Superluminal communication through entanglement

[bascule]

I've been through this before, and despite all the physics experts that seem to be around I never received a satisfactory answer. So someone who actually knows the math behind this and can help out someone venturing into quantum from a purely conceptual perspective, I'd really appreciate your help in pointing out what's wrong with my reasoning here because, as far as everything I've ever read from physics has told me, I must be wrong...

 

Okay, we have Einstein saying that you can't have anything causal occur at superluminal speeds... state changes within the universe propagate at c and that's all there is to it...

 

Yet we have ESR "spooky" action which allows certain things to happen non-locally, quantum entanglement in which the same events can occur "simultaneously" across distances which it would take light (in a vacuum) a considerable amount of time (interpret that as how you see fit) to traverse.

 

We have the delayed choice quantum eraser experiment demonstrating that collapsing the probability wave of an entangled particle simultaneously collapses the probability wave of the particle it's entangled with.

 

Okay, let me say that again, and if it's not correct, please let me know why, because that statement is the crux of this entire question:

 

Collapsing the probability wave of an entangled particle simultaneously collapses the probability wave of the particle it's entangled with.

 

If this is so, do we not then have superluminal causality through ESR spooky action? I'm aware that there is an occurance of spontaneous collapse of an otherwise uncollapsed probability wave, but as far as I'm aware it is not possible for the opposite to occur, namely a collapsed probability wave spontaneously reverting to uncollapsed.

 

So if this is the case, shouldn't statistical analysis of the behavior of an entangled particle on one "side" reveal whether or not the other "side" has chosen to collapse the probability wave or not? Can we not sample for awhile and discover that the probability wave was uncollapsed a whole bunch of times, leading us to the conclusion that the other "side" was sending the "uncollapsed" signal? Similarly, if we sample over a window in which we expect the probability wave to remain uncollapsed at least a few times and observe only the collpased probability wave, can we not infer that the other side is sending the "uncollapsed" signal? Sure, that inference may be lossy, but computer science has been dealing with sending signals reliably over lossy media for decades, and could they not work out reliable communication over such a medium?

 

I detailed my idea of how this could actually work based on the delayed choice quantum eraser experiment, and never received the debunking I had hoped for.

 

Please, I'm crazy, I'm saying that causation can occur at superluminal speeds! Debunk me!

[Zero Wing]

Although i'm not sure exactly what you're asking, this is something i believed could happen. Are you asking whether or not it's possible to take full control over the system by manipulating it's amplitude, or is it something else? Cause that's something i'd like to know..

 

Is it possible to 'create' your own unique entangled system for this purpose?

[swansont]

Collapsing the probability wave of an entangled particle simultaneously collapses the probability wave of the particle it's entangled with.

 

If this is so' date=' do we not then have superluminal causality through ESR spooky action?[/quote']

 

But you are the only one who know the state of the particles when you do whatever measurement collapses the wave function. You can't convey that information to anyone else faster than c. So causality is not violated.

[bascule]

But you are the only one who know the state of the particles when you do whatever measurement collapses the wave function. You can't convey that information to anyone else faster than c. So causality is not violated.

 

Okay, getting back to the actual setup I was trying to detail:

 

You have a down converter which spits out a pair of entangled photon beams (a photon at a time at very small intervals), and you cause these to "self-interfere" ala the double slit experiment by providing two possible paths for the photons to take and making the two paths cross. Let's call one of the two entangled photon beams A and the other B.

 

At B we have a CCD or some other device that registers the locations where individual photons land. When the wave function for A is not collapsed, the photons at B trace out the expected interference pattern on the CCD. However, if the wave function is collapsed before the corresponding entangled photon strikes the CCD at B, the wave function collapses and the interference pattern is destroyed.

 

So if you use a few trillion photons to send each "bit", couldn't you tell if beam A's probability wave was collapsed by whether or not it traces out a corresponding interference pattern at B? Wouldn't the presence or absence of an interference pattern at B let you know if the probability wave was being collapsed at A?

[swansont]

I don't see how the "which-path" information has anything to do with the entanglement, unless the path information is measured by a polarizer. Then you have to convey how the polarizer was set up to get the measurement, which has to be transmitted via a classical communication channel.

 

It's always going to boil down to a similar situation. The trick may be in identifying the detail, but its going to be there somewhere.

 

And, though I don't wish to be curt, I need to say this: I'm not going to get into a scenario of "why doesn't this setup work?" Intellectually it's "topologically" the same as explaining (yet again) why a perpetual motion machine doesn't work, or why one hasn't demonstrated that relativity is false. To me it gets rather tedious and uninteresting. Perhaps someone else is interested in that, but I choose not to spend my limited time here on problems of that nature.

[bascule]

Okay back up for a second here... is this statement incorrect?

 

"Collapsing the probability wave of an entangled particle simultaneously collapses the probability wave of the particle it's entangled with."

[Zero Wing]

Okay back up for a second here... is this statement incorrect?

 

"Collapsing the probability wave of an entangled particle simultaneously collapses the probability wave of the particle it's entangled with."

Yes' date=' and knowing that, what is the exact question you're getting at? [b']What is it about entanglement that you want to know?[/b] Whether or not you can use the up/down spins to replace our current have/have-not communications in order to achieve long-range communication? If so, i believe that's very possible, but if not, please detail the question concisely. Not trying to be pushy, just want us all to be on the same page.

[5614]

Okay back up for a second here... is this statement incorrect?

 

"Collapsing the probability wave of an entangled particle simultaneously collapses the probability wave of the particle it's entangled with."

You know the particle's spin yes, but your friend who is miles away with the other particle doesn't know what you've got, until you tell him (using a classical channel of communication, so it's limited by c) he doesn't know what his particle's spin is.

[J.C.MacSwell]

You know the particle's spin yes, but your friend who is miles away with the other particle doesn't know what you've got, until you tell him (using a classical channel of communication, so it's limited by c) he doesn't know what his particle's spin is.

 

If he knew how you would test the particles (need many) could he use statistical information gained from testing the remote particles to ascertain whether you had run the tests? (regardless of the results, just could he tell that the tests had taken place?)

[bascule]

I'm not sure why everyone is focusing on spin... that would be the Aspect experiment as opposed to the delayed choice quantum eraser...

 

I guess I didn't make it clear enough that I was talking in using delayed choice to send information, not spin...

 

In a delayed choice experiment, the interference pattern is selectively destroyed depending on whether or not the other side observes the entangled particle (and thus collapses the probability wave), no?

 

Why can't this be used to send information?

[Zero Wing]

In a delayed choice experiment' date=' the interference pattern is selectively destroyed depending on whether or not the other side observes the entangled particle (and thus collapses the probability wave), no?

 

Why can't this be used to send information?[/quote']

Can you explain "delayed choice"? Not familiar with that one. As for it working or not, from what i can tell it doesn't seem as practical, but i don't have all the marbles yet, so please fill in.

[J.C.MacSwell]

I'm not sure why everyone is focusing on spin... that would be the Aspect experiment as opposed to the delayed choice quantum eraser...

 

I guess I didn't make it clear enough that I was talking in using delayed choice to send information' date=' not spin...

 

In a delayed choice experiment, the interference pattern is selectively destroyed depending on whether or not the other side observes the entangled particle (and thus collapses the probability wave), no?

 

Why can't this be used to send information?[/quote']

 

By forcing a particle to "decide" on a slit (say slit "A") you cannot correlate it's remote "twin" to remote slit "B", as they are not "quantum opposites" whereas certain spin properties are.

 

Another way of looking at this is if you tested any particle to see which slit it went through and it continued "through" another set of slits without checking testing it would still result in (add to) an interference pattern. (over many tests)

 

Hopefully Swansont might confirm or correct this!

[bascule]

Can you explain "delayed choice"? Not familiar with that one. As for it working or not, from what i can tell it doesn't seem as practical, but i don't have all the marbles yet, so please fill in.

 

Here's a highly technical writeup on the delayed choice quantum eraser experiment:

 

http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm

 

Of course quantum erasure doesn't come into play in this, merely the delayed choice aspect. Hopefully one of the real physics people here can give you a layman's description, as I wouldn't want to try to explain it incorrectly (obviously I'm misunderstanding something here, or I wouldn't be asking this question)

 

By forcing a particle to "decide" on a slit (say slit "A") you cannot correlate it's remote "twin" to remote slit "B", as they are not "quantum opposites" whereas certain spin properties are.

 

They don't have to. Guess I need to bring out my diagram again:

 

 

Okay, here we have a device at B involving a laser (which is magically spitting out single photons at a fixed interval) whose beam is passing through a beam splitter (BS), both outputs of which are passing through down converters (DC) which generate entangled photons.

 

These entangled photons travel along until they arrive at destinations A and C. At A, we have "which path detector" (WPD). At C, we have optics which cause the two beams to cross back over each other.

 

Now, (assuming the WPD is off) we know from the traditional double slit experiment that the probability waves of the two paths will interfere. Thus if the little blue bar at C is a CCD hooked to a computer which registers where each photon hits and builds a composite image, the photons will eventually trace out a pattern that looks somewhat like this:

 

 

(note: This image assumes that the probability wave will collapse due to outside interference about 50% of the time)

 

Now, if we were to switch on the WPD, the interference pattern would be destroyed, and we'd get a picture more like this:

 

 

Is this correct? Am I misunderstanding something? Because if so, it would seem that we could use these two very different looking pictures to differentiate between two different types of signals. Would this not be an example of using delayed choice to send information?

[Zero Wing]

Just found this on another site..

Q: Can we use entanglement for instantaneous communications?

No. Current theories suggest that the information passed between particles in an entangled system must remain a secret. However, experiments involving the decay of radioactive particles may hold promise for detecting when one half of an entangled pair has changed.

To illustrate, if an entangled photon meets a vertical polarizing filter (analagous to the fence in Figure 4.4) , the photon may or may not pass through. If it does, then its entangled partner will not because the instant that the first photon's polarization is known, the second photon's polarization will be the exact opposite.

So no go?

 

But can't you just backwards-interperate the oppisite state in order to make a successful communication?

[J.C.MacSwell]

This may be a case of "a bridge too far". There are two sets of entangled photons, each set created by different parts of the same probability function and we're asking if they will interfere with each other. If they will then we're asking if testing for the prescence or non presence of one of the sets, by testing for a result of it, can collapse the interference, which may have already manifested itself as a pattern before the decision to test or not was made. (that must be the "eraser" part).

 

Interesting set up even without the remote part. What experiments have been done? Can the offspring of two separate parts of a probability function interfere with one another?

 

Edit: when I say probability function I don't mean the math, just what it represents real or unreal (whatever that means)

[bascule]

This may be a case of "a bridge too far". There are two sets of entangled photons, each set created by different parts of the same probability function and we're asking if they will interfere with each other. If they will then we're asking if testing for the prescence or non presence of one of the sets, by testing for a result of it, can collapse the interference, which may have already manifested itself as a pattern before the decision to test or not was made. (that must be the "eraser" part).

 

In the above setup, if we observe the "which path" information for the entangled particle then we know which path its partner took, and consequently the probability waves for both particles are collapsed and the interference pattern is destroyed.

 

The twist of the quantum eraser experiments is that if the which path information is "erased", that is to say that other quantum events are used to destroy the which path information, the interference pattern reappears.

[5614]

If he knew how[/b'] you would test the particles (need many) could he use statistical information gained from testing the remote particles to ascertain whether you had run the tests? (regardless of the results, just could he tell that the tests had taken place?)

He couldn't know the actual data he would need to make this faster than c. He won't know if your particle is up or down spin unless you tell him (slower than or equal to c) and he wouldn't know the axis you use to measure the polarisation of a photon.

 

He might know how I would test the particle, that is, in the same way that I can read on the internet how to test a particle's spin, but that isn't going to help him as far as trying to make it faster than c is concerned.

 

No he cannot tell if you have run the tests or not.

[bascule]

It's always going to boil down to a similar situation. The trick may be in identifying the detail' date=' but its going to be there somewhere.

 

And, though I don't wish to be curt, I need to say this: I'm not going to get into a scenario of "why doesn't this setup work?" Intellectually it's "topologically" the same as explaining (yet again) why a perpetual motion machine doesn't work, or why one hasn't demonstrated that relativity is false. To me it gets rather tedious and uninteresting. Perhaps someone else is interested in that, but I choose not to spend my limited time here on problems of that nature.[/quote']

 

So you don't believe this question has any merit whatsoever? I'm sorry you feel that way, but this is a problem which I've confronted several of my physics student friends with, they have asked their professors, and no one has ever come up with the specific flaw which would render this impossible.

 

The best I've heard is "spontaneous probability wave collapse." If you were trying to use this over, say, interplanetary distances, I could definitely see it being possible that the probability waves of one of the two entangled particles would be collapsed through collisions with other particles floating around in space. But would this happen to every photon (entangled pair) every time, or would a certain percentage of them still get through? The figure assumes 50%, and that could be the fatal flaw. But is this really the case?

 

I think the real difference in this setup is that the information is revealed through trends observed in the behavior of innumerable photons rather than observing a specific property of a specific particle...

[swansont]

So you don't believe this question has any merit whatsoever? I'm sorry you feel that way' date=' but this is a problem which I've confronted several of my physics student friends with, they have asked their professors, and no one has ever come up with the specific flaw which would render this impossible.

 

The best I've heard is "spontaneous probability wave collapse." If you were trying to use this over, say, interplanetary distances, I could definitely see it being possible that the probability waves of one of the two entangled particles would be collapsed through collisions with other particles floating around in space. But would this happen to [i']every[/i] photon (entangled pair) every time, or would a certain percentage of them still get through? The figure assumes 50%, and that could be the fatal flaw. But is this really the case?

 

I think the real difference in this setup is that the information is revealed through trends observed in the behavior of innumerable photons rather than observing a specific property of a specific particle...

 

When you make the system complex, finding the flaw becomes tougher. I can't tell from your diagram what's going on - where are the slits, and why do you have two sets of DCs? The top photons don't appear to be entangled with the bottom ones. Are the pictures the results of an actual experiment (if so, what is the citation) or is this what you expect to see?

 

These are the kinds of things that I won't necessarily be able to take the time to try and figure out, and there's no gurantee that I can anyway.

[J.C.MacSwell]

He couldn't know the actual data he would need to make this faster than c. He won't know if your particle is up or down spin unless you tell him (slower than or equal to c) and he wouldn't know the axis you use to measure the polarisation of a photon.

 

He might know how I would test the particle' date=' that is, in the same way that I can read on the internet how to test a particle's spin, but that isn't going to help him as far as trying to make it faster than c is concerned.

 

No he cannot tell if you have run the tests or not.[/quote']

 

I agree with the first part of the bold but not the second as this has been agreed to prior to the "test or no test" time period. The actual results of the test are not relevant as they will arrive "too late" (cannot be faster than c).

 

But knowing how the test is to be done, knowing the axis that will be used, should there not be detectable a "statistical shift" in tests of the "remotes" when they are tested on a different axis? Isn't this how Bell's inequality is tested? (though so far not remotely/outside the "causality cone zone" of the first particles or photons)

[J.C.MacSwell]

So you don't believe this question has any merit whatsoever? .

 

For my part (I'm no swansont), I think it is good bit of thinking regardless of whether there is a flaw in your assumptions.

 

I see no flaw in your logic and I don't see any unnecessary complexity.

 

My first question on your assumptions would be on the "local" part of the setup. Has this been done and produced interference patterns?

 

I can understand Swansont's reluctance to look at all these "setups". Most of them are just "noise" and should be dismissed out of hand. Maybe he will look at it closer if noone else finds the "flaw" or possibly it merits an experiment.

 

Where did this set up come from?

[bascule]

Well, I found this:

 

http://www.madsci.org/posts/archives/2004-11/1099321986.Ph.r.html

 

Hi Bill, sorry about the slow response. This was a tough question, and it took me a while to figure out the right approach. The details of this paper still go a bit over my head---but I think I can clear up the causality question.

 

From your question you seem to understand ordinary quantum entanglement experiments. In these experiments, you can take two entangled particles and separate them by any distance---in practical terms, say, you could send two entangled photons down a kilometer-long fiber optic cable. A measurement done on one photon has the effect of "collapsing" the wavefunction of the other photon, instantaneously. The finite speed of light does not come into play. However, it turns out that these experiments cannot actually transmit any information, so you can't use them for faster-than-light communication. The "recipient" physicist can't tell the difference between random photons and "signal" photons, until he or she talks to the "sender" and correlates data from both stations. Check out Nick Herbert's book "Superluminal loopholes in physics" to learn more about this.

 

Now, consider the experiment I just described---two entangled photons, collapsing simultaneously a few kilometers apart---and suppose that you watch it happen from a spaceship moving at 1/2 of light speed. The wavefunction-collapse events which the (stationary) scientists would call "simultaneous" are not simultaneous when seen by the spaceship. The spaceship, depending on its trajectory, might see the "recipient" making measurements before the "sender" generates them. Since there is no information being exchanged, the spaceship should not see anything unusual going on. A real information transfer (at the speed of light or below) can never be seen as acausal, for any real spaceship (at the speed of light or below).

 

Since quantum entanglement experiments always involve this "wavefunction collapse" which seems to move faster than light, it will always run into odd behavior where "effects" happen before their "cause". If it doesn't look odd to a stationary observer, it will look odd to a moving one. This paper seems to be an experiment which manages to look odd to observers at rest---maybe it would look normal to someone moving very fast!! But, again, no information is transferred and causality is maintained.

 

It's pretty weird stuff, though, and I enjoyed thinking about it---thanks for asking! In retrospect, I guess we should not be surprised that such experiments are possible!

 

-Ben

 

Perhaps I'll take a look at that "Superluminal Loopholes in Physics" book to see what it says...

[5614]

But knowing how the test is to be done, knowing the axis that will be used, should there not be detectable a "statistical shift" in tests of the "remotes" when they are tested on a different axis? Isn't this how Bell's inequality is tested?

Remember that you are not forcing the photon onto your axis, the photon will come in at whatever angle it wants to, you can't fix an axis for it.

 

I'm not yet familiar with Bell's inequality to comment otherwise though, what I'm saying may be totaly wrong, not sure.

[J.C.MacSwell]

Remember that you are not forcing the photon onto your axis' date=' the photon will come in at whatever angle it wants to, you can't fix an axis for it.

 

I'm not yet familiar with Bell's inequality to comment otherwise though, what I'm saying may be totaly wrong, not sure.[/quote']

 

 

I don't think you don't force the photon onto your axis as it is already there in a superposition of states.

 

I think you do get to choose the test axis and it chooses the rotation (up or down, clockwise or counterclockwise etc. or some quantum version of what we are picturing), wrt that particular axiswhich will end the superposition of states for that particular axis and in so doing "partially collapse" the wave function of other interdependant axis's (sp?) in a way that can be measured experimentally/statistically as per Bell's Inequality.

[5614]

OK, so this post is guesswork on my behalf.

 

Wouldn't you have the photon coming towards you, let's just say it could either be oscilating up/down or left/right, thats the two polarisations, it's one or the other.

 

When it enters the measuring equipment it is one of those, now if you chose axis which didn't align with that then you'd get your light oscillating along the line y=x on your graph or whatever.

 

As for this Bell's Inequaltiy, I dunno, the wiki article seems a bit iffy, it says:

"Bell's theorem is often presented in the context of the foundations of quantum mechanics, since quantum mechanics predicts cases in which a Bell inequality is infringed. However, neither the formulation of Bell's theorem, nor its derivation, have anything to do with quantum mechanics."

 

Are those bold parts correct? QM doesn't break Bell's Theorem, that was the EPR Paradox which would if it were true... and it was specifaclly formulated to solve the EPR Paradox which has a lot to do with QM!