1
$\begingroup$

The double slit experiment with a single electron provides a different result when observed and when unobserved. When unobserved we see an interference pattern, and when observed we just see 2 lines. When the experiment is unobserved we see an interference pattern at all points, e.g exactly between the 2 slits we may see a fringe, but when observed there is nothing there.

If when observed and the wave function is collapsed, the wave function stays collapsed, that would explain this result. When collapsed the photon behaves instead like a particle, meaning that it does not diffract around the slit but can only travel in a straight line. Unfortunately, this doesnt seem to be correct. A google search says that the wave function is only briefly collapsed (distance of 1 hydrogen atom), so the photon should go back to acting as a wave.

Is the reason just that the common illustrations are exaggerated and there is less of a obvious 2 line pattern?

$\endgroup$
3
  • $\begingroup$ see the experiment for photons in this answer physics.stackexchange.com/questions/90646/… also my answer here might help physics.stackexchange.com/questions/577312/… $\endgroup$
    – anna v
    May 10, 2023 at 10:44
  • $\begingroup$ Comments have been moved to chat; please do not continue the discussion here. Before posting a comment below this one, please review the purposes of comments. Comments that do not request clarification or suggest improvements usually belong as an answer, on Physics Meta, or in Physics Chat. Comments continuing discussion may be removed. $\endgroup$
    – Buzz
    May 12, 2023 at 7:00
  • $\begingroup$ What I can't understand is how is there a way to know what light is doing when not being observed. I take "observed" to include all forms of monitoring, including video recording. So maybe I need some clarification as to what "observed" means in this case and how "unobserved" is actually being "observed." :-/ $\endgroup$
    – Zahhar
    Apr 17 at 16:46

5 Answers 5

3
$\begingroup$

When unobserved we see an interference pattern, and when observed we just see 2 lines. When the experiment is unobserved [...] e.g exactly between the 2 slits we may see a fringe, but when observed there is nothing there.

That's not true, though it's a common mistake. Here's an incorrect picture of the experiment from Sean Carroll's blog:

Here's a correct illustration from The Feynman Lectures:

What you get when there's a detector at one or both slits is the sum of $I_1$ and $I_2$, and what you get without a detector is $I_{12}$. Note that $I_1+I_2$ is large wherever $I_{12}$ is large; in fact, $I_{12} \le 2(I_1+I_2)$ everywhere. So it's not possible for a point on the screen where there is almost no light when the detector is present to receive a large amount of light when the detector is absent.

$\endgroup$
1
  • $\begingroup$ beautiful and correct. take my upvote $\endgroup$ May 11, 2023 at 8:22
3
$\begingroup$

It does not matter that the thing you are studying is a photon, so I will speak of the general particle.

There is no Hamiltonian whereby position localisation are eigenfunctions (but somewhat spread out ones are, e.g. QHO, H atom solutions, etc). As long as you have travelling solutions, they would be somewhat wavy.

When you perform a localisation observation on the particle, the eventual evolution of the particle will look as if it were to start from a wavefunction that was localised to the place you looked at. In this case, that is one of the slits. The particle would behave as if it were everywhere in that one slit, and nowhere else. Then it will propagate as a wave from then on.

That is how we are supposed to get two Gaußian spreads, what you call as two lines. Those are not sharp lines, but the important part is that you do not see the interference pattern of many peaks and troughs. The lack of interference pattern is particularly obvious.

It can be exaggerated, as in, it can be one big blob with a slight variation in brightness, but it is not too exaggerated. You can sometimes see the two lines separately. The important part is, again, that you do not see the interference fringes.

$\endgroup$
8
  • 1
    $\begingroup$ As you pointed out yourself in your comment to my answer, there is nothing unusual about light exhibiting an interference pattern - it proves its wave nature, but not its quantum nature. So discussing all particles in the same way is misleading. $\endgroup$
    – Roger V.
    May 10, 2023 at 10:21
  • 1
    $\begingroup$ The terminology of "wave" v.s. "particle" is not the pedagogically wise one. It is much better to describe all particles' behaviour as they quantum mechanically are. Yes, I know the irony of calling them particles right after saying they are not---it is not my fault that we only have that term for the "(asymptotic) excitations of underlying quantum fields" that is being meant in the 2nd usage. For what people would classically think of as particles, e.g. electrons, interference pattern is the surprising quantum behaviour. Then we should say that photons doing that is also worth emphasising. $\endgroup$ May 10, 2023 at 10:30
  • $\begingroup$ Similarly, classically we think photons should be waves, but they exhibit particle-like behaviour, and that is also worth emphasising. In particular, the behaviour of a large wavepacket that supposedly went through both slits and is later, after localisation observation, behaves as if the wavepacket is much smaller than before, is a quantum behaviour worth emphasising, and that is what I have done in my answer. $\endgroup$ May 10, 2023 at 10:32
  • $\begingroup$ I get that the interference fringes are gone, but I still don't get why you could ever see the 2 lines separately. Or what might make it more obvious, the spread for the interference pattern when unobserved is larger than the sum of the 2 spreads when observed. If modelled as 2 waves propagated from each slit, then the resulting overall spread should be the same (although with no interference) as the unobserved case, but most diagrams and explanations seem to disagree. $\endgroup$
    – Mercury
    May 10, 2023 at 10:44
  • 1
    $\begingroup$ @Mercury even with one slit one can have Fresnel diffraction. But this is not the same interference pattern that comes from the two slits. $\endgroup$
    – Roger V.
    May 10, 2023 at 11:34
0
$\begingroup$

I think we need to clarify what do you mean by "observed". The outcome of the two-slit experiment is the same regardless if you look at it or not. The only thing that matters is the experimental setup.

If you want to know which slit the particle passed through you need to place some sort of detector at the slits. This modified setup, with detectors at the slits would not generate an interference pattern. I assume that you have such a setup in mind when you refer to an "observed" experiment.

OK, so we have two different experiments, Experiment 1 consisting of a particle source, a barrier with two slits and a screen and Experiment 2 consisting of a particle source, a barrier with two slits, two detectors placed at the slits and a screen.

The pattern we see on the screen is determined by the way the incoming electrons interact with the barrier. In the Experiment 1 the barrier does not have detectors and some electrons are scattered between the slits. In the Experiment 2 the barrier has detectors and no electron is scattered between the slits. We can therefore conclude that the presence of those detectors is causing the electrons to change their paths.

I don't think that imagining the electron as changing from a particle to a wave and back is the right way to think about what happens, since such a change needs to be non-local (the wave needs to "collapse" instantly at a point) which is in conflict with relativity.

$\endgroup$
0
$\begingroup$

I agree that the Double Slit (DS) with photons is conceptually similar to the DS with an electron (or other quantum particle types). However, there are experimental & theoretical variations on the DS with photons that cannot be easily performed for other particle types. And this gives us more insight. Let's take the case of photons one at a time to build up a pattern, and vary our knowledge of "which slit" information. We CAN do this without affecting the total number of photons which will hit the screen.

As you know, the general rule is that if you could NOT know which slit the photon goes through, there will be interference. If you could know which slit the photon goes through, there will NOT be interference. It does not matter if you actually know the which slit information, it is enough that you could know it. And it is possible to vary continuously between these two extremes and get corresponding variation in the pattern (i.e. a mixture of interference and no interference).

The method we consider is when we place polarizers over each slit. a) If the polarizers are aligned parallel, there WILL be interference. b) If the polarizers are aligned orthogonal (crossed, perpendicular), there will NOT be interference. Parallel: it is NOT possible to extract which slit information. Orthogonal: it IS possible to extract which slit information. Either way, 50% of the photons will be blocked by the polarizers. So the total light detected remains constant.

Also, you may choose to look at the b) scenario as follows: there can be no interference between the slits if they are sending through the "portion" of the possible "paths" (for lack of a better description - insert your preferred description here) that are themselves now orthogonal. By definition pretty much, these cannot self interfere. Of course they can self interfere in the a) scenario.

Note that there is no mention of "collapse" or "localization". There is no need to refer to point particles or similar. So yes with a "*"; the answer is that in the DS with photons (or whatever), detection will occur much greater in some places when there is self-interference ("unobserved") than when there is no such interference ("observed"). As pointed out by @benrg, the diffraction that occurs when there is no self-interference places some light everywhere. So it might be more accurate to say: When observed, we see light in places we do not see when unobserved. You can compare Figures 8 and 9 in the experimental reference below to see this is the case. References:

Theory: Polarized Light https://books.google.com/books?id=w6PMBQAAQBAJ&pg=PA271&lpg=PA271&dq=double+slit+interference+polarizers+Fresnel&source=bl&ots=MumGUCvweh&sig=7wit8bq5xXkunYoHpnbB3hL3YXw&hl=en&sa=X&ved=0CB0Q6AEwAGoVChMI-tm_ueGNyQIVBigeCh1SLwwt#v=onepage&q=double%20slit%20interference%20polarizers%20Fresnel&f=false

Experiment: Young’s double-slit experiment with single photons and quantum eraser https://sciencedemonstrations.fas.harvard.edu/files/science-demonstrations/files/single_photon_paper.pdf

$\endgroup$
10
  • $\begingroup$ All photons are self interfering all the time per Dirac and Feynman ... they determine their own paths ..... whether they go thru single or double slits. In your (a),(b) polarization examples a photon can choose either path in (a) but can only choose one path in (b). So I would not mix up the terminology of "self interference" with the term "interference pattern". $\endgroup$ May 10, 2023 at 22:37
  • $\begingroup$ @PhysicsDave Interference is the absence of interaction. It's a phenomenon that can only exist in (almost) linear systems. To say that light interferes with itself is a misunderstanding of both classical and quantum superposition. That "photons" are objects that have a path is an old mistake that goes back to Einstein's 1905 paper about the photoelectric effect. It's time to abandon it. A photon is a (small) amount of electromagnetic field energy. $\endgroup$ May 11, 2023 at 5:24
  • $\begingroup$ Feynman and his path integral (which he applied to light) were/are instrumental to understanding and are the basis for the modern physics theories we have today. Yes ... "interference" is a broadly misused/misunderstood term .... as is evident in the above answer. Dirac and Feynman may have even used the term perjoratively. In modern field theory energy can have a direction. $\endgroup$ May 11, 2023 at 11:47
  • $\begingroup$ @PhysicsDave In the referenced setup: Photons are free to go through both slits in both the a) and b) scenarios. Hopefully we agree that when the polarizers are orthogonal, the effects of possible paths (or however you choose to refer to them) through one slit cannot affect the outcome of possible paths through the other. I think my use of the terms "interference (or interference pattern)" and "self-interference" are standard, and certainly follow the usage in both of my references. $\endgroup$
    – DrChinese
    May 11, 2023 at 13:40
  • $\begingroup$ @FlatterMann I recognize the nuance that in orthodox QFT, photons don't have a "path" or "position" in any traditional sense. Yet, as PhysicsDave points out, the Feynman path integral model is still widely used in both theoretical and experimental papers. I follow common usage by saying that a photon's position is represented by a click on a detector and its path is however/whenever it went from its source to that detector. For discussion purposes, writers operate "as if" these photon properties are present. My references use this lingo exactly as I have. $\endgroup$
    – DrChinese
    May 11, 2023 at 13:53
-4
$\begingroup$

The double slit experiment is a mostly a thought experiment, intended to illustrate the wave and particle properties of particles/light. There exist some literal physical realizations of such experiments, but most discussions are still aim at being pedagogical, rather than taking into account all the possible effects that may take place. In this sense, quantitative statements like the wave function is only briefly collapsed (distance of 1 hydrogen atom), so the photon should go back to acting as a wave are generally out of context.

When discussed as an experiment with electrons, the experiment simply aims at showing the interference: how electrons, classically thought of as particles, actually represent wave properties, if described quantum mechanically.

Light, on the other hand, is classically a wave, which exhibits interference and diffraction. Notably, diffraction can be even observed with a single slit/pinhole/whatever (image source): enter image description here

What makes double slit experiment special in case of light is when the light intensity is very low (so low that photons pass one by one) - in which case the interference pattern may vanish, demonstrating that photon is a particle. (Note that saying photons pass one by one makes sense only after we proved that photon is a particle.) There is nothing wrong with photon still interfering from a single slit - the interference pattern is clearly different from the light passing through two slits.

$\endgroup$
11
  • 4
    $\begingroup$ The interference pattern does not vanish when the photons pass one-by-one. We know that the case of photons passing one-by-one gives the same interference pattern as with many at once, even as strongly as laser light. Which in itself is a problem because we know that at much higher occupation numbers/temperatures etc, there should be discernible differences, but we have yet to experimentally observe the theoretically expected discrepancies. $\endgroup$ May 10, 2023 at 10:10
  • $\begingroup$ @naturallyInconsistent However, such experiments demonstrate that particles do not form the interference pattern if one detects which slit they pass through. See Double-slit experiment $\endgroup$
    – Roger V.
    May 10, 2023 at 10:18
  • 2
    $\begingroup$ That behaviour is shared between the many photon case and one-by-one case, so why did you phrase it that way in your answer? $\endgroup$ May 10, 2023 at 10:24
  • 4
    $\begingroup$ I am not disputing that. I am only specifically pointing out that you are insinuating in your answer's phrasing that if you take a many-photon-thus-interfering case and lower the intensity until there is only one photon at a time, the interference pattern may vanish, when in fact we know it does not vanish. $\endgroup$ May 10, 2023 at 10:34
  • 3
    $\begingroup$ No, the experiment is totally different and that is why the interference pattern changes. You are supposed to discard the photons that evaded the slit detection and only take data when the slit detection happened, and in doing this properly, the many-photon case agrees with the one-photon case. $\endgroup$ May 10, 2023 at 10:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.