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

How does one type of detector determine path of photon?

  1. Apr 6, 2010 #1
    How does a screen make a particle behave differently than a "detector"?

    I was reading a 2007 Newsweek article Putting Time in a (Leaky) Bottle. In one version of the double slit experiment, if you fire photons through double slits and have a screen on the other side, your results are a wave interference (zebra) pattern on the screen- nothing special there.

    However, if instead of a screen, you have "detectors peeking at the slits", the photons travel as particles going through one of the two slits at a time.

    The article goes on to say, in the experiment, the screen can open up like blinds with the detectors behind it. If it stays closed after the photon goes through the slits, you get a wave pattern, if it opens AFTER the photon supposedly goes through slit, the detectors get a peak and suddenly the photon behaves as if it's been traveling as a particle even before the blinds open- as if it 'knows' the future.

    Here is what I don't understand...

    How does the screen make the photon behave differently than a "detector"? If the detector is passive (it doesn't fire anything out), then it should act just like a wall of particles that is the same as the screen, right? But the screen makes the photon act as a wave, and the detector results in the photon behaving as a particle. Where is the difference? Location of the detector?

    If the detector is active (so it is emitting other particles to detect it), then the photon is just hitting more particles and that should collapse the wave function just like hitting the screen, right?

    I am not seeing how the interaction of the detector and photon vs. a screen fundamentally changes the way it travels or behaves from the double slits. Both should collapse the wave function in the same way.

    If interested here is the original article:
  2. jcsd
  3. Apr 6, 2010 #2


    User Avatar
    Science Advisor

    They do not collapse the wave function in the same way in the sense that they do not collapse the wave function at the same PLACE. Namely, the screen collapses the wave function AT the screen, while the detector at the slit collapses the wave function AT the slit. Since interference occurs AFTER the slit, only the measurement at the slit, and not the measurement after the slit (at the screen) can destroy the ability of wave function to create interference.

    Does it help?
  4. Apr 6, 2010 #3
    Thanks for the reply! In this experiment described by the Putting Time in a (Leaky) Bottle article, the detectors are actually placed BEHIND the screen. The screen is able to open and close like "venetian blinds". After the photon has already gone through the double slits, if the screen remains closed, the photon results in wave pattern, naturally. However, if the blinds are open after the photon's gone through, the detectors show the photon as behaving as particle.

    So the detectors are basically at the same location as the screen (blinds) but one is "on" while the other is "off", changing what the particle DID after the slits.

    How does the detector differ from the screen in how it changes the behavior of the photon? They both should collapse the wave field in the same way.

    Here's a quote from the article:

    In the meantime, experiments have put detectors on the far side of the blinds. If the blinds are open and the detectors peek at the slits, photons fly through only one slit and no zebra stripes form. If the blinds are closed so the detectors cannot see the slits, photons fly through both and form the stripes. Here's the twist: if the blinds open only after photons have passed the slits but before they reach the blinds, the stripes fail to form even though the photons have seemingly done what they must to form stripes—namely, fly through two slits, as they always do when unobserved. The act of observing alters what the photons did earlier, somehow changing things so they passed through one slit and not two.

  5. Apr 6, 2010 #4


    User Avatar
    Science Advisor
    Gold Member

    Welcome to PhysicsForums, enduril!

    I looked at the article, and it is not completely clear about one crucial point: the detectors are able to infer "which-slit" the photon went through. So you have 2 scenarios:

    a) the blinds are closed, no which-path info is possible, and the interference pattern forms; or

    b) the blinds are open, the detectors know which-path information and no interference is registered.

    The important issue is whether the which path information is known (or knowable, in principle). It does not matter when it is learned. (Or where for that matter.) Only the context of the entire experiment matters, and that can extend over spacetime in a way which is not immediately obvious. So the context of the detector matters AT THE TIME OF DETECTION. That is why the decision to open the blinds or not "appears" to affect something that already happened. Actually, it is the nature of the observation which is being altered.

    There are a number of interesting "delayed choice" scenarios that work similar to the one described in the cited article. They always yield results consistent with the predictions of quantum mechanics. But they do challenge our ideas of spacetime.
  6. Apr 6, 2010 #5
    Okay, so the time of detection is what matters from what you are saying. So from the experiment, both the detector and the screen should "come into contact" with the photon AFTER the photon has traveled through slit(s). However, as long as it is AFTER the slits, there should not be any difference in how they determine the path of the photon (as wave or particle), right? Both detector and screen are just a wall or bundle of particles hitting the photon sometime after it has supposedly decided it's path.

    Right, so the nature of the observation has changed. I guess I am just asking how is the observation from the screen different from the observation from the detector?

    Thanks from the response!
  7. Apr 6, 2010 #6


    User Avatar
    Science Advisor
    Gold Member

    It is not really a question of when in the normal sense of the term. Nor, as I mentioned, where. It is the effective context of the experiment. The blinds don't matter when they are open, as that is my context b). In that context, the observer knows which slit the photon took to get to the detector. So there is NO interference.

    When the blinds are closed, you have a different context - my a). In this context, it is not possible to determine which slit the photon traversed. So there IS interference.

    The switching from one context to another is a matter of sleight of hand. There is only 1 context at a time - ever. Now, keep in mind this is simply what the experiment is designed to highlight: QM is context sensitive in its predictions. That context simply defies a classical description.
  8. Apr 6, 2010 #7
    If I understand correctly, the difference between when the detector is "looking at" the photon vs when the screen is in place is that it changes whether/what we know about which slit the photon traverses to. If we know which slit, particle behavior results, if we do not, wave results. As you put it, this changes the context of the experiment?

    So maybe this is what I am not getting. How does a detector allow us to determine which slit the photon traverses through whereas the screen does not? It may just be that I don't understand how a detector works.

    If a detector is just a wall of atoms that react to the photon, how is it different from a screen in knowing which slit the photon goes through? If it is an emitter of particles (like a flashlight), how does the detection of the photon through the "ping" of the emission determine which slit the photon goes through?

    I understand it is all about what information we have or don't have from the exp. that is the difference.

    But, at the particle level, how does a detector allow us to know which path the photon took, whereas a screen does not.

    Thanks for your great replies!
  9. Apr 6, 2010 #8


    User Avatar
    Science Advisor
    Gold Member

    It's not you... it is the description of the experiment that is lacking. Suffice it to say, they specifically setup the detector in such a way as they could determine the which slit information.

    So first, just ask this question: can you know the path the photon traversed? Regardless of the setup, if the answer is YES then there will be NO interference. And vice versa.

    Once you accept this as being a prediction of quantum mechanics, then you are free to try and figure out HOW this works. If you figure it out, let us know. :biggrin:
  10. Apr 6, 2010 #9
    Okay, now I am starting to understand the difference. So its not that there is something inherently different in the detector vs the screen. Its that the detectors are simply set up in a way that will give info on which path the photon took, whereas a screen can not obtain this info.

    So how exactly can detectors be set up to know which path a particle takes? Two detectors, one facing each slit? If this setup is viable, why would the photon not go straight down the middle as a wave?
  11. Apr 6, 2010 #10


    User Avatar
    Science Advisor
    Gold Member

    I assume that there is the photonic equivalent of a rifle barrel before the detector. And it doesn't really make any difference how it is constructed at all.

    Now you ask: what about photons going down the middle? Actually there might be ones that do that. The detectors won't necessarily obtain the exact same set of photons - this conclusion though is dependent on the exact setup.

    There is another well known setup which involves entangled photon pairs (members Alice and Bob make a pair) and a screen on one side - let's say the Alice stream. By observing Bob, you can sometimes determine the which path information (by which you can deduce Alice's behavior). For such Bobs, the matching Alices do not show interference. For the other Bobs - where which path information cannot be determined - the matching Alices show interference.

    So this example is why I say the where and when does not really matter. Because you can observe the Bobs anytime (and therefore anywhere) and Alice will act in accordance with what you learn from Bob. (An important note is that the Bobs cannot be controlled as to determining their which path information in such a way as to send a signal.)
  12. Apr 6, 2010 #11
    Thanks for all your replies. I know it doesn't matter how the setup is constructed. But in order to really understand what's going on, at least in my head, I need to understand where is that crucial point, that difference that causes a particle to behave one way and then suddenly and fundamentally in another. Its finding that exact point of change where you can say, that's what changes the results.

    For example, saying, "a particle changes behavior when it knows it is being observed," is not enough. What does it mean to be observed, does there have to be a consciousness? (No) Is it simply a disturbance of the particle? (No, there's fundamentally more to it). It seems its all about information. And how do you know it knows?

    In order to understand these questions I have to dissect these experiments and find that one variable or change that determines one outcome vs the other.

    I'm still not sure I understand how in one setup a detector would provide info on which path over a screen. But the difference does allow me to gain a better understanding of how info is obtained or not obtained when it comes to particles affecting the result of the experiment.

    Thanks again for all your info!
  13. Apr 6, 2010 #12


    User Avatar
    Science Advisor
    Gold Member

    Then perhaps you would be interested in some of the techniques for learning which path information. This experiment is one. Another I described above with Alice and Bob, who are isolated from each other. There are several other techniques.

    One is to place a polarizer in front of each slit, placed at 90 degrees apart. In that way, it would be possible to later analyze the photon polarization and determine which slit the particle passed through. And sure enough, the interference pattern disappears. But move the polarizers progressively closer together, and the interference returns - as the knowledge of the which slit disappears.

    So my point is that you can do anything you want in a particular setup to dissect what is going on. But there isn't much more to be learned. There are a number of interpretations that attempt to explain the underlying mechanism, but that all have issues of one kind or another.
  14. Apr 6, 2010 #13
    Would it be about the "Wheeler double slit delayed choice experiment" ? If it is indeed the experiment described below, then the double-slit picture is completely misleading, and should actually behave as enduril says. The real experiment has nothing to do with putting a detector behind a screen :


    At least, enduril is right when he says that a screen is nothing more than a wall of detectors.

    I have not studied the Wheeler experiment, but I know about Scully's delayed choice experiment. It does not exhibit backwards causation, nor necessarily backwards modification of a quantum state.
    It indeed exhibits a "spooky" distant correlation between entangled particles that no one can intuitively explain, and in EPR experiments, this correlation, stunningly, is non local.

    But in delayed choice experiments, everything can be described with normal wave function collapses that occur from past to future : in Alice / Bob experiments, it is equivalent to say that one or the other is actually collapsing the wave function. So is it with delayed choice experiments too, exept that Bob makes his measurement after Alice.

    For a mysterious reason, authors love to say that it is Bob who collapsed Alice's state after her measurement is finished, in a "delayed" way, while the opposite point of view, much more logical, is exactly equivalent.
  15. Apr 6, 2010 #14
    That article is full of interpretations, which is the real problem in quantum mechanics. If we can stick with only the formalism and the experimental results, in my opinion, things are less confusing.

    The screen does not make the particle behave differently from the detector, which determines which slit the photon went through. In fact, we do not need those detectors. There are other ways to determine "which way" information that "destroys" the interference. It all depends on the preparation procedure, which determines the quantum state vector. The preparation procedure corresponds to the experimental configuration in place at the instant that the photon is detected. Even if we make changes after we think the photon has already passed the slits (called "delayed choice"), the experimental results always agree with the experiment that exists at the moment the photon triggers the detector.

    If we do an experiment in which it is impossible, even in principle, to determine which slit the photon goes through then the state vector is the superposition state [tex]
    \left| \psi \right\rangle = \frac{1}{{\sqrt 2 }}\left[ {\left| {y_1 } \right\rangle + \left| {y_2 } \right\rangle } \right]
    where we assumed one slit is at [tex]
    y = y_1
    [/tex] and the other slit is at [tex]
    y = y_2
    [/tex]. The probability distribution for this state is identical to the interference pattern seen in wave optics.

    On the other hand, if we do an experiment in which we know, and it does not matter how we know, which slit the photon goes through then the state vector is an eigenvector, either [tex]
    \left| \psi \right\rangle = \left| {y_1 } \right\rangle
    [/tex] or [tex]
    \left| \psi \right\rangle = \left| {y_2 } \right\rangle
    [/tex]. Such eigenstates do not show any interference effects.

    If you are familiar with the formalism, Dr. Chinese' experiment with polarizers covering each slit is, I think, an easier example that demonstrates these concepts.

    Best wishes
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook