Heat engines are machines that take heat from a hot source, transform part of it in usable energy (work),and dump the rest to a cold source. Their size can range from the nano-scale (e.g. a single tr apped ion engines), to tens of meter and beyond (e.g. a geothermal power plant), but all share the same limitations. The efficiency of conversion of heat into work cannot overcome a limit (found by Sadi Carnot in 1824), that depends only on the temperatures of the two sources.

Common knowledge is that the maximal efficiency (called Carnot efficiency) can be reached only at zero power output, but none of the known laws of thermodynamics seems to forbid this. In a research just published in Nature Communications, Michele Campisi and Rosario Fazio take a close look at this problem and find that it is in fact possible in principle to get closer and closer to Carnot efficiency without giving up power. According to their analysis this should happen when the working fluid inside the engine, exchanging heat with the reservoirs, is close to a phase transition.

The new mechanism  might help designing more efficient and more powerful energy conversion devices, e.g. in solid state nanotechnology.