A team at Lawrence Berkeley National Laboratory kept a compact free-electron laser running for more than eight hours without a single manual adjustment, a milestone that moves the technology from fleeting lab demonstrations toward the kind of sustained, hands-off operation that working scientists actually need. The result, published in Physical Review Accelerators and Beams, documents how a laser-plasma accelerator at Berkeley’s BELLA Center delivered 100 MeV electron beams at one pulse per second for over 10 hours, producing coherent violet light at 420 nanometers for more than eight of those hours in what physicists call the self-amplified spontaneous emission (SASE) regime.
For context, conventional free-electron lasers occupy entire campuses. The Linac Coherent Light Source at SLAC National Accelerator Laboratory stretches roughly a kilometer. The BELLA system fits inside a single large room. If compact machines like it can eventually reach shorter wavelengths and higher repetition rates, universities and hospitals could host their own intense light sources rather than competing for scarce beam time at a handful of national facilities.
Why eight hours matters
Free-electron lasers generate extraordinarily bright, tunable light by sending high-energy electrons through a series of magnets called an undulator. The electrons wiggle, emit radiation, and under the right conditions that radiation feeds back on the beam to produce a powerful, coherent pulse. The catch is that the electron beam must stay within extremely tight tolerances of energy, alignment, and charge from shot to shot. Drift outside those bounds and lasing stops.
Previous plasma-based FEL experiments had demonstrated impressive physics in short bursts. A 2021 study in Nature achieved lasing at 27 nanometers, deep in the extreme ultraviolet, using a laser wakefield accelerator, but stability from shot to shot remained a major hurdle. A 2022 Nature paper showed that a beam-driven plasma wakefield accelerator could also power an FEL, broadening the menu of driver technologies. And a subsequent Nature Photonics study demonstrated a seeded FEL driven by a compact laser-plasma accelerator, prioritizing spectral control over endurance.
What none of those experiments showed was the ability to hold lasing conditions steady for anything close to a working day. The BELLA team’s eight-hour autonomous run fills that gap. “Once the machine was tuned, it required no manual adjustment for the duration of the eight-hour lasing run,” the researchers wrote in the Physical Review Accelerators and Beams paper, underscoring that the system maintained beam parameters well enough to sustain lasing without human intervention for the full session.
Building on earlier gains
The new record did not appear out of nowhere. An earlier study from the same collaboration, published in Physical Review Letters, reported more than 1,000-fold amplification of light at 420 nm, with gain lengths between 16.7 and 22.5 centimeters. During that campaign, the accelerator delivered electron bunches with better than 90 percent shot-to-shot reliability over roughly an hour at one pulse per second, proving that beam quality and undulator alignment were sufficient for substantial FEL amplification even if the system could not yet run all day.
A Berkeley Lab news release describes how the team extended those results by refining the stability of the laser-plasma accelerator and improving how the beam couples into the undulator. The release positions the BELLA facility as a testbed for advancing compact FEL concepts and credits earlier theoretical and experimental work that established key scaling laws for plasma-driven beams. A December 2024 seminar at SLAC cited roughly 5,000-fold SASE gain from the same facility, consistent with the upward trajectory from the earlier 1,000-fold result. The progression followed a deliberate strategy: first achieve high gain in short intervals, then hold beam parameters within tight tolerances for hours at a time.
Open questions and caveats
Several important details remain unsettled. One involves the repetition rate. The peer-reviewed paper describes the laser-plasma accelerator operating at 1 Hz, meaning one high-quality electron bunch per second. A Phys.org summary of the same work refers to roughly 1,000 bunches per second, a figure three orders of magnitude higher. The discrepancy has not been publicly reconciled by the research team. It may reflect a different operating mode, a planned upgrade, or a misinterpretation in secondary coverage. Until clarified, the 1 Hz figure from the journal article should be treated as the reliable description of the conditions during the eight-hour run.
There is also a subtle but meaningful distinction between accelerator uptime and FEL uptime. The plasma accelerator produced usable electron beams for more than 10 hours, but the FEL itself lased autonomously for just over eight. The gap likely reflects intervals when beam parameters drifted outside the narrow window required for lasing even though the accelerator kept firing. Without a full time-resolved breakdown of energy spread and emittance over the entire session, outside observers cannot pin down exactly how often the system hovered near, but not quite at, lasing conditions.
The biggest caveat is wavelength. At 420 nm, the BELLA FEL operates in the visible range. Many of the most sought-after applications, from imaging individual protein molecules to probing the electronic structure of quantum materials, require soft or hard X-rays with wavelengths below 10 nanometers. The earlier Nature and Nature Photonics experiments show that plasma-based accelerators can in principle reach far shorter wavelengths, but none of those demonstrations combined short wavelengths with multi-hour reliability. Claiming that compact plasma-based FELs are ready to replace kilometer-scale X-ray facilities would outrun the published evidence.
Visible-light endurance versus the X-ray horizon
Cost, power consumption, and deployment timelines remain largely unquantified. Berkeley Lab and SLAC communications emphasize compactness as a major advantage, implying that shorter accelerators could be cheaper to build and easier to staff. But none of the available sources provide concrete capital-cost estimates or energy budgets for a mature plasma-based FEL facility, so any specific savings figures would be speculative at this stage.
What the BELLA result does establish, as of spring 2026, is that a compact laser-plasma-driven FEL can hold lasing conditions stable for most of a working day at visible wavelengths and a modest repetition rate, under carefully controlled laboratory conditions. That is a necessary, though not sufficient, step toward broadly accessible user facilities. Reaching X-ray wavelengths, pushing repetition rates from one pulse per second toward thousands, and engineering systems that non-specialists can operate will each demand their own breakthroughs.
For now, the eight-hour run stands as the clearest evidence yet that plasma-based free-electron lasers are maturing from fragile, one-shot experiments into instruments that can hold steady long enough to do real science. The distance between a single room at Berkeley and a kilometer-long tunnel at SLAC is still measured in physics problems yet to be solved, but it is shorter than it was a year ago.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
