In a dark room, a resonator catches light the way a spiderweb catches dew drops. A laser pulse bursts in, and the mechanical membrane starts to vibrate in sync, leaving a trace in the very fabric of the system. That is how memory is born: not in a silicon cell, but in the dance of photons and phonons.
Laser pulses act as a conductor, making light and the resonator's mechanics dance while leaving a trace in the system's memory. If the conductor flicks the baton sharply, the orchestra cannot respond in time: energy dissipates, and the dance ends with the last note. But if the baton moves smoothly, keeping the rhythm, the musicians fall into resonance: even when the conductor falls silent, the strings still vibrate. That residual sound is memory.
In the new work, physicists showed that the shape of laser pulses determines whether memory accumulates. With narrow Gaussian wave packets, the system returns to its initial state after each cycle, like a pendulum that stops if not pushed. With square-sinusoidal pumping, the output remains nonzero even at zero input: energy accumulates, and the resonator 'remembers' previous pulses. Hysteresis, the dependence of the state on its history, becomes controllable.
To measure memory, the authors introduced a geometric form factor. This is a number from 0 to 1 that shows how close the input-output loop is to a perfect circle. The closer to 1, the better the memory. For photons, the value reached 0.965; for phonons, 0.963. That is much higher than for photonic quantum memristors (0.58) and polaritonic systems (0.2).
The practical benefits are huge. Optomechanical memristors could become elements of neuromorphic processors, where each resonator acts as an artificial synapse with memory. Already today, the proposed protocols can be implemented on existing setups, such as microtoroidal resonators. The next step is experimental verification and studying robustness to thermal noise, to bring the technology closer to room temperature.
Looking ahead, arrays of coupled optomechanical resonators could weave a network of memories with light and mechanics. This is not science fiction but tomorrow's engineering, where quantum memory is born from the rhythm of laser pulses.
The form factor used to assess memory is based on the isoperimetric inequality: among all closed curves of a given length, the circle encloses the maximum area. Therefore, an ideal circular hysteresis loop corresponds to maximum memory efficiency.