benchmark for comparison with fast neutron radiographs. The neutron radiographs were generated as a function of fast neutron producing reaction (neutron energy spectrum), scintillator type and thickness, charged particle energy, imaging geometry and sample material. Both a CCD camera operating at room temperature and cooled CCD were used in the imaging process.

Optimisation of light output from zinc sulphide scintillators for fast neutron radiography

Abstract Fast neutron radiography (FNA) is a promising new application for small accelerators. The potential effectiveness of this technique depends on the development of suitable imaging detectors for fast neutrons. Zinc sulphide based scintillators have the largest light output per event in the family of imaging scintillators used so far in fast neutron radiography. but zinc sulphide is not transparent to its own light. This paper investigates different aspects of this scintillator in order to establish the factors affecting the light output. Zinc sulphide screens were prepared by suspending ZnS(Ag) particles at different concentration in a hydrogen rich matrix. The light output of these scintillators have been tested in fast neutron fields generated by the 7Li(p,n)7Be and D(d,n)3He reactions. The light output was detected using a CCD camera coupled to the scintillator screen by an optical fiber through an image intensifier. The results are presented in the form of graphs for different sets of particle sizes and concentrations. A comparison has been made with a simple theoretical model.

DEVELOPMENT OF HIGH PRESSURE DEUTERIUM GAS TARGETS FOR THE GENERATION OF INTENSE MONO-ENERGETIC FAST NEUTRON BEAMS

Abstract Two different technical solutions to the problem of generation of mono-energetic fast neutron beams on the gaseous targets are presented here. A simple and cost-effective design of a cooled windowed gas target system is described in the first part of this paper. It utilises a thin metallic foil window and circulating deuterium gas cooled down to 100 K. The ultimate beam handling capability of such target is determined by the properties of the window. Reliable performance of this gas target system was achieved at 1 bar of deuterium gas, when exposed to a 45 μA beam of 5 MeV deuterons, for periods in excess of 6 h. Cooling of the target gas resulted in increased fast neutron output and improved neutron to gamma-ray ratio. The second part of this paper discusses the design of a high pressure, windowless gas target for use with pulsed, low duty cycle accelerators. A rotating seal concept was applied to reduce the gas load in a differentially pumped system. This allows operation at 1.23 bar of deuterium gas pressure in the gas cell region. Such a gas target system is free from the limitations of the windowed target but special attention has to be paid to the heat dissipation capability of the beam dump, due to the use of a thin target. The rotating seal concept is particularly suitable for use with accelerators such as radio-frequency quadrupole (RFQ) linacs that operate with a very high peak current at low duty cycle. The performance of both target systems was comprehensively characterized using the time-of-flight (TOF) technique. This demonstrated that very good quality mono-energetic fast neutron beams were produced with the slow neutron and gamma-ray component below 10% of the total target output.

A Monte Carlo study of the effect of neutron scattering in a fast neutron radiography facility

Abstract Monte Carlo random sampling methods offer a powerful tool for the investigation of how neutron scattering can affect image formation in a fast neutron radiography facility. A laboratory developed for accelerator based fast neutron radiography contains various apparatus which may scatter neutrons. These include vacuum pumps, the accelerator tube, target used for generating the fast neutrons and the laboratory walls and ceiling composed of high density concrete bricks used to shield the outside environment from high energy neutrons. The Monte Carlo package MCNP-4A has been used to investigate the effect of scattering on image formation for a particular neutron source and geometry. In this study the known variation of neutron source strength and energy spectrum as a function of scattering angle has been included in the modelling process in order to produce an accurate assessment of the effect of scattered neutrons in any fast neutron radiography facility.

Characterization of the 9Be(d,n)10B reaction as a source of neutrons employing commercially available radio frequency quadrupole (RFQ) linacs

Compact high current RFQ linacs have proven to be reliable and robust accelerators for the generation of neutrons for a variety of applications. The 9Be(d,n)10B reaction is often preferred in such applications due to its high cross section for neutron production and the good thermal properties of metallic beryllium targets allowing for high beam power. Thick and thin beryllium target neutron and gamma spectra have been investigated in detail using the time-of-flight technique for deuteron energies of 0.9, 1.5 and 4.0 MeV. These energies were chosen to correspond to the energies of commercially available RFQ deuteron linacs. The results show characteristics of neutron and gamma ray production of importance for the applications of this neutron source.

Optimisation of resolution in accelerator-based fast neutron radiography

Abstract In fast neutron radiography, imaging geometry, neutron scattering, the fast neutron scintillator and the position-sensitive detector all influence feature contrast, resolution and the signal-to-noise ratio in the image. The effect of imaging geometry can be explored by using a ray-tracing method. This requires following the path of neutrons through the imaging field, which includes the sample of interest. A relationship between imaging geometry and feature detectability can be developed. Monte Carlo methods can be used to explore the effect of neutron scattering on the results obtained with the ray-tracing technique. Fast neutrons are detected indirectly via neutron–nucleon scattering reactions. Using hydrogen-rich scintillators and relying on the recoil protons to ionise the scintillator material is the most sensitive technique available. The efficiency, geometry and composition of these scintillators influence the detectability of features in fast neutron radiography. These scintillator properties have a direct bearing on the specifications of the position-sensitive detector chosen to detect the light emitted.

The detector problem in fast neutron radiography

Fast neutron radiography has several advantages including the inherent brightness of an accelerator based neutron source in comparison to a reactor. In addition it can be used for the selective imaging of certain elements using fast neutron resonances (resonance imaging). One of the greatest problems in fast neutron radiography is the efficient detection of the fast neutrons. Such a system must consist of a neutron sensitive screen and a way of converting the image into an electrical signal. In most cases the method uses a screen that is based on the elastic scattering of neutrons by protons followed by the conversion of the proton energy into light in a scintillator material. This light subsequently produces an electrical signal using a CCD, or an amorphous silicon or other semiconductor screen. All such techniques involve a trade-off between image quality and efficiency. As the screen thickness increases, its efficiency also increases, linearly for small screen thicknesses, but saturating exponentially....

Physics of image formation in accelerator-based fast neutron radiography

Fast neutron radiography is an element sensitive non- destructive testing method with potential applications in industry and the detection of contraband and explosives. The physical processes that control image formation can be examined individually by a variety of analytical and experimental methods in order to determine their impact on contrast, resolution and detectability.

A Monte Carlo study of image formation in accelerator fast neutron radiography

Fast neutron radiography offers a number of advantages in comparison with the usual methods of thermal neutron radiography. It does also present a number of problems. Perhaps the most important of these problems lies in the development of a suitable imaging detector of good efficiency. The plastic fibre based scintillator, provides an efficient detector without the disadvantages of light-scatter and spreading associated with planar plastic scintillators. In this work, Monte Carlo methods have been used to study in detail the physics of image formation and degradation in scintillators, with particular reference to fibre-based scintillators, as a function of fibre length, material composition and diameter as well as neutron energy and imaging geometry. The results are being compared with experimental determinations of the edge function for both fast neutron and gamma-ray radiography.

The new Schonland PIXE microprobe and applications to geological and archaeological samples

Abstract The new positive ion microprobe at the Schonland Centre is described and details are given of its configuration, its imaging system and software. In a unique configuration, the microprobe has access to two accelerators, a 6 MV EN tandem and a 2.5 MV single-ended machine. The imaging system uses a CAMAC based system and runs on a personal computer under OS/2. The implementation of PIXE and treatment of thick-target data are presented. Applications discussed are analysis of tin artifacts from the prehistory of metallurgy in South Africa; a diamond-containing rock; carbonado, gold contained in pyrite and cogenetic minerals included in diamond.