Physicists in Germany have steered beams of X-rays around corners by carefully shining them along extremely narrow channels carved into a thin layer of tantalum – a feat that had not been expected, given how X-rays are known to undergo optical refraction. The waveguides were able to deflect X-ray beams by up to 30°, and the team says that the thumbnail-sized devices might be used for ultrahigh-resolution imaging or for analysing pulses from free-electron lasers.
Building optical devices for light at X-ray wavelengths is extremely difficult because the index of refraction for X-rays in solid materials is slightly less than it is in air. This makes it difficult to send X-rays along waveguides such as optical fibres, which exploit total internal reflection. A ray of visible light travelling from glass to air is refracted away from the normal to a significant degree because the indices of refraction of glass and air are very different. This means that the light will reflect back into the glass waveguide, even when the beam is at relatively large angles to the air–matter interface. In contrast, X-rays need to be within about a thousandth of a degree of the interface to undergo total internal reflection.
Scientists have been able to make straight X-ray waveguides by very precisely shining a beam into a long, narrow channel of air cut into a medium such as silicon or a metal. However, the tiny difference in the refractive indices meant that no-one had tried to build a curved version.
Narrow and precise
Now, Tim Salditt and colleagues at the University of Göttingen have overcome this problem. They reasoned that an X-ray beam can be sent along a curved waveguide as long as the device is sufficiently narrow and the optics used to focus the beam sufficiently precise. When a beam enters a curved channel, the part of the beam on the inside of the curve is deflected through the greatest angle when bouncing off the far side of the channel (the outside of the curve). This means that the channel entrance must be made as narrow as possible to limit the beam’s maximum angle of deflection. On this basis, the team calculated the maximum width a waveguide could have if it is to transport a beam of a given wavelength through a given angle.
To put their idea to the test, the researchers used electron-beam lithography to carve out five curved channels on a 500 nm-thick tantalum “chip” measuring 5 × 5 mm. Each channel is 100 nm wide and the waveguides have radii of curvature of 10–80 mm. They placed the chip in a 1.5 nm-wavelength X-ray beam produced by the PETRA III synchrotron source at the DESY laboratory in Hamburg. The beam was fired into each waveguide in turn and the team measured the pattern of radiation emerging on the far side of the chip. The researchers found that they could transport a significant fraction of the beam’s intensity through quite large angles – over 18° in the case of the waveguide with 10 mm radius of curvature.
Emboldened by these results, they carved out a second chip on an identical-sized piece of tantalum but with the channels in this case having radii of curvature of 1–30 mm. They then placed this chip in a 1.5 nm beam produced at the European Synchrotron Radiation Facility in Grenoble, France. Although, as before, much of the beam leaked out along the way, a significant portion of it nevertheless made it to the end of the waveguide with the 10 mm curvature – turning through 30° in the process.
According to Salditt, chips guiding X-rays through 30° could have a number of applications, including interferometers that split X-ray beams and then recombine them. Such devices could be used to measure the length of the very short pulses produced by free-electron lasers. Another possibility, he says, would be to direct X-rays to carry out very high-resolution imaging of objects such as living cells.
David Paganin of Monash University in Melbourne, Australia, agrees that curved X-ray waveguides could have a number of practical applications, and adds X-ray-based information processing and the probing of ultrafast molecular dynamics to Salditt’s list. “This development could be to X-rays what the optical fibre is to visible light,” he says.
According to Salditt, the waveguides could be made more efficient by covering the channels – currently they are open to the air above – and by replacing the tantalum with a material that absorbs fewer X-rays, such as diamond or beryllium. With the right combination of materials and beam wavelength, the team believes it may be possible to deflect X-rays through 90°, or even 180° for a waveguide having a radius of curvature of around 1 mm.
The research is reported in Physical Review Letters.
About the author
Edwin Cartlidge is a science writer based in Rome