Two position measurements let researchers draw additional work from a microscopic particle surrounded by a liquid that remembers earlier motion. A short interval preserved useful correlations; a long wait erased the effect. The controlled experiment accessed energy stored in the surroundings, so the gain over a memoryless comparison remains compatible with thermodynamics and offers no demonstration of large-scale power generation.
A microscopic silica particle trapped by light sat in a viscoelastic liquid, whose network retained effects of earlier movement. In the Nature Physics experiment, two position measurements therefore carried information about both the particle and its surroundings. The apparatus used this history when moving the trap to extract work, rather than relying solely on the final position.[1]
The particle was 2.7 micrometres across. Its surrounding network relaxed over approximately six seconds, making measurements about one second apart useful for retaining memory. After a two-hundred-second wait, relaxation no longer depended on the first observation. Controls matched initial position distributions while discarding history, helping separate environmental memory from a simple difference in starting positions.[1]
For certain selected measurement outcomes, the extracted work reached roughly three times the corresponding memoryless optimum. That ratio applies to those conditioned ensembles. It is not an overall efficiency figure for the apparatus, and the greatest relative increase occurred under different conditions from the greatest absolute work.[1]
This single peer-reviewed demonstration accessed energy stored in otherwise unobserved environmental motion. Accounting for the complete particle–environment system preserves thermodynamic limits. The experiment translated a trap of fixed shape; more elaborate controls and large-scale energy production were outside its measured scope.[1]
