Maxwell’s demon learns to exploit hidden information

For more than 150 years, Maxwell’s demon has haunted the boundary between thermodynamics and information theory. The thought experiment, proposed by James Clerk Maxwell in 1867, imagines a microscopic gatekeeper that sorts fast and slow gas molecules between two compartments of a vessel, using nothing but knowledge of their individual motions. By opening a tiny door for speedy molecules moving one way and sluggish ones moving the other, the demon gradually heats one chamber while cooling the other, conjuring a temperature difference — and therefore usable work — out of a system that was supposedly at thermal equilibrium. To generations of physicists, this seemed like a violation of the second law of thermodynamics, a free lunch served by an invisible imp.

The paradox was eventually resolved by a profound realization: information itself is a physical resource. The demon does not create energy from nothing; it converts information about microscopic fluctuations into work, and the act of measuring, remembering and erasing that information carries its own thermodynamic cost. This insight has spawned an entire research field devoted to information engines — microscopic machines that, unlike conventional heat engines, are driven not by temperature gradients but by knowledge. Now, physicists at the University of Konstanz and the Max Planck Institute for Dynamics and Self-Organization in Göttingen have pushed the concept into territory that is far more realistic, and far more strange, than the demon’s original haunts.

In a combined experimental and theoretical study published in the journal Nature Physics, the team, led by Clemens Bechinger, professor of physics at the University of Konstanz, reports that a demon does not need to see everything to profit from what it cannot see. Even when crucial information about a system remains hidden, the researchers show, a cleverly designed measurement strategy can partially recover that concealed knowledge and convert it into usable energy. The work, titled ‘The extraction of work from hidden degrees of freedom’, represents an important step toward cyclic microscopic information engines that could operate without any macroscopic energy gradient at all.

The challenge the researchers tackled stems from a simple fact of real-world physics: no demon ever has complete information. In the classical scenario, the gatekeeper enjoys perfect knowledge of every molecule’s position and velocity. In practice, even tracking a single microscopic particle with perfect accuracy is impossible, and the problem becomes dramatically worse when the particle does not swim in a simple liquid like water but in a far more complex medium — a polymer solution, a biological fluid, or some other structured material with an internal molecular microstructure. Such surroundings are not passive backdrops. They actively influence the particle’s motion, yet their internal dynamics remain invisible to any observer who can only follow the particle itself.

To make this concrete, the Konstanz and Göttingen team designed an elegant experimental system. A microscopic particle was confined by a tightly focused laser beam — an optical trap — inside a complex liquid. Unlike a simple Newtonian fluid, this medium forms a dynamic, spaghetti-like network of microscopic structures that continuously rearrange, entangle and interact with the particle. As the particle moves, it deforms this network, and the deformed network in turn exerts forces back on the particle, shaping its future motion. The particle’s position can be tracked with high precision, but the internal state of the surrounding mesh cannot be observed directly. From the demon’s perspective, the environment is only partially visible: some of the information needed for optimal control is locked away in the medium itself.

This concealment has real thermodynamic consequences. Without access to the hidden state of the network, the demon’s actions cannot be perfectly aligned with the actual condition of the system, and the amount of work that can be extracted is correspondingly reduced. The medium, deformed by the particle and relaxing only gradually, stores a memory of the particle’s past motion and influences its future dynamics — a memory the demon cannot read directly. In effect, the demon wears an eye patch, seeing the particle but not the forces that the particle’s own history has imprinted on its surroundings.

The team’s solution is as simple as it is powerful: measure twice. Instead of relying on a single observation of the particle’s position, the researchers perform two measurements separated by a controlled time delay. By comparing these two snapshots, they obtain indirect information about the hidden state of the surrounding medium. The logic is subtle but physically transparent: because the medium’s slow internal degrees of freedom respond to the particle’s motion and relax over time, the difference between two position measurements taken at different moments encodes traces of that hidden dynamics. The demon, in other words, can eavesdrop on the environment’s memory without ever seeing it directly.

Crucially, the researchers found that by carefully tuning the delay between the two measurements, the energy stored in the hidden dynamics of the environment can be harvested and converted into useful work. Under suitable conditions, exploiting this hidden information allows the demon to extract more work than would be possible from a single measurement alone. The results demonstrate that information encoded in hidden environmental degrees of freedom can be partially recovered and exploited as a genuine thermodynamic resource. By exploiting temporal correlations and environmental memory, it becomes possible to harness energy from degrees of freedom that remain inaccessible to direct observation.

The implications reach well beyond the laboratory curiosity of a laser-trapped bead in a polymer soup. At microscopic length scales, maintaining stable thermal gradients — the driving force of every conventional heat engine — becomes increasingly difficult, because thermal fluctuations grow ever more dominant as systems shrink. Information engines offer an alternative: machines that use knowledge of microscopic fluctuations as their thermodynamic fuel. The new study shows that even the information hidden in complex environments, which previously would have been dismissed as inaccessible noise, can be systematically mined for work. This is particularly relevant for microscopic machines intended to operate in biological fluids, polymer solutions and other structured materials, where internal microstructures are the rule rather than the exception.

The study, carried out by first author Lokesh Muruga and Felix Ginot, both physicists at the University of Konstanz, together with Sarah Loos of the Max Planck Institute for Dynamics and Self-Organization, under the direction of Clemens Bechinger — who is affiliated with the Centre for the Advanced Study of Collective Behaviour and whose ERC Advanced Grant on Brownian particles in non-equilibrium baths funded the research — thus reframes an old thought experiment for the messy reality of complex matter. Maxwell’s demon, once a paradoxical troublemaker threatening the second law, has become a practical blueprint: a gatekeeper that thrives on partial knowledge, squeezes energy from memory it cannot see, and points the way toward microscopic engines powered by information in environments where temperature differences are nowhere to be found.

Subject of Research: Extraction of work from hidden degrees of freedom in microscopic information engines

Article Title: Maxwell’s demon learns to exploit hidden information

Article References: Maxwell’s demon learns to exploit hidden information. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: Maxwell's demon, information engines, thermodynamics, hidden degrees of freedom, optical tweezers, complex fluids, microscopic machines, Nature Physics, University of Konstanz, Max Planck Institute, work extraction, statistical physics