Make your own free website on

Molecular valve

We don't need demon.
Following doc. is proposal for making one experiment related to second law of thermodynamic. I don't want to defy or violate that law, I want to prove that it is mistake. By second law there is no flow of gas from area with lower pressure to area with higher pressure. Entropy will always increase in isolated system.

Above picture is molecular valve. To work, it must be very low size, smaller hole should have diameter in range of tens of nanometers. Experiment should work with much bigger holes, but then we need to lower gas density (pressure). Whole principle is based on statistic.
Picture is just small part of wall between two gas chambers with same gas, and at start at same pressure. Hole has cone shape, so if diameter of left side hole is 5 times bigger than right hole's one, area is 25 times bigger. By statistic, there will 25 times more gas particle (molecule) enter left hole than right one. Of course not all will pass through to right side, because it depends much from angle. Dash lines on picture show aprox. range of direction of moleculas which will likely pass on other side. Angle at right side is much bigger, it's obvious when we calculate angles of repulsion. But, there will still much more molecule pass from left to right than reverse because of much bigger area of hole.
Why it works not by bigger (easy makeable) holes? Because then collisions between gas moleculas will prevail, and not 'direction effect' of hole.

Depending on number of holes and much many parameter gas will gradually move to right chamber until point of balance. Ratio of pressures at that balance depends mostly of number and efficiency of holes (how small and accurate are they). But any small, reliable measurable ratio above 1 will mean that second law is false.

About realisation of wall filled with molecular valves:
I think that it is possible to make today such wall. I don't know much about nano-technology, this is probably the righ direction. Microchip manufacturing operates at desired sizes currently, so maybe following is realizable:
Cone can be achieved by several layers with gradually bigger hole. We can make millions of holes on 1 square cm area now. Main problem is that there is substrat under every silicon slice. But we need no substrat, but opened space there (this should be right side of picture). With some additional processes it could be possible to make some carrier of tiny valve wall. Normally it will be on right side, because there is more space (smaller holes).

Above picture shows idea of principle for realisation. Holes in right side dicker wall are in shape of circle, while attachment (horizontal on pic) elements form square grid. This is very rough idea, and could be improved certainly.

Some thoughts about Leo Szilard's 'prove' why Maxwell's Demon not denies second law: I think that whole explanation with information needed for demon to conduct moleculas is totally wrong. I don't see any simple relation between energy and information.

Finally, if this idea of molecular valve will show itself as correct, we will certainly ask, can we make something useful with that? To make some significant gas flow from lower pressure to higher we will need large area what will cost very much money to produce. Will nanotechnology able to produce it enough cheap in near future? How stable will be such ultratiny wall?
I think that you know how to close circle, and make movement...

This document is written on July 9, year 2004. Petar Putnik