top of page

Information engines

InfAct ERC

LOGO_ERC.jpg

The relationship between information and thermodynamics remains a fundamental thought-provoking issue since the days of Maxwell. The most effective platforms used to explore this relation are information machines: processes that convert measured information about a system to extractable work.  

A crucial question that goes beyond the currently explored regime in the field is how information machines perform out of equilibrium, namely when not in equilibrium with a thermal bath. This is the general scenario of relevance to biological systems and stands at the focus of our current work.

Providing essential, precise, and detailed experimental observations in a field in which they are lacking will pave the way toward the extension of stochastic thermodynamics to active systems. Unraveling the mechanisms that govern the conversion of information to useful work will benefit far-reaching applications. These include macroscopic and microscopic biomimetic robots and machines made of an ensemble of simple agents, analogous to natural phenomena such as cargo transport in ant colonies.

See below our first realization of an information engine based on a single colloidal particle in contact with a heat reservoir.

A single particle information machine

Our information machine is a machine capable of moving a particle against flow without applying direct force on the particle. Specifically our experimental realization of the information machine includes a colloidal particle driven by flow in a channel against an array of repelling optical beams forming a blocking wall. The particle's diffusion is recorded by sequential measurements at a given constant rate. After each measurement a decision is made; if the particle's distance from the wall (due to diffusive motion) is large enough the wall is moved closer to the particle (without direct contact). Otherwise no action is taken. This imaging and feedback cycle continues until the system reaches a steady state. This periodic process drives the particle upstream performing work. Since the wall is created by light it does not interact with the particle via hydrodynamic interactions. The amount of direct work done on the particle due to optical potential changes upon wall motion is controlled and is orders of magnitude smaller than the machine's output.

3.png
sprite.png
information_machine.gif

In the movie the laser light wall's position is indicated by the green light and the trajectory of the particle is drawn in blue. Once the laser is turned off the particle moves with the fluid flow. Tamir Admon,Saar Rahav, and Yael Roichman, Physical Review Letters 121, 180601 (2018).

bottom of page