Observing a double-slit experiment

[oknow]

TL;DR Summary

Can observing the operation a double-slit experiment influence whether or not an interference pattern forms?

Do I have the proper understanding of the following three double-slit experiment situations?

#1 While a standard double-slit experiment is run via a Mach-Zehnder interferometer apparatus that completely lacks which-way detectors, can an ordinary human experimenter be present and watch that apparatus operate without causing the loss of the familiar interference pattern? My understanding is that under such operation, yes, an interference pattern will form on the detector screen.

#2 Next, rerun the same experiment as in #1, but purely as a thought experiment, consider an experimenter that has super-human vision such that he/she can see each interferometer mirror separately vibrate/recoil each time a photon bounces off it. In this case, would the interference pattern form? My understanding is it would not. Because which-way information would be known to this experimenter, an interference pattern would not form on the detector screen.

#3 What if the experimenters from #1 and #2 (the one with ordinary vision and the one with super-human vision) jointly, and in view of each other, watch the experiment operate? My understanding is both experimenters will see no interference pattern form. Is that correct?

 

[PeroK]

#2 Next, rerun the same experiment as in #1, but purely as a thought experiment, consider an experimenter that has super-human vision such that he/she can see each interferometer mirror separately vibrate/recoil each time a photon bounces off it. In this case, would the interference pattern form? My understanding is it would not. Because which-way information would be known to this experimenter, an interference pattern would not form on the detector screen.

#3 What if the experimenters from #1 and #2 (the one with ordinary vision and the one with super-human vision) jointly, and in view of each other, watch the experiment operate? My understanding is both experimenters will see no interference pattern form. Is that correct?

In both cases you effectively have a measuring device, where in an unspecified way a macroscopic object irreversibly records which-way information.

 

[DrChinese]

Summary:: Can observing the operation a double-slit experiment influence whether or not an interference pattern forms?

Do I have the proper understanding of the following three double-slit experiment situations?

#1 While a standard double-slit experiment is run via a Mach-Zehnder interferometer apparatus that completely lacks which-way detectors, can an ordinary human experimenter be present and watch that apparatus operate without causing the loss of the familiar interference pattern? My understanding is that under such operation, yes, an interference pattern will form on the detector screen.

#2 Next, rerun the same experiment as in #1, but purely as a thought experiment, consider an experimenter that has super-human vision such that he/she can see each interferometer mirror separately vibrate/recoil each time a photon bounces off it. In this case, would the interference pattern form? My understanding is it would not. Because which-way information would be known to this experimenter, an interference pattern would not form on the detector screen.

#3 What if the experimenters from #1 and #2 (the one with ordinary vision and the one with super-human vision) jointly, and in view of each other, watch the experiment operate? My understanding is both experimenters will see no interference pattern form. Is that correct?

1. This is not a standard double slit experiment. The MZ demonstrates some of the same issues, but is completely different (there are no slits, for example). There is no "familiar" interference pattern in the MZ. Humans watch these experiments, and of course that does not change the results.

2. There is no recoil of the type you are considering. If there were, then you *could* in principle know which-path information; and so there would be no interference. The presence of an observer makes no difference.

 

[oknow]

OK, thanks for the replies, they are helping. Please consider scenario #4:

#4) A visual of this experiment's detector screen, and only that, is being streamed live via the interwebs. Those watching this live stream are provided no way in advance to know whether experimenter #1 (with ordinary vision) will be observing the interferometer operate, or instead experimenter #2 (with super-human vision) will do so. Will those watching this live streaming broadcast from afar see an interference pattern or not?

 

[DrChinese]

OK, thanks for the replies, they are helping. Please consider scenario #4:

#4) A visual of this experiment's detector screen, and only that, is being streamed live via the interwebs. Those watching this live stream are provided no way in advance to know whether experimenter #1 (with ordinary vision) will be observing the interferometer operate, or instead experimenter #2 (with super-human vision) will do so. Will those watching this live streaming broadcast from afar see an interference pattern or not?

There are already true double slit experiments that demonstrate conclusively that the observer's role makes no difference to the outcome. The critical factor is whether the which-path information potentially exists in the environment, regardless of whether it is recorded, analyzed or otherwise made into information that could be observed and used to determine which-path. Here is one:

https://sciencedemonstrations.fas.h...-demonstrations/files/single_photon_paper.pdf

 

[oknow]

OK, thanks, what about the following more-complex scenario #5:

A double-slit experiment isolated within an ideal box is activated by a timer, and runs for a brief but sufficient period to build an interference pattern, or not, on its detector screen. The screen saves the pattern, or lack thereof. The apparatus is equipped with which-way monitors that have a 50% chance of being enabled (for the full run) by a radioactive decay after the box is sealed but prior to the start of the run. If these monitors are not enabled, no which-way information is observed. If the monitors happen to be enabled for the run, none of the observed which-way information is intentionally stored by the device. After the run period has ended, an experimenter opens the box. What are the odds the experimenter will see an interference pattern recorded on the detector screen? If this type of experiment actually has been performed, can someone recommend a paper online that describes it?

 

[PeroK]

OK, thanks, what about the following more-complex scenario #5:

A double-slit experiment isolated within an ideal box is activated by a timer, and runs for a brief but sufficient period to build an interference pattern, or not, on its detector screen. The screen saves the pattern, or lack thereof. The apparatus is equipped with which-way monitors that have a 50% chance of being enabled (for the full run) by a radioactive decay after the box is sealed but prior to the start of the run. If these monitors are not enabled, no which-way information is observed. If the monitors happen to be enabled for the run, none of the observed which-way information is intentionally stored by the device. After the run period has ended, an experimenter opens the box. What are the odds the experimenter will see an interference pattern recorded on the detector screen? If this type of experiment actually has been performed, can someone recommend a paper online that describes it?

There's no mystery here. The QM system interacts with its macroscopic environment, which leads to decoherence and essentially classical probabilities. It's just an interference pattern instead of a dead cat; and, no interference pattern instead of a live cat.

 

[DrChinese]

OK, thanks, what about the following more-complex scenario #5:

...If the monitors happen to be enabled for the run, none of the observed which-way information is intentionally stored by the device. ...

It does not matter if the which-way information is stored or not. There will be NO interference pattern if the monitors are on.

The reference I provided is an experimental example of exactly this scenario. The polarizers (when perpendicular) act as "monitors" to encode which-slit information in the photons themselves, and this information is ignored. There is no interference.

 

[Jarek 31]

Weak measurement allows to measure in weakly enough way not to destroy interference, but still allowing to get e.g. mean trajectories from statistics: https://science.sciencemag.org/content/332/6034/1170

On the top you can see the 2 slits, on the bottom interference pattern:

 

[Cthugha]

Weak measurement allows to measure in weakly enough way not to destroy interference, but still allowing to get e.g. mean trajectories from statistics: https://science.sciencemag.org/content/332/6034/1170

The strong emphasis here is on "mean". The weak measurement does not provide information on which path any individual photon took.

 

[Jarek 31]

Exactly - strong measurement would destroy interference, weak allows to avoid it - finding mean trajectories from statistics, which agree with dBB:

Single-particle trajectories measured in this fashion reproduce those predicted by the Bohm-de Broglie interpretation of quantum mechanics, although the reconstruction is in no way dependent on a choice of interpretation.”

 

[Cthugha]

Exactly - strong measurement would destroy interference, weak allows to avoid it - finding mean trajectories from statistics, which agree with dBB:

"Allows to find it" is a misleading statement. This is the same result you get from tracking the Poynting vector of a classical wave. You could "find" these trajectories just as well by putting the observation screen at different distances to the slit and recording the interference patterns at the different distances. Actually, this is what has been done in the figure you posted. The figure shows the strongly measured interference patterns. The results from the weak measurement are shown by the black lines for comparison. Sure, performing this strong measurement does not carry any information about the individual photons, but neither does the weak measurement.

Weak measurements have their uses, most prominently in open systems. It is also important to understand how they differ from strong measurements to avoid some pit traps in thought experiments involving, e.g. tiny momentum exchange. However, in this thread, they may be somewhat off-topic.

 

[Jarek 31]

Ok, to be more specific: they encode which-path information in polarization (using birefringent crystal), then use polarization beam displacer for the photons to get to one of 2 (L,R) CCDs, and calculate mean momentums from differences of their statistics:

 

[Cthugha]

How is your response related to anything I wrote?

It is also not correct. They do not encode which-path information anywhere in the polarization. All the photons are prepared with diagonal polarization. They then use an off-axis oriented piece of calcite to slightly rotate the polarization of the photons. The rotation depends on the path through the calcite crystal and thus depends on the angle with which the photons arrive - which corresponds to momentum. They then go on to use a set of lenses to image individual planes at different distance between the double slit and the calcite crystal onto a CCD. This yields the position information. The polarization one finds at each position provides (due to the calcite) a weak measurement of the mean momentum of photons arriving at this very position.

It is the weak momentum information that gets encoded in the polarization, not any "which path" information. Still, I do not see how all of this is relevant to the topic at hand. Could you explain why you think it is?

 

[Jarek 31]

You have written I have made a misleading statement, so I wanted to clarify.
My mistake - you are right: calcite is to encode momentum, not which-path information.
Here is the setting: