X-Rated Interferometry

The appearance of an x-ray interference pattern in a Fabry-Perot interferometer has been achieved, for the first time, by a group of physicists at the University of Hamburg (Yuri Shvyd'ko, 49-40-8998-2200). This might lead to a new generation of x-ray optical devices, such as high-resolution x-ray spectral filters, or x-ray clocks, and, more important still, a new way of calibrating length measurements at the atomic scale. X-rays are a potent type of electromagnetic radiation, with a much higher energy and smaller wavelength than visible light. But because x-rays are so potent and because they see various materials as having essentially the same indices of refraction, x-rays are much harder to reflect at a surface. Indeed, x-ray telescopes in orbit use only grazing-incidence (reflected through an angle of a milliradian or less) mirrors to focus x-rays on a detector.

In the last few years, though, the scientists in Hamburg have succeeded in reflecting x-ray light directly backwards with special sapphire (Al₂O₃) mirrors; the price for this high-angle reflectivity (other than the difficulty of preparing faultless crystalline mirrors) is that the reflection occurs only for an extremely narrow spectral range (x-ray color), precluding the mirrors' use in telescopes, where x-radiation over a broad range is important. In the Hamburg device, an x-ray version of a Fabry-Perot Interferometer (FPI), the reflecting waves will resonate if the cavity between two exquisitely polished mirrors is a multiple of the radiation half-wavelength. Light entering the cavity bounces back and forth between the mirrors producing multiple sub-waves emerging from the cavity. Their interference shows up as a modulation in the radiation that exits the cavity, both on time and wavelength scales. The Fabry-Perot interference pattern provides a means of measuring of the x-ray wavelength, and this provides an opportunity for creating a new, higher-precision, length standard. Currently the most accurate way to measure x-ray wavelength is to produce a Bragg scattering pattern by sending x rays into a silicon crystal, whose lattice spacing (the distance between atoms) is known with a relative uncertainty of about 6 x 10^-8.

There is, however, a nuclear process related to the Mossbauer effect which produces x-rays (better known as Mossbauer radiation) with an extraordinarily narrow spectral line. The most familiar is the Mossbauer radiation originating from the decay of the first excited state of 57-Fe nuclei. The radiation wavelength of about 0.086 nm is perfectly suited for atomic scale measurements. Its stability, about 10^-15, is comparable to the best cesium fountain clocks. If Mossbauer x rays could be used to calibrate an FPI device capable of operating in both x-ray and visible ranges, then this could facilitate a stable, reproducible, wavelength (and hence length) standard far better than is possible (about 3 x 10^-11) with, say, helium-neon lasers.

An important step toward this goal has now been attained in the experiments of the Hamburg group conducted at synchrotron radiation facilities including the Advanced Photon Source at Argonne (near Chicago) and HASYLAB at DESY (near Hamburg). The x-rays, from the synchrotron-radiation sources, were chosen to be as similar to Mossbauer rays as possible. For the first time, interference patterns in a Fabry-Perot interferometer have been observed for x-rays. From the attenuation time of the multiple sub-waves emerging from the cavity, the spectral sharpness of the Fabry-Perot interference fringes was estimated to be less than a micro-electron-volt. This is more than 100 times better than the best x-ray crystal monochromators can do. (Shvyd'ko et al., Physical Review Letters, 10 January 2003; also see figure and related Physical Review Focus article, 17 July 2000).