Delayed choice quantum eraser experiment interference pattern question

[livethere]

If I understand the delayed choice quantum eraser experiment correctly, it's essentially set up in the manner as the double slit. The only difference is that the detectors are set up in front of the double slits.

It was thought that if there was an interference pattern that appeared then that means that the probability wave had passed through both slits, however if it did not appear then that meant that the probability wave had only passed through one slit.

It is my understanding that once this test was conducted, it was shown that there was no interference pattern. It seemed like the probability wave had gone through both slits, saw that there was a detector, and then went back in time to only enter through one slit.

To avoid this paradox, physicists said that it could be resolved by assuming that the probability wave was in a superposition of passing through both slits prior to observation and then the observation generated in phenomenon in which the wave only passed through one slit, and so no interference pattern.

This would make sense to me, however if this was true then why do we still witness the interference pattern in the double slit experiment? It seems like the same situation should be produced normally.

I would also like to say that I'm asking this to gain understanding, and I'm not trying to suggest that the physicists are not correct.

 

[andrewkirk]

There is no paradox to be resolved and no need for any 'going back in time'.
The experiment is complex and cannot be simply summarised, but it well repays reading the original paper, which is quite clearly written.

There is no interference pattern amongst the set of all 'signal' photons striking the detector D0. However, the file listing the locations of all such photon detections can be partitioned into four files, based on which of four other detectors D1-D4 detected the 'idler' photon that was the entangled twin of a signal photon.

Two of those files will show interference patterns, and two will not. When the results of the first two are combined (superimposed) the interference pattern disappears because the bright stripes of one correspond to dark stripes of the other.

 

[whitsona2]

There is no paradox to be resolved and no need for any 'going back in time'.
The experiment is complex and cannot be simply summarised, but it well repays reading the original paper, which is quite clearly written.

There is no interference pattern amongst the set of all 'signal' photons striking the detector D0. However, the file listing the locations of all such photon detections can be partitioned into four files, based on which of four other detectors D1-D4 detected the 'idler' photon that was the entangled twin of a signal photon.

Two of those files will show interference patterns, and two will not. When the results of the first two are combined (superimposed) the interference pattern disappears because the bright stripes of one correspond to dark stripes of the other.

Thanks for this-- what is the original paper? I run a maker space for middle and high school kids and two are doing the SciAm quantum entanglement right now and all I can tell them is that some physicists have a model for the eraser that doesn't require retro causality. Need more. Middle school minds are insatiable.

 

[andrewkirk]

Thanks for this-- what is the original paper?

Hello. This link is to the original paper by Kim et al, that reported the experiment. I think it's fairly readable for a physics paper.

 

[Viktor74]

I have a question that probably was already understood by physicists and has an answer. What would happen if the delay time between the idler and signal photons would be not 8ns but actually 1 minute. In other words, the photons arriving at D0 path will be faster my one minute than to other detectors. A whole set of photons arrive at D0 before anyone arrives at any other detectors. Suppose the human observer sees the pattern at D0 (since delay allows that) now before any other photons arrive at other locations. The pattern will be only interference at first (since path information has not been determined yet), idler photos en route. But what would happen to that pattern (that has already been observed and remembered before any other photons reached path/no path detectors) once the other photons finally reach the other detectors and take "path known" or "path erased" routes (50/50 chance according to experiment).

 

[andrewkirk]

@Viktor74 There is no screen in this experiment, but a detector D0 that moves over the target space. Images are built up in a computer by matching 'signal' photon strikes recorded by that detector with those of their 'idler' partners at D1 - D4. No image at all is available from the computer until the matching is done. So the time taken for the idler photons to reach the detectors D0-D4 does not make a difference to the outcome - it just means we have to wait a bit longer to see it.