Silicon Can't Bend. These Crystals Can.

Silicon cannot bend on command. These perovskite crystals can.

Researchers at the University of California, Davis and ETH Zürich have demonstrated that halide perovskite crystals reversibly change shape when exposed to light, flexing and snapping back without damage. The work, published March 3 in Advanced Materials, shows this photostriction effect is tunable, and crucially, hysteresis-free: the crystal returns exactly to its starting point every time. The team demonstrated 20 distinct modulation states by varying light intensity.

"This is a dramatic change in the lattice when you shine light on it, a unique phenomenon that you don't see with silicon or gallium arsenide," Marina Leite, a professor of materials science and engineering at UC Davis and senior author on the paper, said in a university news release.

The team tested three perovskite compositions. Methylammonium lead bromide (MAPbBr3) showed the strongest response, with its crystal lattice expanding up to 0.3% under above-bandgap light, small in absolute terms but measurable and reversible. Cesium lead bromide (CsPbBr3) proved more rigid, changing by only 0.062%. The variation means researchers can tune the effect by chemistry: different compositions absorb different wavelengths, and the mechanical response scales with how much light the material absorbs.

The perovskite lattice expands under light because photogenerated charge carriers accumulate in the material and interact with its ionic lattice. Positive and negative ions physically shift in response. It's a solid-state effect with no moving parts. "It's not a binary on/off effect," Leite said. "It can be a scaled response, like a dimmer, depending on the light you shine on it."

Leite's group received $1 million from DARPA, the Pentagon's research arm, in November 2023 specifically to develop switchable photonic devices based on perovskites. The agency is interested in beam steering, hyperspectral imaging, and smart windows that can toggle between transparent and opaque states using only light as the trigger, no electrical contacts, no mechanical actuators. The Advanced Materials paper is a direct output of that program.

Perovskites are already further along in solar cells than in photonic switches. As Chemical & Engineering News reported, four Chinese startups are selling perovskite solar panels at megawatt scale, more than the rest of the world combined. Wonder Solar has two 200-megawatt plants operational and a 3-gigawatt facility under construction. UtmoLight launched a 1-gigawatt production line in February 2025, according to market research. The materials science of making perovskites at scale is being worked out in parallel.

The photostriction work is earlier. The experiments were done on single crystals, not integrated devices: what happens when you try to pattern these crystals into a waveguide, apply continuous illumination cycling, or run them at elevated temperature over months is still unknown. Perovskites have a long history of promising laboratory results that did not survive contact with manufacturing reality. Moisture sensitivity, thermal instability, and lead content have repeatedly stalled commercialization in areas that looked straightforward on paper.

The upside, if the physics holds at device scale, is something genuinely new: a material that changes its mechanical dimensions in response to light alone, reversibly, at room temperature, without power rails or moving components. That would be useful for optical filtering, tunable resonant cavities, or microscale actuators where wires are impractical. Whether anyone can actually build a reliable device from this effect is a separate question, and that has historically been where perovskite research goes to die.

The paper is solid. The gap from crystal to device is where stories like this typically end.