SLAC Fired Lasers Into a Common Semiconductor and Watched the Charges Clump Instead of Spread

A team working at SLAC National Accelerator Laboratory has caught a common class of semiconductor doing something its own textbook description does not predict: when light knocks charges loose inside it, those charges do not spread out evenly. They bunch into dense local pockets, and the pockets are far denser than the models said they should be.

The materials are II-VI semiconductors, a family that sits inside infrared detectors and photovoltaic solar cells. The measurement was done with megaelectronvolt ultrafast electron diffraction, or MeV-UED, an instrument at SLAC that fires a very short pulse of high-energy electrons through a sample and reads how they scatter. Do it repeatedly, with the electron pulse arriving at a slightly different delay after a laser pulse each time, and you assemble a stop-motion film of atoms and charges rearranging themselves. The team synchronized laser pulses at different wavelengths with the electron probe to see how the response changed with the color of the light going in.

"The densities were much higher than we calculated," said Diana Jeong of Stanford University, an instructor of radiology and the study's corresponding author. The work also involved Patrick Kramer, a SLAC staff scientist who leads laser science for MeV-UED, and Tom Hopper, now an assistant professor of chemistry at the University of Central Florida and a postdoctoral scholar at SLAC when the measurements were taken. The results were published in Nature Photonics, DOI 10.1038/s41566-026-01894-3.

The clumping has a downstream consequence that matters more than the clumping itself. Where the carrier density spikes, the material's band gap shifts substantially — enough that the crystal starts transmitting light it would ordinarily block. A semiconductor's band gap is the energy threshold that decides which photons it absorbs and which pass through, and it is normally treated as a fixed property of the material. Here it is being pushed around by the light already inside.

That is a problem for anyone modeling these devices with the standard assumption of a uniform excited state, and an opportunity for anyone who wants a material whose optical properties can be switched on a picosecond timescale by a control pulse. The same effect that scrambles a simple model is the effect an engineer would want to exploit.

Jeong's stated interest is medical. "Our hope is that this leads to new sensing platforms that can form images right away and lead to real-time intervention," she said — detectors that produce a usable image during a procedure rather than after it.

The broader lesson is one that ultrafast instruments keep delivering. Averaged, time-integrated measurements of a material describe an average that may not exist anywhere inside it. Watch fast enough and finely enough, and the uniform excited state turns out to be a bookkeeping convenience.