X-ray imaging technique could unlock small crystal properties

Using computing techniques that model the radiation from the X-ray source, a team from the London Centre for Nanotechnology at UCL revealed the shape of gold nanocrystals in more detail than ever before.

‘Gold is really just a model material for these studies; we’ve been looking at crystals of metal oxides, steels and cement,’ said Prof Ian Robinson, who has written a paper on the research for Nature Communications with Jesse Clark. ‘Looking at the smallest sizes of crystals is a good way to tackle questions about how cement works, for example, and where these materials’ properties come from.’

For example, he said, the strength of metals depends on stopping the materials’ crystal grains from sliding against each other. ‘One of the things that makes nanomaterial such an exciting prospect is that smaller is stronger,’ Robinson said. ‘Our imaging method will allow us to see the strain in the grains as they interact with each other, which will help us to find ways to use that strength.’

X-ray techniques are valuable for imaging nanomaterials, because X-rays can penetrate the materials, potentially giving more detail than electron microscopy. But unlike other forms of radiation, such as light, X-rays cannot be focused efficiently; lenses that focus X-rays are very inefficient, Robinson said.

Instead, X-ray imaging uses a technique called coherence diffraction imaging, first suggested by the pioneer of X-ray crystallograpy, Lawrence Bragg, in 1939. Shining an X-ray through a crystal creates a diffraction pattern as the radiation interacts with the material’s lattice of atoms. Measuring this pattern and inverting it — effectively tracing the path of the radiation through the crystal and working out what arrangement would have caused the observed diffraction — would allow an image of the crystal to be reconstructed.

‘Bragg couldn’t do that in 1939,’ Robinson said. ‘But you can do things with computers now which weren’t possible then.’

CDI is limited, however, by the properties of the X-ray beam, particularly its coherence — a property that depends on the range of X-ray frequencies and the way the beam spreads in space as it passes through the sample. This is known as the coherence function, and it’s modelling this accurately that is the key to Clark and Robinson’s technique.

‘Our results are a combination of better computing and the fact that we have better sources than in Bragg’s day,’ Robinson said. ‘He had a few photons to play with; the beam from the Diamond Light Source gives conservatively 10**8 [10 to the eighth power] photons per second in the coherence path through the sample.’

Understanding nanocrystal structure could help improve the performance of fast electronic devices, Robinson said. ‘One necessary place for these materials would be with semiconductors, where your metal wiring has to be on the same scale as the 45nm size of the semiconductor components themselves,’ he said. ‘If the wiring doesn’t connect properly, or if it crumbles because of nanocrystals forming in the metal, then you get circuit failures.’