A Free-Electron Laser Normally Needs a 2-Kilometer Accelerator. Dresden and Paris Just Ran One Off a Few Millimeters of Plasma, Stably, in the High-Gain Regime for the First Time.

Physicists at the Helmholtz-Zentrum Dresden-Rossendorf in Germany and Synchrotron SOLEIL outside Paris have operated a free-electron laser driven by a laser-plasma accelerator stably and reproducibly in the high-gain regime for the first time, a milestone the field has chased for more than a decade. The result, published in Physical Review Letters, produced ultraviolet flashes at a wavelength of 272 nanometers with high pulse energy and the exponential power growth that distinguishes a true free-electron laser from a merely bright light source.

Free-electron lasers, or FELs, are the brightest sources of short-wavelength light on Earth. They work by driving electrons to nearly the speed of light and then sending them through an undulator, a long row of alternating magnets that forces the beam into a wiggling path. The wiggling electrons radiate, the radiation acts back on the beam and bunches it into microscopic slices, and the slices emit in lockstep, producing coherent flashes lasting femtoseconds. Facilities such as Europe's XFEL in Hamburg and the LCLS at SLAC in California use these flashes to film chemical reactions and image proteins, but they require accelerators up to about 2 kilometers long, and beam time at the handful of such machines worldwide is booked out for months.

A laser-plasma accelerator shrinks the first stage by a factor of roughly a thousand. An intense laser pulse fired into a millimeter-thin jet of gas rips electrons from atoms and drives a wave through the resulting plasma; electrons trapped on that wave surf it, gaining in a few millimeters the energy a conventional machine needs many meters to deliver. "The electrons effectively surf on a wave of plasma," said Marie Labat of Synchrotron SOLEIL, the paper's lead author. The catch has always been quality. Plasma-accelerated beams have tended to vary in energy and spread from shot to shot, and an FEL amplifies only if the beam is consistent enough for the bunching to build up over the length of the undulator.

"Stability, that is, light flashes that consistently maintain high quality over hours or even days, is extremely important for all experiments involving an FEL," said Arie Irman of HZDR's Institute of Radiation Physics. "Conventional large-scale facilities have long achieved this stability, but laser-plasma FELs have not." The team solved it by controlling the laser-plasma interaction with HZDR's high-power DRACO laser, whose infrared pulses create the plasma in the gas jet. "We managed to precisely tune the laser to match the plasma," said HZDR physicist Susanne Schöbel, a task complicated by the nonlinear processes that govern the interaction. A seed laser pulse injected into the undulator alongside the electrons enhanced the radiation and enabled the higher pulse energies.

The group had demonstrated a laser-plasma FEL in 2023, but not in the high-gain regime and not with this reproducibility. "This is significant progress in comparison with the results we published in 2023," Irman said. The next step is to push from ultraviolet into the extreme ultraviolet, the 13.5-nanometer band used to pattern and inspect the most advanced computer chips, where a compact, bright, coherent source would have immediate industrial demand.

Irman is careful to say that compact laser-plasma FELs will take "a few years" before they can complement the kilometer-scale facilities rather than replace them. But a laser that fits in a university basement rather than a purpose-built tunnel would change who gets to use one.