Going beyond X-rays
Atomic Weapons Establishment scientists Giles Aldrich-Smith and Nicholas Bazin* describe the development and validation of compact room scale fast neutron imaging capabilities.
X-rays are widely used across industries for non-destructive examination of components as a means of monitoring defects arising during manufacturing, or resulting from ageing processes while in service. In some cases, X-rays are insufficient for this purpose, for example, in large and/or high-density components. In these situations, neutrons offer an alternative and complementary approach.
Neutrons interact with matter differently to X-rays, meaning that high-density materials are often transparent, and low-density ones opaque. This characteristic allows low-density materials to be imaged, for example, when stored in high-density containers. Multi-modal imaging using both X-rays and neutrons optimises the amount of information that can be obtained from the non-destructive evaluation (NDE) of difficult components. However, neutron imaging usually requires access to a major national facility.
Choose your light source
The nuclear industry manufactures specialised components that come with an associated design life based on construction, application and environmental exposure. Assessing and underwriting the integrity of components non-destructively can be performed using a range of techniques including X-ray radiography, ultrasound and eddy currents. While NDE is beneficial both in terms of cost and waste avoidance, it can also be essential on safety grounds. As a consequence, the opportunity for developing neutron imaging as a complementary NDE technique for more challenging components has arisen.
Whether or not it is possible to resolve sub-surface features in a component non-destructively is determined by the interaction of the illuminating medium – visible light, X-rays, neutrons – and the atomistic structure of the object to be illuminated. Thus, X-rays that interact with the electronic clouds of atoms can readily penetrate small, low atomic number atoms with fewer electrons, but are stopped by higher atomic number atoms with denser electron clouds. This makes them ideal for seeing higher opacity broken bones through lower opacity flesh. With X-rays, the bigger the atom the larger the degree of interaction – known as the cross-section – and the more opaque the material is. Whereas for neutrons the cross-section is not related directly to atomic mass. Neutrons, having an overall neutral charge, do not interact with the electronic clouds of atoms and instead interact with the very small nuclei at the centre of atoms. There is a complex relationship between the size of the atom and the cross-section. Neutrons have a unique ability to image low-density materials screened by high-density materials – the opposite of X-rays.
Neutron cross-sections are measured in barns, where 1 barn = 10-28m2. The term is said to have originated from the Manhattan Project during World War II, when physicists needed a secretive term for the unit describing the approximate cross-sectional area presented by the typical nucleus. They decided on barn, appropriating the American idiom, ‘couldn’t hit the broad side of a barn’, referring to someone whose aim is terrible, as they considered this a large target for particle accelerators that needed to have direct strikes on nuclei.
When discussing neutron energies, there are a number of different, and often interchangeable, units used. A beam of neutrons can be described as cold, epithermal, thermal, or fast, depending on its energy. At lower energies – colder, slower neutrons – there exists a 1V relationship between the material cross-section and the neutron velocity – the faster the neutron travels the lower the likelihood of it interacting with a nucleus. At higher energies there is a resonance region where the nuclear cross-section fluctuates intensely with small changes in neutron energy. Interrogation of this phenomenon yields a very powerful tool for neutron imaging where parts of a component can switch between transparent and opaque with very small shifts in neutron energy. Above this region (>1 megaelectronvolt) neutron cross-sections tend to plateau, with library data often inaccurate or unknown for many materials.
The UK does not currently have a high flux fast neutron imaging capability. Professor Tom Scott, a William Penney Fellow at University of Bristol, UK, said, ‘The development of such a capability at Atomic Weapons Establishment (AWE) could provide the basis of a UK-wide facility supporting many other government and commercial requirements including waste package assessment for nuclear storage facilities and other high value R&D areas - such a capability would be incredibly useful for the UK.’
How to make a neutron
Free neutrons are unstable in nature, decaying into a proton or electron with a half-life of about 15 minutes. A large amount of energy, in the form of radioactive nuclei, is required to release them, either from nuclear power stations, or by accelerating protons to very high energies and smashing them into dense metals, at a spallation source like the ISIS Neutron and Muon facility at the Rutherford Appleton Laboratory (RAL) in Oxfordshire, UK. However, both these processes are very expensive, large, and produce a wide energy spectrum of omni-directional neutrons.
For an affordable imaging capability, a compact source of, ideally, monoenergetic and directional neutrons is required. This can be achieved by utilising ionic charges and exploiting nuclear reactions that produce free neutrons. If the correct energies and parameters are chosen, a point-like beam of neutrons can be produced with a relatively efficient use of space.
A compact neutron beam for imaging is being developed at the Lawrence Livermore National Laboratory (LLNL) in California, USA. LLNL Scientist Brian Rusnak described their approach, saying,
‘The system starts with an ion source that ionises deuterium gas giving it a positive charge.
These ions are bunched and accelerated with a combination of radio frequency quadrupoles and drift tube linear accelerator structures then transported
and focused using magnets on to a deuterium gas target. The deuterons in the beam interact with the nuclei of the deuterium in the target material via a myriad of processes.’
The dominant reaction is shown (above, opposite etc.), and gives maximum neutron ejectile energy in the forward angle from the accelerator. The neutron energy and flux decrease as the accelerator off-axis angle increases. This opens up the possibility of using different monoenergetic neutrons for imaging by changing the location of the imaging station relative to the accelerator. For the LLNL example, their 7 megaelectronvolts (MEV) deuteron beam will produce maximum neutron energy of 10.2MeV along the 0-degree. This approach is also being considered as part of the high energy NDE project at AWE.
Seeing a neutron
One of the challenges of fast neutron imaging is that it is very difficult to detect fast neutrons as they have weaker interactions than slow ones. This is beneficial when trying to shine them through thick and dense objects, but less helpful when trying to convert them to photons via a scintillator – a material that converts incoming ionising radiation to visible light. Polymers are classically used for neutron scintillators as they are very rich in hydrogen and present a large cross-section to an incoming neutron, increasing the probability of interaction. The incoming neutron hits a proton – a hydrogen nucleus, knocking it out – a process known as proton recoil. This elastically scattered proton interacts with surrounding electronic clouds, causing a number of photonic emissions and re-absorptions, culminating in visible light being emitted very close to where the neutron first interacted with the proton.
These photons are detected either via a flat panel CCD, or preferably, a lens-coupled camera located outside of the neutron beam. As such, there is a trade-off to be made – thicker scintillators will have a greater chance of interacting with neutrons, but the resulting image will have a lower resolution than a thinner scintillator, as the photon origin is less defined.
Shields to maximum
High-energy neutrons and X-rays present radiological hazards and need appropriate levels of shielding to protect the users. They can also significantly degrade the quality of an image due to scattering off the target, component, camera, and the walls/ceiling/floor of the room. These scattered and reflected neutrons can interact with the scintillator – such emitted light causes a room return phenomena leading to image ghosting and resolution degradation.
The use of particle transport modelling is being assessed to investigate interactions of neutrons with materials in the neutron path, such as translational and rotational stages, cameras and concrete walls. This allows for visualisation of the likely room return scatter and shielding requirements to be assessed for an AWE facility. This technique is being validated on the cold neutron Imaging and Materials Science & Engineering (IMAT) facility at ISIS. A laser scanner was used to map the entire experimental hutch to a nominal resolution of 2mm and a higher resolution, 20µm, point cloud – a set of data points in a space – of the sample area was also collected. This data is being used to recreate a geometric reconstruction of the IMAT hutch and sample area compatible with a particle interaction software programme.
Seeing is believing
When assessing an imaging technique for quantitative feature data it must be validated. All forms of measurement have error associated with them and imaging is no exception. Here the challenge is to underwrite the resolution of the capability with respect to all possible variables including feature size, shape, composition and location within the component. For example, the same feature buried deeper in a component or within a more opaque material, may be more difficult to see. AWE scientist Nick Bazin explained, ‘Ultimately, we need to not only demonstrate that if a feature is there that we will see it, but also that if we don’t see it, it is because it not there and not just because we might have missed it.’
Verification and validation is being undertaken using a number of image quality indicators (IQIs). The simplest is the kaleidoscope test object – a tungsten block with varying pitch slots machined into it.
To further assess the ability of neutron imaging to resolve across the full spectrum of parameters, a novel conceptual designed IQI is being developed at AWE. This concept utilises an off-axis, multi-rotor-based system, where a number of concentric rotors can be orientated independently of each other. The inner rotor houses a number of pucks, which can be made to represent engineered features such as joints, inclusions, voids and cracks. Systematic rotation of the IQI can position a feature at any location within a component.
Safety first, and last
Finally, the most important aspects that must be demonstrated are the safety and non-destructive elements of a proposed fast neutron imaging capability. These are assessed against the neutron flux and energies that will be used and the different materials to be imaged. Modelling of the energy deposited, material activation or production of daughter products, or other induced degradation mechanisms, will all be evaluated against previous experimental data and assessments. Only when all of these factors have been addressed and validated will the techniques be considered safe and non-destructive.