Backward Heat Flow Bends the Law a Bit

Physicists have for the first time accomplished what the second law of thermodynamics seems to forbid: transfer of heat from something cold to something hot. But don't expect to replicate the experiment at home. The weird result was witnessed last year onboard the Mir space station.

In a normal liquid, heat usually diffuses from hot regions to cold regions. If a liquid boils, heat can also move via bubbles of hot gas or liquid. These two processes generally overwhelm a third, little-known process, called the piston effect: When the liquid surrounding a bubble is heated, it expands, which compresses the bubble and warms the gas.

But there is no convection in the microgravity of the Mir space station. So a team of scientists set out to look for the piston effect. They warmed a copper-and-sapphire-walled cell filled with a drop of liquid sulfur hexafluoride and one bubble of sulfur hexafluoride. Within seconds of heating the container, the bubble temperature had risen 23% above that of the wall, the team reports in the 1 May Physical Review Letters. For this to happen, heat must have been transferred from the cooler walls to the hotter gas, they say--seemingly violating the second law.

How is this possible? The key is that the mixture of bubbles and liquid becomes acutely compressible, says John Hegseth, a physicist at the University of New Orleans in Louisiana who took part in the experiment. Heat from the walls raises the liquid pressure, instantaneously compressing and heating the gas in the bubbles. Because the fluid is so squishy, the piston effect heats the gas much faster than diffusion can carry energy back to the liquid, and the bubble temperature overshoots both that of the wall and the liquid.

Don't worry, the experiment didn't really break the second law of thermodynamics. Strictly speaking, the law applies not to changes in temperature but to changes in entropy--a related but different property. What's more, it applies only to systems in thermodynamic equilibrium. Inside the cell on board Mir, conditions were changing so fast that they left equilibrium behind. Only temporarily, however: After about 2 minutes, the second law reasserted itself, and the bubble cooled back down to the same temperature as the wall and the fluid.