Physicists Fired X-Rays So Short They Outran the Damage They Were Causing — and Partly Undid It

Physicists have found that X-ray pulses can be made short enough to partly outrun the damage they inflict on whatever they are imaging — and, stranger still, that the same light can push some of that damage back.

The result, published Monday in Nature Communications by a team led by Anatoli Ulmer of the University of Hamburg and involving SLAC National Accelerator Laboratory, attacks the central compromise of X-ray imaging. X-rays reveal structure at atomic scale because their wavelength is atomic scale. But they also ionize whatever they hit, stripping electrons off atoms in a cascade that converts a solid structure into a cloud of ions and free electrons — a process called electronic bleaching. The sample stops looking like itself before the picture is finished. Every experiment is therefore a tradeoff: enough dose to get a signal, not so much that the signal describes a ruin.

The Hamburg-SLAC team illuminated neon nanoparticles with X-ray pulses lasting only a few hundred attoseconds. An attosecond is a quintillionth of a second; these pulses were 100 to 1,000 times shorter than those used in earlier experiments of this kind. They also tuned the photon energy to sit near neon's K-edge, the threshold at which the innermost electrons absorb strongly.

Two things happened. The first was expected in principle: the pulse ends before the damage cascade has time to develop. "Attosecond-scale pulses break this cycle by outrunning the usual damage cascade," Ulmer said. The second was not. "In addition, we observed a competing process: stimulated emission."

Stimulated emission is the mechanism behind lasers — an incoming photon prompts an excited atom to drop back down and emit a matching photon. Here it was running in the middle of the ionization, actively pushing electrons back toward their original states while the X-rays were still on. The light causing the damage was simultaneously undoing part of it.

The measured effect was substantial. Comparing 300-attosecond pulses against pulses 50 times longer at comparable brightness, the ultrafast pulses produced stronger X-ray signals, images that more accurately represented the sample's structure, and significantly reduced ionization damage.

"We were very surprised to see that brighter images come with less damage," said Tais Gorkhover of the University of Hamburg's Cluster of Excellence CUI. "It's a bit like reflecting more sunlight off a metal roof without the roof getting any hotter."

The conceptual shift matters more than the specific numbers. X-ray imaging has been organized for decades around minimizing exposure, a passive strategy of harm reduction. This result suggests the response of matter to X-rays can be actively steered by choosing pulse duration and photon energy, turning damage from a fixed cost into a parameter an experimenter controls.

The obvious application is single-particle imaging at free-electron laser facilities, where researchers try to photograph individual viruses, proteins and catalysts without first freezing them into crystals. Those experiments have always been limited by how much structure survives the flash. If the flash can be made to partly heal what it breaks, the ceiling moves. Neon nanoparticles are a clean test case; biological samples are far messier, and whether the effect survives that complexity is the next question.