Some materials just bend when you tell them to. These bend when you shine light on them.
Researchers at the University of California Davis and ETH Zurich have demonstrated that halide perovskites reversibly distort their crystal lattice when exposed to above-bandgap light — not through heating, not through some second-order effect, but as a direct photomechanical response. The work appears in Advanced Materials (DOI: 10.1002/adma.202521800).
The team tested three perovskite compositions. Methylammonium lead bromide, known as MAPbBr3, showed the strongest response: up to a 0.3 percent change in its lattice parameter under X-ray probe measurement (eScholarship). Formamidinium lead bromide came in lower. Cesium lead bromide, CsPbBr3, proved most resistant to deformation, shifting just 0.062 percent. The team demonstrated the effect across 20 distinct excitation states and repeated cycles without measurable degradation or hysteresis — the kind of repeatability that matters when you're talking about a material property rather than a one-off observation (PubMed).
Marina S. Leite, the UC Davis materials scientist who led the work, put it plainly in a university statement picked up by EurekAlert: there is a dramatic change in the lattice when you shine light on it, and it is not something you see with silicon or gallium arsenide (EurekAlert). That comparison matters. Silicon doesn't do this. GaAs doesn't do this. The perovskite chemistry — tolerant of point defects that would kill a conventional semiconductor — appears to enable a photomechanical coupling that mainstream materials simply don't have.
The crystals were grown by collaborators Bekir Turedi, Andrii Kanak, and Maksym V. Kovalenko at ETH Zurich. Kovalenko's group is well-known in the halide perovskite space; this is not a one-off result from an obscure lab. The work was funded by DARPA (award HR00112190008) and the National Science Foundation. No conflicts of interest were declared.
Leite has said the effect could open new device categories — light-controlled sensors, tunable photonic components, actuators with no moving parts. The dimmer analogy comes up in Phys.org coverage: not binary on/off but a scaled response tunable by light frequency and power (Phys.org Tech). That is a plausible long-term vision.
But.
The 0.3 percent figure is where this story deserves some friction. A sub-percent lattice change in a lab crystal is real physics. It is also a different thing than a working device. The paper demonstrates the effect convincingly. It does not demonstrate a switch, an actuator, a photonic modulator, or anything you could hold in your hand. The applications are projected, not fabricated. There is a distance between "we measured this in a crystal under X-rays" and "you will buy a device built on this in 2030." That distance is not trivial to cross.
The history of halide perovskites in photovoltaics is instructive here. The materials have been "on the verge" of transforming solar for a decade. Oxford PV, which once held the record for perovskite-silicon tandem efficiency, has yet to reach mass commercial deployment. The stability problem — moisture, heat, ion migration — has been the wall every time. Photostriction faces its own version of this: whether the effect can be integrated into a thin film architecture, whether it survives packaging, whether it operates fast enough for the device someone wants to build. None of that is answered yet.
This is basic research wearing engineering language. The science is solid. The device projections are the usual press-release extrapolation from a result that works in a crystal but has not yet found a form factor that anyone has shown works in practice. Leite and her collaborators have found something genuinely unusual — a material that bends when you shine light on it, reversibly, repeatably, in measurable degrees. Whether it becomes anything more than an interesting property of a perovskite crystal is a question the paper cannot answer.
That question is worth watching. It is not answered yet.
