CSI and Maxwell’s Demon

scordova

So no, I won’t take what I said, back and I leave it to the readers to decide who has stated their case better based on physics. If they believe you, fine, but I’ve stated my reasons.

You couldn't state your reasons better. I bet rhampton7 is an evolutionist. Evolutionists are used to get more from less, no wonder. Usually they do that spectacularly with Darwin's demons, but we see here rhampton7 tries also with Maxwell's demons, maybe hoping that physics helps evolution. Unfortunately all hard sciences have the bad habit to destroy the dreams and illusions of evolutionism. niwrad

But it seems very likely that an MD does not require “more energy.”

No Rhampton, you're not ever reading the passage you cited:

There has emerged a kind of consensus, based largely on the works of Landauer and Bennett, and independently Penrose, according to which a mechanical demon can indeed deliver work by rectifying fluctuations, but in doing so it gathers information that must be written to physical memory. The eventual erasure of this information carries a thermodynamic cost, no less than kBT ln 2 per bit (Landauer’s principle), which eliminates any gains obtained from the recti cation of fluctuations…

Without erasing the memory, you essentially have an irreversible process, so your interpretation is all messed up. I've referred to Landauer's principle here at UD or TSZ somewhere... Without erasing memory, you run out of memory, so in essence, the unusable memory storage provides a finite limit on how much work the demon can do, and thus it needs more energy than it can create because it runs out of information storage (which in this case is a proxy for energy). And to erase the memory it has to expend energy! Further, by taking on the assumption that you have a memory bank that has yet to be erased, you've subtly said: 1. the system wasn't really closed, it is open 2. the system is not in equilibrium 3. the system is both open and not in equilibrium

but in order to lift the mass the device requires a memory register to which it can write information.

And writing information will eventually cost energy (if not immediately, at time of erasure). But this is little different than saying I have a battery that will power the imp, but the battery will run out of energy unless it is recharged. The non-erasure of memory storage is like an uncharged battery. Lifting things, by the way, with minimal energy is no big deal, you don't even need a memory storage device to do it. For example, to lift a 10,000 kg object 1 meter, requires 98,100 joules of energy, but how can a human do this expending a few joules of energy? Simple, have a pulley system that has an object slightly more than 10,000 kg on the other side. You won't need a lot of added mechanical energy on the part of the human to perform the task. So the autonomous lifting of objects is not really a big deal is it? Now if your supposed MD is writing and erasing information this is no more than the MD acting like a frictionless pendulum in a vacuum, except in this case the gas molecules are swinging from one container to the other -- not very interesting is it? You can accomplish this by having just a few molecules (say 3 or 4) in a chamber and an open gate. The molecules will spontaneously aggregate on one side of the chamber vs. the other (somewhat like a chaotic pendulum). So what you describe is no big deal and can be accomplished with a lot less effort. The 2nd Law is appropriate to large scale statistics, not sparse gas examples. You don't need demons to get spontaneous aggregation of molecules from one chamber to another, just sparse gases. This has been well known since Boltzmann's time, but the issue is for systems with many particles. By the way, a liter of gas has on the order of 2.78 x 10^22 molecules based on Avogadro's number. Good luck building a computer memory storage device that will process 2.78 x 10^22 bits just to manage a quantity of gas molecules that will fill 1 litre bottle! If I made the generous assumption it takes .01 watt of power to maintain 100 gigs of memory, I estimate it will take 347,656,250 watts of power to manage your imp device for a mere liter of air using current technology. I can do better with a simple air compressor connected to my household outlet! :-) And its a moot point anyway because your citation doesn't refute Maxwell's Demon. All this to point out, what you cite might not have the effect on technology that you might suppose, and also, you can do the same without a demon, you just need a sparse gas. So no, I won't take what I said, back and I leave it to the readers to decide who has stated their case better based on physics. If they believe you, fine, but I've stated my reasons. scordova

scordova, The problem is that you made those three statements under the incorrect assumption that they applied to Maxwell's Demon. But it seems very likely that an MD does not require "more energy." Unlike a perpetual motion machine, it is not impossible for an MD to exist. Furthermore, an MD can perform logically irreversible operations such as AND, OR and erasure as well as encode information. At the very you would want to restate that the 'free' information available to an MD is the dissipated energy in a given system. rhampton7

Since I brought up the topic "Brownian windmill" consider wind generators that exist at the macroscopic scale. You'll see plenty of them all around the USA. Energy isn't being created, what is happening we are converting the kinetic energy in the breeze into electrical energy. The source of the electrical energy is the wind energy, and the source of the wind energy is the sun. Now, gas in a container still has kinetic energy, and if you could build a windmills small enough you could in principle tap into that energy, but that would be merely converting the kinetic energy into electrical energy and the exchange of momentum as the gas molecule hits the windmill "blade" will cause the gas molecule to slow down, and thus heat energy (kinentic energy) is converted into electrical energy and the gas cools. But this is not perpetual motion nor creation of energy anymore than WindGenerators of today, the ultimate source of the energy is the heat energy of the sun -- thus, such a Brownian windmill is nothing more than a glorified solar energy converter. scordova

rhampton, I'm not trying to poo-poo your idea, actually it is really smart, but what if such a device works, what is happening is the thermodynamic boundaries are being redrawn. Let's suppose the demon is able to get pressure on one side. How can you now use this energy? If you attach a generator to the system, the kinetic energy of the gas molecules will now be converted to the electrical energy that flows out of the original gas system. Eventually, the demon will deplete the heat energy out of the gas, and like a battery it will go dead. Where such a device will obviously have advantage is in an OPEN system where the heat can be replenished easily from the environment. But Szilard analysis was for a closed system. I've often thought about the notion of "Brownian windmills" and I'm not unfamiliar with the research in nano-engineering on the topic, but I think you're mischaracterizing the significance of Maxwell's demon. I hope these devices work, but they will not prove: 1. 2nd law is wrong 2. perpetual motion machines work 3. Darwinian evolution is possible in principle. Best of luck to these guys! Thank you for your comments. What you said needs to be discussed. Sal scordova

1. that information is inaccessibile from the environment

Most definitely not in the case of complex systems.

3. that, like a perpetual motion machine, it is impossible

You want to advocate perpetual motion machines, you are more than welcome to. :-) At best, the kinetic energy in the system will be reconfigured, more kinetic energy will NOT be created (1st law of thermodynamics). So no thank you, I will not take that back.

2. that more energy is needed to power the demon than the demon can create

At best, the kinetic energy in the system will be reconfigured, more kinetic energy will NOT be created (1st law of thermodynamics). So no thank you, I will not take that back. At issue with the 2nd law is USABLE energy, not energy that exists to begin with (that is the 1st law of thermodynamics). scordova

Maybe very soon, as others are proposing to test Jarzynski's work with quantum dots; "Granted, they haven’t actually built such an experiment, but the researchers are optimistic that it should be possible. So Maxwell’s Demon need not be all that smart, or even sentient — just very well designed." However it should be noted that you made some specific claims that you may want to walk back: 1. that information is inaccessibile from the environment 2. that more energy is needed to power the demon than the demon can create 3. that, like a perpetual motion machine, it is impossible rhampton7

Thanks for the info rhampton! I hope an intelligently designed device will be able to succeed where Maxwell's demon failed. scordova

See also: Maxwell’s Refrigerator: An Exactly Solvable Model Unified Jarzynski and Sagawa-Ueda relations for Maxwell's demon Maxwell’s demon gets scribbling Truly work-like work extraction via a single-shot analysis Prof. Jukka Pekola: Thermodynamics and Maxwell Demon in an Electronic Circuit Thermodynamics of a Physical Model Implementing a Maxwell Demon rhampton7

"Since just a free-lunch Maxwell demon is refuted by thermodynamics in principle" Then again... Work and information processing in a solvable model of Maxwell's demon, Dibyendu Mandal and Christopher Jarzynski (2012) A system in thermal equilibrium undergoes random microscopic fluctuations, and it is tempting to speculate that an ingeniously designed device could deliver useful work by rectifying these fluctuations. The suspicion that this would violate the second law of thermodynamics has inspired nearly 150 years of provocative thought experiments, leading to discussions of the thermodynamic implications of information processing. Although both Maxwell and Szilard famously took the rectifying agent to be an intelligent being, later analyses have explored the feasibility of a fully mechanical "demon". There has emerged a kind of consensus, based largely on the works of Landauer and Bennett, and independently Penrose, according to which a mechanical demon can indeed deliver work by rectifying fluctuations, but in doing so it gathers information that must be written to physical memory. The eventual erasure of this information carries a thermodynamic cost, no less than kBT ln 2 per bit (Landauer's principle), which eliminates any gains obtained from the rectication of fluctuations... Experiments have been performed with the explicit aim of testing theoretical predictions, including Landauer's principle. Moreover the consensus or "favored explanation" described above is widely but not universally accepted, as suspicions persist that it assigns an unwarranted thermodynamic significance to random data... In this paper we propose an explicit, solvable model of a system that behaves as a Maxwell demon. Our device, which extracts energy from a single thermal reservoir and delivers it to raise a mass against gravity, is fully autonomous -- it is neither manipulated by an external agent nor driven by an explicit thermodynamic force -- but in order to lift the mass the device requires a memory register to which it can write information. rhampton7

far more complex in the former than in the latter.

If DNA and other features of the cell and even human nervous system involve quantum entanglement and spooky action at a distance, then evolving a cell has definitely far more information processing aspects than would be required of Maxwell's demon. scordova

Conservation of information proves thermodynamics, not the other way around, imho. Thermal information in classical thermodynamics is merely information about position and momentum of the particles, whereas organizational information (such as found in a biological organism) involves far more than position and momentum of particles. The kind of information in a biological cell seems to me far richer than the information of interest for Maxwell's demon. This is just a qualitative opinion on my part however, nothing I'm willing to defend yet rigorously.

far more complex in the former than in the latter.

I'm sympathetic to saying that, but you know me, I won't swear by it until I can work out the details of a proof. :-) scordova

Conceptually a Darwin's demon is a Maxwell's demon. Mainly what changes is the amount of information processing, far more complex in the former than in the latter. Evolution example: a Darwin's demon must construct a protein in the right box by selecting amino acids from the left box through the hole in the wall. Since just a free-lunch Maxwell demon is refuted by thermodynamics in principle, a fortiori is refuted a free-lunch Darwin's demon, whose work is far more complex. Clear, no? niwrad