Article - Issue 51, June 2012
Imaging in Four Dimensions
Professor Phil Withers FREng and Professor Peter D Lee
Advances in X-ray tomography allow researchers to capture three-dimensional images of the interior of objects. By collecting a series of these images, a 3D movie can be made. Known as 4D imaging – where time is the fourth dimension – the movie can show how the internal structure of an engineering component evolves under stress or how a cancer drug acts on a tumour. Professors Phil Withers FREng and Peter Lee of the Manchester X-ray Imaging Facility explain.
How to take a 3D picture. A series of 2D radiographs is collected, viewing the object at different angles as it is rotated, here illuminated by a parallel beam of X-rays
X-ray imaging has come a long way since the discovery of X-rays in 1895. Working with a cathode ray tube in his laboratory, German physicist Wilhelm Röntgen discovered that when a high voltage was applied, his instrument could produce “mysterious rays” capable of passing through most things, including human tissue.
A week later, Röntgen took the first X-ray image, a radiograph of his wife’s hand, clearly showing her wedding ring and bones, and published a paper on his discovery. Within months of publication, scientists worldwide were producing images, and even battlefield physicians were using X-rays to locate bullets in wounded soldiers.
In the decades that followed, two-dimensional X-ray imaging and X-rays continued to be widely used. In 1912, Max von Laue, Walter Friedrich and Paul Knipping obtained the first diffraction pattern of a crystal using X-rays, and by 1953 James Watson and Francis Crick had resolved the structure of DNA, again using X-rays.
Now, scientists and engineers worldwide are using 4D imaging to understand how and why events take place: be it chemical plant corrosion, turbine blade failure or the behaviour of cancerous tumours.
Advances in imaging
A two-dimensional projection (such as a radiograph) shows the lateral relationship between constituents of an object, say, the contents of your airport luggage, or the length of a bone fracture, but provides no information about the depth of features in the object.
Thanks to advances in computer processing power, it is now possible to reconstruct a 3D image from a series of two-dimensional X-ray radiographs of an object viewed from different angles. For there to be enough information to recover a 3D image with high spatial resolution, we generally need between 200 to 2,000 radiographs, taken as the object is rotated so as to view it from many angles.
Radiographs can be collected in various ways, depending on the application. For medical imaging, the X-ray source and detector move around the patient in a spiral motion. For inanimate objects, the object tends to be rotated in front of a point source of X-rays, or illuminated with a powerful, parallel beam of rays from a synchrotron source. With the set of radiographs collected, the 3D structure of the object is then reconstructed by a computer (see panel: Reconstructing a 3D image from a series of 2D views).
Provided the X-rays can penetrate the object, the imaging is non-destructive. X-ray penetration increases with energy, so if, for example, a geologist wishes to image fossils within thick rock, as long as the X-ray energy is high enough, this is possible. With today’s imaging systems, objects up to one metre in length can be imaged with great clarity. At the other end of the scale, smaller objects can be observed to spatial resolutions as high as 50 nm.
Until recently, the focus was on collecting 3D images, but since the imaging process can be repeated over and over again these techniques can now be used to resolve processes over time – the fourth dimension. This recent step from 3D to 4D X-ray imaging is partly due to improvements in computing processing power, as many thousands of radiographs need to be analysed, but recent developments in X-ray instrumentation have also helped.
Synchrotron sources such as the European Synchrotron Radiation Facility in Grenoble and the Diamond Light Source at Harwell are incredibly powerful and bright, enabling researchers to capture thousands of images a second.
Operating at a slower pace, laboratory X-ray systems can monitor processes taking place over periods from minutes to years. The latest X-ray systems at the Manchester X-ray Imaging Facility have been designed as walk-in scanners or with large sample spaces. These are large enough to install the experimental equipment necessary to study how specimens evolve over time under stimuli. Sophisticated loading rigs or environmental cells can ensure sample conditions are carefully controlled and facilitate an uninterrupted view of the sample as it rotates during data acquisition.
Left: carbonate pore system (colour coded) in microporous carbonate rock (field of view ~2 x 1 x1.5 mm). Right: extracted pore connectivity map with the pipe diameters reflecting the pore channel sizes, from which permeability, or the ease of extracting oil, can be calculated
Imaging natural structures
Engineers and researchers from a range of industries are now using 4D X-ray imaging. As the oil and gas industry explores more remote regions in a bid to discover and exploit untapped reserves, the technique is proving indispensable. Although rocks, soils and sands are normally very heterogeneous, petrophysicists have frequently relied on idealised, homogenised microstructures to predict strength and permeability. A recently developed tension/compression rig with built in rotation allows real-time imaging of the evolution of cracks and damage in large rock and soil samples.
The rig has been combined with a purpose-built permafrost simulation chamber to image permafrost soils. This allows direct calculation of their permeability and to estimate greenhouse gas release as they thaw. Experiments performed in collaboration with scientists from the Civil Engineering Department at Imperial College London produced unexpected results that could save the petroleum industry vast sums of money.
Our imaging revealed micro-cracking in the ice during freeze-thaw cycles. Melting will occur at the micro-crack surfaces first, and geologists believe these might be the initiation sites for so-called ice lenses, localised layers of ice that form in permafrost soils causing heave as they grow and subsidence when they melt, destroying roads, buildings and pipelines. Ice lens formation leads to permafrost heave which reshapes local landscapes and has already cost the petroleum industry millions of pounds.
Frost heave aside, the efficient recovery of hydrocarbons is crucial to the petroleum industry and relies on a clear understanding of the petro-physical behaviour of rocks. These same imaging techniques are now being honed to perform nano-scale imaging on shale rock, to better understand why some shale formations yield more gas than others.
It is also being used to understand the impact of pore size, shape and connectivity on, say, fluid flow, through a rock formation. Researchers can now image the porosity within samples of heterogeneous carbonate rocks down to a micron and can extract data to describe how the pores are connected. From the 3D image of the porosity network one can calculate petro-physical properties with the resulting data fed into flow simulators, to reconstruct fluid flow in real time. The researchers soon hope to construct a set of rules that demonstrate how different pore geometries affect hydrocarbon recovery.
In many cases, optimising the structure and microstructure of materials is critical to optimising performance; 3D movies are starting to help refine and optimise manufacturing processes. The production of aluminium alloy components for automotive and aerospace industries is an example. Components are manufactured via liquid metal casting, but defects within their structure
at a microscopic level often form during the process, with the finished piece then having to be scrapped, or worse, failing in service.
Using the Diamond Light Source facility in Oxfordshire, we have worked to create 3D movies of what takes place during the casting of an aluminium alloy. We have been able to see how defects form, grow and then finally combine or coalesce leading to final fracture (see image below). Armed with
this information, we are developing models of how these, and other defects form. These models are being used by automotive and aerospace companies to design their component manufacture to avoid the formation of such defects, for lower cost and more energy-efficient production.
Another manufacturing process being examined is powder metallurgy. This is used to make high performance metallic parts by pressing metal powder in a die before sintering (fusing the powder) at high temperature to form a fully dense product. Movement of the particles during powder transfer as the powder takes up the shape of the product can lead to regions of high and low density giving rise to non-uniform contraction during sintering and in some cases cracking. By taking a series of images it is possible to track the powder movement within the die.
In many cases it is critical to understand the events that occur leading up to failure. Take, for example, thermal barrier coatings (TBC) used to thermally protect turbine blades from the extreme temperatures experienced during combustion in an aeroengine. Here, the single crystal nickel superalloy blade is coated with a metallic layer, which acts as a bond-coating. The ceramic thermal barrier coating is then deposited onto the metal bond-coat, providing thermal insulation.
Within the harsh operating conditions of an engine, an aluminium oxide layer grows between the metal bond-coat and ceramic thermal barrier layer. 4D X-ray imaging allows us to see how it forms beneath the coating. Indeed, recent experiments reveal various characteristic defects that arise as the thermally grown oxide develops (circled in Imaging turbine blades figure, bottom image on following page). Understanding the way in which it forms, as well as its morphology, is critical to delaying coating failure and preventing catastrophic damage to the entire component.
Movement of powder in a cylindrical die determined from a series of 3D images: here powder is transferred by the downward movement of the punch. Red regions are becoming more densely packed, those dilating in blue and the arrows show the particle flow
Repair processes in the human body
Medical science and biomedical engineering are massive growth areas for high resolution 4D imaging. Indeed, one of the first applications of 4D imaging was to study the response of cancers to treatment. Ex-vivo four-dimensional imaging is now being used to investigate the action of cancer treatments on the vasculature on which cancers depend. Without a functional vascular supply to provide nutrition and remove waste products, solid tumours cannot grow larger than 2mm3.
X-ray tomography is also an excellent way to study repair processes non-invasively. Working with a group of researchers at Imperial College London, who are designing tissue scaffolds to stimulate
and support bone growth, we have used 4D imaging to measure how bioactive glass scaffolds dissolve when implanted into bone.
Using a profusion bioreactor and simulated body fluid, imaging reveals the glass dissolving, with different colours indicating exactly how the ions dissolve and redistribute depending on preferential flow patterns. This is key information for the successful replacement of synthetic biomaterial with natural tissue during recovery.
The use of four-dimensional imaging is likely to grow significantly over the coming years, revolutionising the level of information available to the engineer. Further dimensions are being added to the images by superimposing information gleaned from other imaging modalities; for example, positron emission tomography (PET) and X-ray computed tomography (CT) are commonly combined in medical imaging. Using multiple modalities it is possible not only to produce 3D images showing regions of different density, but to image the chemical composition, oxidation state, crystal structure, and crystal orientations of different features.
This will be accelerated by the development of detectors that can see in colour. A Manchester-led consortium is developing detectors that can tell you not just that an X-ray photon has arrived on a given pixel but also the energy it had. In one potential application, this will enable engineers to chemically fingerprint materials in a 3D image, perhaps to highlight a security threat or the presence of drugs in baggage.
In parallel, computer reconstruction algorithms are now being introduced that will reduce the number of radiographs needed to obtain a high resolution 3D image. This will increase the 3D image frame rate, allowing the capture of shorter timescale events and also reduce the dose of radiation required to collect a 3D movie, which is important when tracking cancers.
In the future, 3D glasses might be just as useful to engineers and doctors as to cinemagoers.
More details about the MXIF and further images and movies can be found at www.imaging.manchester.ac.uk