## Entangled Particles Break Classical Law of Thermodynamics, Say Physicists 222

Posted
by
Soulskill

from the not-in-homer's-house dept.

from the not-in-homer's-house dept.

New submitter Zex_Suik writes

*"Japanese physicists have used one of Maxwell's thought experiments and the ability to turn information into energy to extract more energy from an entangled system than should be possible according to the laws of thermodynamics (abstract). From the article: 'Imagine two boxes of particles with trap door between them. You want to use the trap door to guide the faster particles into one box and the slower particles into the other. In a classical experiment you would have to measure the particles in both boxes to do this experiment. But things are different if the particles in one box are entangled with the particles in the other. In that case, measurements on the particles in one box give you info about both sets of particles. In essence, you're getting information for nothing. And since you can convert that information into energy, there is clear advantage when entanglement plays a role. That's hugely significant. It means that the laws of thermodynamics depend not only on classical phenomenon and information but on quantum effects too.'"*
## Re:Fail (Score:5, Interesting)

No information is gained, for the same reason that separating entangled particles by a great distance and then measuring one doesn't result in information traveling faster than the speed of light.

This is like saying putting a red ball in one bag and putting a blue ball in an identical bag, then shuffling the bags around, then looking in one bag gives you free information about the other bag. It doesn't.

Not quite. The latter scenario is affectionately called "Bertlmann's socks"; once you separate the bag, it's true that one has the red ball and one has the blue ball but we don't know which until we look.

However, with a pair of entangled particles of spin state (up + down) for example, it's not the case that one is up and one is down. If you measure one particle in the "east" direction and find that it is pointing east, then the other one will be found to be pointing "west". It's been proven (Bell's inequalities etc.) that there is no possible "hidden state" that would account for the fact that the two measurements can be taken in arbitrary directions and still correlate.

## Re:Fail (Score:4, Interesting)

"that there is no possible "hidden state" that would account for the fact that the two measurements can be taken in arbitrary directions and still correlate."

Not quite. Bell's theorem, and the experiments inspired by it, suggest that any classical theory (or hidden variable theory) would have to be non-local. The non-locality can be quite mild though.

Also, both the results of the experiments that show Bell's inequality is violated, and the theorem itself, are being challenged.

## Re:Soooo (Score:5, Interesting)

According to my understanding of the article (IANAP), this has nothing to do with memory, and use of memory would not impact the system in any significant way in any case (the initial energy required to take the measurements to store into memory would offset the reduction in entropy during the experiment).

The fundamental issue with the classical scenario of Maxwell's Demon is that in order to know if/when to open/close the gate you need to measure each particle in the system at least once. The number of measurements >= The number of particles. The basic implication is that you introduce entropy via taking measurements at least as much as you reduce it via segregating particles according to energy differential.

If you consider quantum entanglement, however, the rule that number of measurement >= the number of particles is no longer necessarily true. E.g., if each particle in the system is entangled with another particle in the system, the number of measurements could be as low as 1/2 the number of particles since one measurement gives you information about both of the paired particles. It is also possible for more than 2 particles to be entangled, so to generalize, you could have N-way entanglement between sets of particles in the system, and the minimum number of measurements becomes number of particles / N.

The fundamental question I have is if it's possible to determine entanglement relationships between particles in the system for less energy than independently measuring each particle. If not, then you offset the entropy reduction of only measuring one particle from each entangled set by the energy required to identify entanglement relationships.

## Re:Energy from information? (Score:5, Interesting)

Interesting question. I used to ask a related question, "How much does a bit weigh?" I learned a couple of years ago that the proper question is, "What is the area of a bit?" See the Holographic Principle [wikipedia.org], and/or an article in Scientific American two or three years ago. It has to do with the requirement that the Universe can never lose, but must always gain, entropy. When mass is sucked into a black hole, the entropy of the Universe would lose entropy, so the entropy must be left behind at the event horizon. This somehow forces the surface area of the event horizon to expand according to the mass of the black hole. Since mass entropy can be equated to information entropy, after some shenanigans I don't understand, it turns out that the area of one bit is 2x2 planck lengths.

But I suspect, since that area is related to the mass that has been sucked in, wouldn't that imply that one bit is related to that amount of mass? Which means it is related to that mass, or equivalently that energy. :D I don't think that means that the mass 'represents' one bit though - rather the opposite, one bit represents that amount of mass or energy.

## Re:Any cost to entangled particles ? (Score:5, Interesting)

So, in order to get particles that are already in the "entangled" state, something must have happened to ordinary particles, first, right?

If so, what's the cost (in term of energy) to get originally un-entangled particles to be "entangled"?

Assuming, e.g., a photon decays into two entangled photons, there is not really an energy cost associated with the fact that those two photons are entangled. There are different methods of producing entangled photons. One way is by passing a higher energy photon through a special crystal (see here [wikipedia.org], but the conversion efficiency is extraordinarily low, so you'll have to spend a lot of extra energy generating unentangled photos. Another way is to trap an electron and wait until it decays into photons. Again, no extra energy required to get the entangled photons from the electron, but you do expend energy getting and holding the electron in the trap.

My impression is that "entanglement" occurs for free, but verifying that you have entangled particles is always going to cost some energy. The neat thing here is that you can perform the verification before you put the particles in the boxes, so the information you have on the particles is kind of like a battery storage in terms of how it can be re-extracted as energy.