Inside world’s ultimate X-ray machine before it becomes more powerful

The Klystron Gallery, a concrete corridor dotted with evenly spaced metal cylinders, is long enough to extend beyond my field of vision. But as I stand inside, I know something even more spectacular lies beneath my feet.

Beneath the Klystron gallery is a gigantic metal tube that stretches 3.2 kilometers: the Linac Coherent Light Source II (LCLS-II). This machine, located at the SLAC National Accelerator Laboratory in California, generates X-ray pulses more powerful than those produced at any other facility in the world, and I’m visiting it because it recently broke one of its own records. Soon, however, its most powerful components will stop for an upgrade. Once re-ignited, perhaps as early as 2027, its X-rays will have more than double the energy.

“It will be like going from a flicker to a light bulb,” says James Cryan of SLAC.

To describe LCLS-II as a mere flicker is an understatement. In 2024, it produced the most powerful X-ray pulse ever recorded. It lasted only 440 billionths of a billionth of a second, but carried nearly a terawatt of power, which far exceeds the average annual output of a nuclear power plant. Additionally, in 2025, LCLS-II generated 93,000 X-ray pulses in one second – a record for an X-ray laser.

Cryan says this latest record paves the way for researchers to gain unprecedented insight into the behavior of particles inside molecules after they have absorbed energy. It’s comparable to turning a black and white film of their behavior into a sharper, full-color film. Between this achievement and the upcoming upgrade, LCLS-II has a chance to radically improve our understanding of the subatomic behavior of light-sensitive systems, whether photosynthetic plants or candidates for better solar cells.

LCLS-II achieves all of this by accelerating electrons until they approach the speed of light – the ultimate limit of cosmic speed. The cylindrical devices that I saw, which are the klystrons which give its name to the Klystron Gallery, are responsible for producing the microwaves which achieve this acceleration. Once fast enough, the electrons pass through rows of thousands of magnets whose poles are carefully arranged to make the fast electrons tremble. This in turn produces pulses of X-rays. Like medical X-rays, these pulses can then be used to image the interior of materials.

On the day of my visit, I visit one of the many experimental rooms where the X-rays end their journey by crashing into molecules. I take a look at some of the chambers where a molecule and an X-ray meet. They look like something out of a futuristic submarine: thick metal cylinders with round glass windows, all carefully bolted together so as not to let in any stray air molecules that might interfere with the experiment.

Cryan and his colleagues conducted an experiment the day before my visit, studying the movement of protons inside molecules. Imaging methods other than X-rays struggle to determine precisely how protons move, but precise details of the process are important for the development of solar cells, he explains.

What will happen to these investigations once LCLS-II completes its “High Energy” upgrade to LCLS-II-HE? The ability to study the behavior of particles and charges within molecules will increase dramatically, Cryan says. But achieving this will not be an easy task.

CERN and Mont Blanc, dark and frozen matter: Switzerland and France

Prepare to be blown away by CERN, Europe’s center for particle physics, where researchers operate the famous Large Hadron Collider, nestled near the charming Swiss lakeside town of Geneva.

John Schmerge of SLAC says that the more energetic the electron beam becomes, the more the team has to worry about losing just a few particles. He says he once saw an imperfectly controlled beam punch a hole in an instrument at another facility, so there’s little room for error. SLAC’s Yuantao Ding says all the new parts the team will install during the upgrade have been designed to withstand the new higher power of the facility, but it will be crucial to increase the power step by step and verify that everything is working as expected. “We will turn on the beam and carefully observe what happens,” he said.

He and his colleagues will spend most of 2026 doing a big engineering effort to put all the pieces in place, which will then prepare them for this gradual process over the next year or two. If all goes as planned, researchers around the world will be able to use LCLS-II-HE by 2030. Conversations between researchers who use X-rays, like Cryan, and those who control them, like Schmerge and Ding, will also play an important role. “At the end of the day, it’s an important tool and people will learn how to use it well,” Schmerge says. “We’re going to be constantly tweaking it.”