Laws of thermodynamics

Post Reply New Topic

There is no absolute deductive proof for them. However they are very well supported empirically and they have never been falsified. In the domain of scientific hypotheses it does not get better than that.

ruveyn

Yes! That's the little devil.

I had a thought experiment when I was a physics student 30 years ago but at the time I knew nothing about Maxwell's demon. I'd love to know where my theory breaks - assuming it does. The closest thing I've come across is Maxwell's demon, but can someone tell me exactly how this would apply in the following thought experiment:

Just to set the scene, remember Brownian motion? Air molecules knocking little dust particles around due to the very high speeds of the air molecules (typically around 700 mph at room temperature). The little dust particle just bounces around (the random walk) but just think if this could be scaled up so a larger particle could be made to move. Unfortunately this can't happen because the forces tend to cancel out on larger objects and the net momentum imparted to them is negligible.

What is needed is a way to create a difference in the number of collisions on one side of the larger object so that a net momentum can be imparted, but this needs to be done somehow without needing to do expend more energy doing it than would be gained (as kinetic energy) by the target.

This may be the solution: Consider if you had a very thin but strong membrane. Perhaps a polymer sheet. In this polymer are small holes slightly larger than that of the air molecules or other gas used (Argon would be good because it is inert and single atoms rather than lumpy molecules such as O2 or N2).

Covering those holes are small "doors". The doors consist simply of a side chain of polymer consisting of only a few atoms. The doors are only on one side of the polymer sheet. The doors can only swing open one way - the same as a typical house door. (This is because of the way the door is attached to the sheet and the fact the door is slightly larger than the hole). A one way flap.

Normally the doors have a tendency to swing shut, this tendency can be achieved by having atoms that are attracted to the "door frame" atoms, in a similar way to that in which protein molecules fold to a preferred 3 dimensional shape. This could be achieved using atoms that would be electrostatically attracted to one another but not strong enough to form chemical bonds. However, this force holding the doors shut is minimal by design.

Now consider what happens if a fast gas molecule hits the door. If it hits on one side it will knock the door open and pass through. It may lose a little momentum to the door / membrane but not much as it continues in its forward direction. However, if a gas molecule hits the other side of the door it will simply knock the door against the door frame and the gas molecule will rebound. In rebounding it imparts momentum to the door / door frame / membrane.

If you do the maths on this you will find that at room temperature and pressure that such a membrane with only say 1% coverage with such doors will pick up a substantial amount of momentum from the gas rebounds.

So what? You may think.

Well consider arranging three such membranes ate 120 degrees to each other and attached to an axle. You have a turbine. In principle the turbine will turn and is capable of doing work.

This in NOT a perpetual motion machine because there is no net gain of energy. What would actually be happening is the gas molecules would lose momentum as the turbine gained momentum and consequently the gas would cool.
This is contrary to the second law of thermodynamics. Useful work would be obtained from background heat.
The "loophole" that my thought experiment uses to try to cheat is that of random motion. The backbone of Thermodynamics is that of random motion of particles and a statistical treatment of such particles. However, if you deviate from the randomness (with the one way doors) then the principle may well break down.

Anyone understand what I'm saying? Would this work? Maxwell must have had a similar thought experiment to me to come up with his "Demon" but it isn't actually proof it would not work.

In my mind I can see the process, see the atoms bouncing from one side and passing straight through from the other. Why would this not work? What am I missing (if anything)?

I can think of two environmens for your system. A closed system, where the vanes are enclosed in a static chamber of gas, and the open system, where we have an infinite volume of gas. As the later is really just a limit case of the former, we really only need to examine what happens in that former case.

Your trapdoors will selectively allow gas to pass one way, but not the other, giving rise to a force on the vane. However, in that process, the rebounding gas itself will mean that the velocity distribution within the gas changes. It will result in there being more molecules travelling towards the trapdoors from the side where they open. After an initial "kick", I see an equilibrium establishing itself, where the excess bounced molecules are forcing trapdoors to open more often. Although this individually causes less momentum transfer, there are more impacts. The net effect will make the apparent pressure on each side of the vanes equal.

You may be right. It would be nice to get a definitive answer though. I'm half tempted to write a computer simulation but my maths and physics are a bit rusty now and I'm not sure I'd make the model 100% realistic. It would also need a fairly hefty computer to run such a simulation, at least with a visual presentation. The last time I wrote and ran anything like that was on a Cray supercomputer in Fortran and I got told off for slowing it down I was then introduced to the concept of batch processing Happy days.

I had a look in a book by Richard Feynmann, he describes a similar thing with a ratchet and pawl. Instead of pushing a door open, the molecules would push a vane attached to a ratchet, and a pawl on a spring would ensure that the 'turbine' only moved one way. Each collision on the 'clockwise side' would move the system, those on the 'anticlockwise side' wouldn't. This way we'd extract motion from a gas at uniform temperature, breaking the 2nd law.
The catch is in the mechanism that regulates the one-way motion, whether it be your atomic doors or this imaginary ratchet/pawl system. If the objects are perfectly elastic, so no energy is dissipated in their motion, then the door/pawl will bounce open and closed. In this case, molecules/ratchet teeth are as likely to pass through in either direction. If the motion is damped somehow, so that the door stays closed or the pawl stays down, then this heats up the ratchet+pawl/door+frame increasing their thermal motion and the pawl/door will flap open and closed from the increased vibration.
That's roughly the idea, he goes into a great deal of detail and uses a fair bit of maths to prove his point, but the problem is always the same as Maxwell's Demon; the process of seperating the particles always uses more heat than you extract by seperation. The book is brilliant by the way; The Feynmann Lectures Vol 1-3.

Thanks Theory foiled again I'll have to be more creative

I read a /. article sometime within the last two years about a nanoscale structure functioning similarly to Maxwell's Demon.

Here is a Wikipedia blurb about it:

It can not be Maxwell's Demon becuase it was supplied with light, I think that the synthesis of a Maxwell's demon like molecular device is a challenge which might be impossible to overcome.