Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Quantum Entanglement and Communication

  1. Sep 27, 2005 #1
    Can quantumly entangled particles be used to send information? I've read that if they could, the information couldn't be sent at greater than c. However, the ability, if possible, to send information via this mechanism would allow information to be sent to another place even if the place is on the other side of a planet. Or for communication in space without the signal becoming weaker with distance. Am I onto something, or wrong yet again?
     
  2. jcsd
  3. Sep 27, 2005 #2

    selfAdjoint

    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    As entanglement is described by quantum mechanics, it is impossible to send information FTL using it. All you can do is confirm a greater than random correlation between the resullts after the fact, the comparisons being based on information transmitted slower than light.

    There are several threads on this subforum dealing with the technical aspects of this; I just want to affirm for you that quantum mechanics has a definite, coherent, and consistent answer to your question and that the entanglement experiments that have been done are consistent with that answer.
     
  4. Sep 28, 2005 #3
    Thanks for replying, but you replied to what I wasn't asking. I'm not asking if information can be sent FTL via this mechanism; I'm asking if information can be sent period? As in the same speed that electromagnetic radiation moves at, or slower?
     
  5. Sep 28, 2005 #4

    DrChinese

    User Avatar
    Science Advisor
    Gold Member

    Using entangled particles, you can't send information from one member of the pair to the other using their entangled state as the mechanism of information transfer.
     
  6. Sep 28, 2005 #5
    Ok, but why?

    If two particles are entangled, say one is spin up, the is other down. If you make the spin up one become spin down isn't the other one forced to become spin up? If the particle on your end is changing its states, and you're not making it do it, then you know the particle entangled with it is. If the particle entangled with it is changing states that means it's being made to. Thus couldn't some kind of code be send from one place to another?

    Or what about particle haves? They follow the movements of whatever the other half is doing right? No matter the distance between them. Surely this could be used to send data?
     
  7. Sep 28, 2005 #6

    DrChinese

    User Avatar
    Science Advisor
    Gold Member

    You are certainly welcome to try to send data that way, but it just won't work. :smile:

    You cannot make an entangled photon become spin up if it wasn't that way already. You can only force it to take on a specific state *relative* to a polarizer you position as you desire. (When I say "specific state" I mean: it is no longer in a superposition of states. This does not in any way imply you can control the actual result.) The result will be a 50-50 mix on your end. And therefore a 50-50 mix on the other end too.

    To send data, you need to force the photons on the other end into State A or State B at will so that the other person can decode your message. That won't happen. Instead, the person on the other end ALWAYS sees a random mixture of 1's and 0's which cannot be decoded locally.
     
  8. Sep 29, 2005 #7

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Look at my journal (click on the left of this post), as this has been addressed several times here, I tried to put the QM "proof" there.
    In summary:
    all things you can measure locally are determined by the LOCAL reduced density matrix, and this matrix doesn't change when one or another measurement is performed, or not, on the OTHER system. So locally, no result (average, probability....) changes.
    What DOES change is the correlations between the two subsystems, but you can only find that out when you bring the local measurements together through some classical information channel (a telephone, for instance).

    In other words, if I give you a set of photons which are all the "left part" of pairs of entangled photons, and I take the set of "right parts" with me on Betelgeuze, you can now do experiments on your set of photons any way you like, the results of these experiments will not depend on what I do with my photons (results being: averages of measurements on the set).
    It is only when you and I come together afterwards, and compare our individual results, that we will see that there are correlations.

    cheers,
    Patrick.
     
  9. Sep 29, 2005 #8
    So, in essence, I'd have to call you up and ask you about your data you have on betelgeuze which would basically defeat the point of using this method to communicate, right?
     
  10. Sep 29, 2005 #9

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Exactly !

    However, you can use the technique to encrypt a communication, in that someone evesdropping on your telephone call to Betelgeuze (and just getting my data) would not learn anything about the settings of MY detector over there ; however, YOU could derive it from the correlations you find between these data of mine (transmitted through the telephone) and your local measurements. These detector settings then contain the secret message.

    edit: btw I noticed that the rendering of the small text in my journal was screwed up (a part was not displayed) because of the use of > and its inverse (which was interpreted as the opening of an HTML tag). I corrected it now.
     
    Last edited: Sep 29, 2005
  11. Sep 29, 2005 #10
    You say that the correlation "between the two subsystems" changes when they're brought together. Not to be nitpicky, but I don't think that's a correct way of putting it. (Unless, I'm thinking about it incorrectly.)

    The two subsystems, A and B, are never correlated wrt each other. That is, we're not talking about how A varies as B varies, or vice versa. The correlation in the setting being considered involves variables other than A and B.

    This correlation involves pairing up the individual results wrt extrapolation to a common source by referencing a joint detection interval. Then the rate of these joint (A,B) results is correlated to variations in Theta.

    This correlation refers to the linear association between Theta (varying from 0 to 90 degrees) and the rate of coincidental detection. The correlation coefficient defines the strength (the relative linearity) of the correlation.

    Strictly speaking, there's actually a stronger correlation defined by the prototypical lhv correlation coefficient (linear) than by the qm correlation coefficient (nonlinear) and experimental results.

    (Maybe what people are referring to when they say that the qm correlation is stronger is that the qm probability curve more closely approximates the curve described by the adjusted experimental results. I don't know.)

    Anyway, what does change when the individual results are brought together and jointly correlated wrt Theta (ie., when the joint setting is what's being observed) is that the setting can no longer be talked about in terms of individual results.
     
  12. Sep 29, 2005 #11

    DrChinese

    User Avatar
    Science Advisor
    Gold Member

    The absolute value of the QM *statistical* correlation is always greater than or equal to that of any LHV theory. This is true for all theta values. The QM statistical correlation varies from 1 to -1 as the "match" ratio varies from 1 to 0 (as we vary theta from 0 to 90 degrees, for example). Only at 45 degrees they are equal. Bell's Theorem limits the values that a LHV theory can yield. That is why the QM correlation/anti-correlation is considered stronger.
     
  13. Sep 29, 2005 #12

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    No, I said that you NOTICE a correlation when you bring the data together.

    Let us go back to me on Betelgeuze and someone here on earth, Joe, each with 1 million photons in one million little boxes, them being paired up (with each one I have in my box Joe has one entangled one in one of his boxes).

    Now, I put my analyser to angle 0 degree for the first of my 1000 photons (and I record their individual results: 1,0,0,1,1,1,0,...) ; then I put my analyser to 45 degrees for the next 1000 photons (and I record their individual results: 0,1,0,0,1,1...) etc... for each set of 1000 photons I make a choice 0 degrees or 45 degrees ; and this choice is determined by a word of 1000 bits I would like to "send": if it is 0, I put the analyser to 0 degrees, if it is 1 I put the analyser to 45 degrees.

    In each of the series of measurements I have a random train of 1 and 0 (about 500 1 and about 500 0), no matter whether my analyser is 45 or 0 degrees.

    Joe always keeps his analyser to 0 degrees.
    For each series of 1000 measurements he will find a train of about 500 1 and 500 0, completely randomly distributed. He cannot find out individually, whether during that series, I (on Betelgeuze) had put my analyser to 0 or to 45 degrees.

    But the surprise comes when he calls me on the phone: he says: give me your readings for the first series: 1,0,0,1,1,1,0,...
    Lo and behold: that's EXACTLY THE SAME as what he had for the first series !

    Next series: in about half of the photons (500) Joe has a 1 when I have a 0 or vice versa, and the other 500 cases, Joe has the same as I. In other words the results are uncorrelated (50% coincidence).

    Next series...

    Each time when my analyser was put to 0, Joe recorded exactly the same string of 1000 results as I did, and each time when my analyser was put to 45 degrees, the results are uncorrelated.

    So Joe can, from my data, deduce when I put my analyser to 0 and when I put it to 45 degrees, hence reading my "set bits".

    But only Joe's results don't indicate anything: they are a random series of 0 and 1, 50-50 distributed.
    Only MY data (transmitted over the phone) don't indicate anything either: they are ALSO a random series of 0 and 1, 50-50 distributed.

    It is the agreement between both (their correlation) which indicates what was the angular relationship between the polarizers.

    cheers,
    Patrick.
     
  14. Sep 30, 2005 #13

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I would like to add that this particular result is no "proof" for Bell inequality violations. For instance, you can easily construct a LR model which does exactly the same: write pairs of 0-0 and 1-1 on pieces of paper, mix them, and cut them in two, putting one half in one box and the other half in the other box (so that the corresponding boxes contain both a paper with a 0 on it or both a paper with a 1 on it).

    Now, I, on Betelgeuze, will, instead of putting my polarizer to 0 degrees, just open the box and read the paper, and instead of putting my polarizer to 45 degrees, just throw a coin and have 1 for heads, and 0 for tails. It generates the same data. Joe will always open the boxes.

    The Bell kind of situation only arises when Joe picks 3 or more axis positions, and so do I.
     
  15. Sep 30, 2005 #14
    Is that an example of a quantum encryption technique?

    My confusion had to do with what people are talking about when they refer to "the correlations". That is, what exactly is being correlated to what. Does "the correlations" (when used in connection with Bell tests) have several meanings (varying with the observational setting), or always just one (eg., the comparison of individual, paired results).

    Maybe I'm just stuck on some simple thing that is obvious to people with the proper background in statistics and probability. My mission for tomorrow is to get a statistics book and actually read it. (any suggestions) I have a Schaum's Outline on probability. (is that sufficient?) And, I'm in Bohm's chapter on fluctuations, correlations and eigenfunctions. Any shortcuts to an eventual understanding that you can offer will be appreciated.
     
  16. Oct 1, 2005 #15

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I think so, but I'm no expert. In fact, the idea of having a common set of random bits (for instance, a spy, and his parent nation's secret service, who shared a booklet of random bits before the spy's mission) is a classical encryption technique, but the nice thing about the quantum version is that "you cannot make an illegal copy". If someone would steal Joe's (or my) photon boxes, he could not make a "copy" of them and restore them to me without destroying the entanglement which would mess up our system (and we'd notice). On the other hand, you could of course make a photocopy of the spy's booklet without him noticing, and then you could decrypt the messages.
    I think that the other advantage of this quantum encryption is that you can safely send in extra pairs of entangled photons (indeed, when the spy runs out of bits in his booklet, that's the end of the story, and he has to be careful to get a new one): if they are eavesdropped, they loose their entanglement.

    There are small variations in the definition, but I think you can simply state that the correlation of bitstreams (always of course between TWO sets of data, so making up pairs of data) is simply the probability (or the relative number of times) you have the SAME result over the total number of pairs, let us call this number c. c is a number between 0 and 1. In this case, 50% correlation means in fact totally uncorrelated streams. When the probability is higher, you have "correlated" bitstreams, and when the probability is lower, you have "anticorrelated" bitstreams
    (at least if in each stream, the number of 1 and 0 are about equal).

    You can also use the *correlation coefficient* which is given, for instance, by: http://en.wikipedia.org/wiki/Correlation
    This is a number that varies between -1 and 1 (and, for bitstreams, it is related to c by 2 c - 1)

    This is a number that varies between -1 and 1.

    cheers,
    Patrick.
     
  17. Oct 3, 2005 #16
    What about physical motion? Don't particle halfs move in the same way as their other half does if it's made to move? Physical motion, as long as going to the left, right, up, down, forwards and backwards is defined to mean something then couldn't information be sent? If I wave my hand, that tells you "hi" because we choose to agree on what that physical motion means. A jiggeling proton could do the same.
     
  18. Oct 3, 2005 #17

    DrChinese

    User Avatar
    Science Advisor
    Gold Member

    The symmetry of entangled particles can extend to position/momentum. You still cannot use that method to send a message, as the results on the "other" end still looks random and need decoding. You must send a normal message with the code key and now you are not ahead of the game.
     
  19. Oct 3, 2005 #18

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    It is a tricky question, because "movement of quantum particles" can mean several things. For instance, the momenta could be entangled, which means that the two particles don't have a well-defined momentum, until one is measured, in which case the other one takes on a definite value (a la projection postulate). Entangled photons usually only have their spins entangled but their "movement" isn't (that's however, not always the case ! In parametric down conversion, you can have an entanglement between momenta - the rainbow that gets out of it - and it depends on how you set up the experiment).

    However, what DOESN'T happen is the following: if I have 2 entangled particles (protons, say), and I *go and shake one* then the other one DOESN'T START SHAKING. I've even seen silly proposals of a space engine, where the motor would be left on earth and only the "entangled fuel" was taken in the rocket, but that's complete nonsense. This is not what entanglement can do. And luckily so. Otherwise we would be living in a very peculiar world. If my book would be entangled with some other book, if that other book would be opened, then my book would open too ? And the pages would start turning all by themselves ? That would be sheer magic. But again, this is not possible with entanglement. Entanglement just "helps decide" previously undetermined measurements, and even in such a way that it doesn't change the statistics, locally, but just "the order" in which they happen.
     
  20. Oct 4, 2005 #19
    Ah.
    New idea: Even though the states are random, states don't change without being made to, correct? If a state change happens, and you didn't cause it, you could conclude the other end did. If four random state changes per second(or whatever) is defined to mean a "1", two random state changes per second is defined to mean a "0" and no state changes per second means no data is being transmitted, then couldn't information be sent? When I have an idea, I look for any and all loop holes that I can think of to try to get around problems to make it work. If this comes off as irritating, I don't mean to be. I just want to learn.

    Ok.

    I have to disagree that it'd be "sheer magic". It'd only make the universe a little weirder, until you get used to it. Have you ever heard of superfluids? They are liquids that can flow up and out of open containers against the force of gravity; even out of several beakers nested within each other!

    Nuclear bombs, kilograms of matter being split and releasing enough energy to obliterate a large city (or greater). If I were a man in the 1870s and you told me these things were possible, I'd laugh in your face; because I should, until you can prove it. Like wise if proof for the possibility for making book pages turn from a remote location were shown to be possible, it wouldn't be magic.

    Just normal quantum entanglement, which you don't seem to think is weird, is pretty weird to me. Think about it, you can influence what a particle does from as many lightyears away as separate the two. Weird, and not a very different level of weirdness as psychic abilities would be to me, if they were real.
     
  21. Oct 4, 2005 #20

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    The problem is that you can't see the "state". You can only see the results of measurements, which are generated from a certain state but which don't tell you the exact state (in quantum theory, you can never find out the state of a single system!). And - it seems almost a conspiracy - the mix of states you get give you always the SAME statistical distribution of measurements on one side (that's what my little proof of the reduced density matrix tells you in fact).

    That's ok, it is indeed a good attitude. I often have it too :-)

    You're right. But knowing that about all stuff around us (if quantum mechanics works the way we think it works "all the way up" which is yet unknown) is about entangled with everything else in the visible universe, you start to see the mess it would make !

    Believe me, quantum entanglement is about the weirdest thing in physics I know of, and indeed on the verge of true magic. My personal opinion is that in fact nothing "happens" to the other particle, but that we have (even weirder) the experimenters being in superposition (where he/she got one result in one term, and another result in the other term) and that the "magic" only happens when the two experimenters come together and "interfere" there, producing the correlations. This is called the "many worlds view". You don't have to adhere to it :smile:
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Quantum Entanglement and Communication
Loading...