Supercooled liquids age — their structure slowly rearranges toward equilibrium, and the rate depends on temperature. Cool the bath and aging slows. Heat it and aging accelerates. The control knob is thermal.
Unless you put the liquid in an optical cavity. Then light can selectively pump fast vibrational modes while the bath temperature stays unchanged. The structural dynamics — the slow rearrangements that constitute aging — shift without any thermal change. The cavity provides a non-thermal handle on a process everyone assumed was fundamentally thermal.
Note (2026-11-12; read against the paper's full text): three points. The results are from molecular dynamics simulations, not an experiment: a model glass-forming liquid (a dipole-active Kob-Andersen mixture, 250 molecules at 100 K) coupled to a cavity, averaged over 500 trajectories; the paper says the same mechanism should apply to slower liquids and experimentally accessible modulation times, but that is a projection. Second, the direction of the effect: in the paper stronger coupling gives progressively slower structural relaxation and aging-like dependence on waiting time, as the structure moves into lower potential-energy configurations, a deeper well that is slow to escape; it describes the liquid as behaving as if structurally colder than its surroundings. The essay's statement that the liquid ages faster toward equilibrium is not what I found. Third, the cooling protocol: it does not drain energy from structural modes. The cavity is switched on to pump fast vibrational modes, then switched off while the bath temperature is adjusted, and the cycle is repeated, relying on the timescale separation; in the simulations it cools a liquid from room temperature to a Tg of about 32 K in about 50 ps.
The mechanism: cavity photons couple to specific molecular vibrations, redistributing energy among modes in a way that changes the effective potential energy landscape the liquid explores. Fast modes get excited; the slow structural modes respond to the altered landscape. The bath temperature, measured by any thermometer, hasn't changed. But the configurational temperature — the one that matters for structural relaxation — has.
Configurational feedback cooling exploits this gap. By tuning the cavity to selectively drain energy from structural modes, the researchers reach lower effective structural temperatures than the bath would allow. The liquid ages faster toward its equilibrium without getting colder.
The broader implication: “temperature” is not one number. A system can be at one temperature vibrationally and another configurationally. When you have a tool that addresses them independently — like a tuned cavity — the single-number description breaks down, and new dynamics become accessible.