Scientists built a camera fast enough to catch atoms undressing — and discovered the simulations were wrong

At the European XFEL facility outside Hamburg, physicists have a new reason to distrust their simulations. A diagnostic technique they spent years building has finally produced a direct measurement of what happens to copper atoms in the first trillionths of a second after a high-intensity laser hits them — and the result contradicts the models. ScienceDaily covered the work on May 1, 2026. The underlying paper, from Dr. Lingen Huang and colleagues at the Helmholtz-Zentrum Dresden-Rossendorf, appeared in Nature Communications on April 3.

The finding matters to anyone running high-energy-density physics simulations, and especially to those designing laser fusion experiments: the computer codes that model how plasma behaves under extreme conditions overestimated how much the plasma heats up. Whether that is because the models overestimate collisional heating rates, underestimate radiative cooling, or miss some other process entirely remains unresolved. What the experiment shows is that the measurement is now possible — and that the models need calibration.

The setup was a hair-thin copper wire, about one-seventh the thickness of a human hair, positioned at the HED-HiBEF experiment station. The researchers fired a 3-joule, 30-femtosecond pulse from the high-intensity optical laser ReLaX at the wire, then probed the plasma 25 femtoseconds later with 8.2-kiloelectronvolt X-rays from the X-ray free-electron laser, tuned exactly to the copper Cu²²⁺ electronic transition. By varying the delay between the two lasers in one-femtosecond increments, they built a movie of ionization happening at the electron-dynamics timescale.

What they watched: Cu²²⁺ ions appeared within a picosecond of the laser hitting the wire, peaked in abundance at 2.5 picoseconds, and fully recombined and vanished within 10 picoseconds. The entire event is faster than a single camera flash. The wire target geometry, a 10-micrometer-diameter cylinder, solved alignment problems that plagued previous flat-foil experiments, where spatial and temporal jitter in the XFEL beam created significant uncertainties in the measurements.

The off-resonance emissions observed on both sides of the XFEL photon energy, with comparable yields, indicate balanced ionization and recombination rates. That balance is what the models failed to predict — they heated the plasma more than the experiment showed. The technical term for these kinds of models is particle-in-cell codes: computer simulations that track how charged particles move and interact in the electric and magnetic fields inside a plasma. They are the standard tool for designing inertial confinement fusion experiments, and getting their physics right determines whether your hohlraum design, laser pulse shape, and capsule compression will actually work.

NLTE — non-local thermodynamic equilibrium — is the regime where electrons in a plasma are not in thermal equilibrium with each other or with the ions. This is the standard state of matter in laser fusion experiments, in astrophysical plasmas, and in any system where intense laser pulses interact with solid-density targets. Getting the collisional processes wrong in NLTE conditions means your simulation predicts the wrong temperature, the wrong pressure, and the wrong plasma behavior as it expands. The paper states the comparison "reveals that typical models overestimate the plasma heating under the extreme conditions achieved in our experiment, highlighting the requirement for improved modeling of NLTE collisional processes for predictive capabilities."

The wire target geometry, a 10-micrometer-diameter cylinder, solved alignment problems that plagued previous flat-foil experiments, where spatial and temporal jitter in the XFEL beam created significant uncertainties in the measurements. Compared with the copper foils used in earlier work, the cylindrical target eased the alignment and overlap of the optical laser and XFEL on the target, while localizing heating effects in a smaller volume.

Prof. Tom Cowan, former director of the Institute of Radiation Physics at HZDR, was involved in the work. Dr. Ulf Zastrau, who runs the HED-HiBEF station at European XFEL, has been explicit about the applications to laser fusion. The diagnostic technique — resonant X-ray absorption spectroscopy combined with emission tracking at matched photon energies — is now a validated tool for probing ultrafast ionization dynamics that the field has not had before.

For builders working on inertial confinement fusion, this is a calibration problem. The simulations that tell you whether your hohlraum will work, whether your laser pulse shape is right, whether your capsule will compress symmetrically, depend on getting the plasma physics right at the electron dynamics level. A diagnostic that proves those models are wrong, even in a narrow regime, is worth paying attention to. Huang et al. have given the field a measurement that did not exist before. Now the modelers have to catch up.