So just what is a free-electron laser?

A free-electron laser (FEL) is not actually a laser: it is most closely related to vacuum-tube devices. A beam of relativistic electrons is passed through an undulator [a series of magnets that create a sinusoidal field]. The driving mechanism for radiative emission is the instability that develops in the electron beam owing to its interaction with the electromagnetic field in the undulator. Although the electromagnetic wave is always faster than the electrons, a resonant condition occurs such that, after one undulator period, the radiation slips relative to the electrons by a distance equal to the resonance wavelength. The fields produced by the moving charges in one part of the undulator interact with moving charges in another part leading to a growing concentration of particles. Initially the intensity of this radiation is proportional to the number of electrons, N, as all electrons are distributed randomly. But then, because of periodic bunching, the electrons start to radiate in phase, and the intensity becomes proportional to N2, leading to an amplification by many orders of magnitude. It is easy to see where multigigawatt levels of radiation power come from: if you consider a 1 GeV electron beam, with a peak current of 1 kA — multiply these numbers together and the peak power transferred from the electron beam is at the terawatt level. So even though the conversion efficiency is just a fraction of a percent, the power levels involved are impressive.

What are the key properties of FEL light?

FLASH [the free-electron laser in Hamburg] can produce gigawatt-level, laser-like, extreme-UV radiation pulses. The fifth harmonic of the radiation, at a wavelength of 2.75 nm, is shorter than that produced so far by any plasma-based X-ray device, and it is the brightest source of radiation in the world in this wavelength range. At present the pulses are approximately 10 fs long, and it should be possible, in principle, to generate subfemtosecond pulses in the future.

The first FEL reported emitted light in the infrared. How has the operation wavelength been reduced?

The FLASH facility in Hamburg, Germany. Image courtesy of DESY, Hamburg.

The experiment performed by Claudio Pellegrini and colleagues [Phys. Rev. Lett. 81, 4867–4870; 1998] was a milestone in the development of FELs. The physics of light amplification in this experiment is the same as in our machine. The jump in the wavelength to the extreme UV is due to an increase in electron energy by a factor of about 40 (from 18 MeV to 700 MeV). This is one of the advantages of FELs: they are readily tunable by changing the electron energy.

Also, in FELs at visible and infrared wavelengths, feedback can be provided by means of a resonator. This approach is made difficult in the UV region by large absorption in the mirrors, a problem also encountered in conventional UV lasers. Free-electron lasing at such short wavelengths can be achieved with a single-pass, high-gain FEL amplifier. In our machine, for each metre of the undulator length, the radiation power increases by roughly a factor of e (2.718). We reach saturation after only 20 radiation-power-gain lengths. This removes the need for feedback.

What research is done at FLASH using this intense, short-wavelength radiation?

Imaging experiments benefit from the access to shorter-wavelength light and the corresponding improved resolution: femtosecond diffraction imaging using 32-nm radiation was recently reported [Nature Phys. 2, 839–843; 2007] using FLASH. Also, 2.75-nm-wavelength light lies well within the so-called water window where biological systems can be imaged and analysed in vivo. The short pulse duration also helps: opening the door to studying different dynamical processes — biological or chemical, for example — on a timescale approaching possibly even the subfemtosecond regime.

Free-electron lasers are of course rather expensive: the cost of the FLASH facility was €117 million [90% financed from within Germany and 10% by international partners]. So many of their applications tend to focus on the parts of the electromagnetic spectrum not covered by conventional sources: X-rays and the so-called terahertz gap (submillimetre wavelength range) are examples.

A future research direction that is being seriously discussed is UV lithography. Large-scale, high-resolution patterning could provide a novel technique for the semiconductor-electronics industry. Other potential industrial applications of high-average-power FELs may involve material processing (for instance, treatment of polymer surfaces), isotope separation and chemical applications.

What does the future hold for FELs?

FLASH is the pilot facility for future X-ray FELs (XFEL). As the radiation wavelength scales inversely proportionally to the squared energy of the electron beam, another order of magnitude energy increase will bring us into the X-ray regime. The same harmonic generation technique that we describe in our paper can also work in XFELs, shifting the wavelengths below 0.1 nm. But energy is not the only consideration for the effective amplification of X-ray radiation: we also need high-quality electron beams. The electron bunches produced at FLASH are already very close to the quality required for XFELs.

The European XFEL facility, unlike alternative approaches taken in Japan and the USA, will use a superconducting accelerator technology, which will make possible not only a jump in peak brilliance by ten orders of magnitude, but also an increase by five orders of magnitude in average brilliance. There is little doubt that XFELs will provide a unique base for future studies of nature, and FLASH has demonstrated that we have a very solid foundation for the future.

Yurkov and co-workers have an article on FELs on p336 of this issue.