The following is an item by Harold Aspden which appeared in the journal 'Nature' on September 6, 1990 at p. 25 of vol. 347.
Breaching the Second law of Thermodynamics
SIR - Maddox (Editor of Nature), in his discussion of the possibility of breaching the second law of thermodynamics (Nature 345, 109; 1990) raises an issue which deserves more than cursory attention. If Maxwell's demon could act effortlessly to allow more energetic particles to pass from one cavity A to another B, while restricting flow from B to A to less energetic particles, then one could transfer heat up a temperature gradient to cause B to be hotter than A. The question is important because our search for a pollution-free energy source would soon be solved for ever if an army of Maxwell's demons could be put to such use on a commercial scale.
Because we regard photons as carriers of energy and see them as a kind of particle travelling at the speed of light, we hardly need rely on the intelligence of Maxwell's demons to open and close a shutter across a hole between the two cavities. Instead it is sufficient to place a convex mirror located in cavity A and positioned away from but facing the hole. The focusing action of this mirror will ensure that those photons transfer heat from A to B until the walls of cavity B are at a higher temperature than the walls of cavity A.
This should suffice as an apparatus which will breach the second law of thermodynamics. Textbook teaching declares that "In nature heat is never found to proceed up a temperature gradient of its own accord". From this, the textbooks advance to a statement of the law according to which it is impossible for any machine to abstract heat from the coldest body of its surroundings and convert this into useful work surplus to that needed to power the machine. Maxwell's demon can affect that self-accord of the heat transfer and this causes one to wonder if that demon does really do any work in controlling the shutter. But with the mirror discharging this physical task, it is assuredly not doing work itself and those photons proceed up that temperature gradient by their own accord as they bounce from the mirror surface. In principle, therefore, there just has to be a failure of the second law of thermodynamics.
Maddox, in his editorial on mechanical engines driven by light (Nature 342, 13; 1989), suggested that "it would be more than just fun if somebody were to build one". The practical implications are enormous but they highlight the need to develop a miniature thermoelectric power converter which could be incorporated between cavities A and B in a stratified system containing a large superficial heat radiating surface with numerous cavities and numerous mirror focusing elements so that a significant net power per unit volume can be generated.
What is so fascinating about this proposal is that such an energy device would not be subject to the Carnot efficiency limit. All the heat fed into the system to sustain the temperature of cavity A would emerge as electricity even though it might cycle several times between cavity A and cavity B, going one way thanks to the focusing power of the mirror and the other way as heat 'loss' through the therrnoelectric converters. The efficiency of the heat-to-electricity conversion of the thermoelectric power converter is not a factor limiting what has just been said. It is just that the greater this efficiency, the smaller the volume of the structure needed for a given power output and so the smaller the capital expense incurred.
Department of Electrical Engineering,
University of Southampton,
Southampton SO9 5HN, UK
* Note: The above address is no longer applicable, the author having now retired. Mail to the author should be sent c/o Sabberton Publications, P.O. Box 35, Southampton SO16 7RB, England.