UCL Engineering


Detecting the undetectable with x-ray

Transforming the way x-rays are used.

X-ray of a beetle

14 February 2020

X-ray imaging has been based on the differences in x-ray attenuation: to see a feature inside an opaque object with x-rays, this feature has to absorb more or less x-rays than the background it is immersed in, ultimately leading to a “shadow” image.

This is an effective approach – which however has limitations. Whenever detail and background have similar x-ray attenuation properties, or the detail is too thin to generate a sufficient difference in the transmitted x-rays, it may not be visible in the image.

This prevents, for example, detecting tumours before they reach a certain size, and similar limitations are encountered in all areas where x-rays are used.

Speeding to a radical new approach

There is, however, a different physical principle that can be exploited to generate x-ray images. X-rays are electromagnetic waves of very short wavelength, which take different speeds when they travel through different material.

These varying speeds lead to distortions in the x-ray wave front as it traverses an object, and these differences can be orders of magnitude larger than differences in attenuation. The challenge is to develop mechanisms sensitive enough to pick up these distortions (which are invisible to conventional x-ray systems) and convert them into image contrast.

Initially, it was thought that detecting these distortions was only possible at very large, specialised and expensive facilities called synchrotrons, only approximately 50 of which exist in the world. However, the AXIm group at UCL has developed a method that enables doing this with conventional x-ray sources like those used in hospitals.

This provides the basis for widespread use of x-ray phase contrast imaging in a wide range of applications, medical and beyond, enabling e.g. earlier detection of life threatening diseases, better security at airports, improved industrial products through advanced testing and monitoring methods, and more.