Delayed Choice Quantum Eraser Question

[04qWIGk3]

I watched this video that dumbs down and explains very simply how the quantum eraser works

From my understanding: The particle will go thru both slits if not observed (probably wave collapsed). If you setup a device that will detect it (that emits a photon if the particles passes thru) the particle will act like a particle and not a wave... But if you use a quantum eraser (by having the photon go thru a lasersplitter) the particle will once again act as a wave.

So what happens if you delay the amount of time before the photon hits a detector? Just to keep things simple, let's say you setup a bunch of mirrors so the photon took 2 seconds to hit a detector or a lasersplitters. If you then threw a particle and saw that it hit the detector in a spot that would correspond to a particle, and within the 2 seconds you added the lasersplitter, what would happen? Same question about if you did the opposite, you setup lasersplitter, saw the particle acting as a wave, and trolled the universe by removing the lasersplitter within 2 seconds?

Seems I a missing information, because I cannot see how this experiment could work (unless it can work and it can trap the universe in an infinite time loop :P )

 

[.Scott]

From my understanding: The particle will go thru both slits if not observed (probably wave collapsed). If you setup a device that will detect it (that emits a photon if the particles passes thru) the particle will act like a particle and not a wave... But if you use a quantum eraser (by having the photon go thru a lasersplitter) the particle will once again act as a wave.

The delayed choice quantum eraser is a tough piece of logic to "dumb down". Everything that the voice over states it correct - but you didn't catch all of it.

What is going on can be interpreted as a time issue - but it's easier to catch what's going on by interpreting it as it is presented by the voice over. So that is how I will describe it in this response.

The first thing that you need to catch is that the particles go through both detectors even though only one detector reports a detection. So, by this interpretation, when the beam splitter hides the which-way information, it is not actually changing any pattern.

The second thing you have to catch is the phrase "a stripe pattern can be recovered". In this experiment, if you look directly at the screen you will never see the stripe pattern. Instead, you need to pick out those photons from one of the two sets where the eraser has been activated. Remember that when the beam splitter erases the which-way information, it creates two beams - call them A and B. So before you use that beam splitter, you don't have the information you need to see either the A or B stripes. You only seen the combined A and B pattern - which looks exactly like the unerased left plus right pattern.

So what happens if you delay the amount of time before the photon hits a detector? Just to keep things simple, let's say you setup a bunch of mirrors so the photon took 2 seconds to hit a detector or a lasersplitters. If you then threw a particle and saw that it hit the detector in a spot that would correspond to a particle, and within the 2 seconds you added the lasersplitter, what would happen?

Once you have determined which detector was struck, it is too late to move the beam splitter - the photon has already passed through it. The detector emits only one photon - and that one photon is the only evidence in the universe about which way the photon went. So, if you run the experiment as described, you can't know which detector the photon went through without looking at that photon. But if you look at the photon, changing the beam splitter won't affect anything because the photon is no longer on the fly.

If, on the other hand, you changed the experiment so that you knew which path the particle would take even before hitting the detector, then it would never be possible to erase that even with the beam splitter.

Same question about if you did the opposite, you setup lasersplitter, saw the particle acting as a wave, and trolled the universe by removing the lasersplitter within 2 seconds?

Again, once the particle has crossed through the beam splitter, it doesn't matter what you do with that device. Before the particle hits the beam splitter, it is unmeasured and free to assume any of the four states (left, right, erased A, or erased B).

 

[04qWIGk3]

This is all very interesting, but over my head.

The first thing that you need to catch is that the particles go through both detectors even though only one detector reports a detection. So, by this interpretation, when the beam splitter hides the which-way information, it is not actually changing any pattern.

To me that seems to imply the 'particle' is a wave on its way to the detector, turned into a particle to be detected at the detector, turned back into a wave to go thru the slits, interfered with itself as a wave, and back into a particle once it hit the wall.

The second thing you have to catch is the phrase "a stripe pattern can be recovered". In this experiment, if you look directly at the screen you will never see the stripe pattern. Instead, you need to pick out those photons from one of the two sets where the eraser has been activated. Remember that when the beam splitter erases the which-way information, it creates two beams - call them A and B. So before you use that beam splitter, you don't have the information you need to see either the A or B stripes. You only seen the combined A and B pattern - which looks exactly like the unerased left plus right pattern.

How can just looking at the pattern of detector A or B cause a strip pattern? Also the video said there was a 50% it would pass thru, and 50% it would bounce off... If there is a true 50% of the photon hitting either detector, then how does it make sense that any pattern can be recovered?

Once you have determined which detector was struck, it is too late to move the beam splitter - the photon has already passed through it. The detector emits only one photon - and that one photon is the only evidence in the universe about which way the photon went. So, if you run the experiment as described, you can't know which detector the photon went through without looking at that photon. But if you look at the photon, changing the beam splitter won't affect anything because the photon is no longer on the fly.

If, on the other hand, you changed the experiment so that you knew which path the particle would take even before hitting the detector, then it would never be possible to erase that even with the beam splitter.Again, once the particle has crossed through the beam splitter, it doesn't matter what you do with that device. Before the particle hits the beam splitter, it is unmeasured and free to assume any of the four states (left, right, erased A, or erased B).

What I am proposing is you do not look at the photon. You setup in such a way that the photon is traveling for 2 seconds bouncing off mirrors (completely undetected). During that 2 seconds the observer would see the location the particle hit the wall. At that point within 2 seconds (like the photon is still traveling) he could either slide in a beam splitter or slide it out.

Looks like this general idea has already come up tho https://en.wikipedia.org/wiki/Wheeler's_delayed_choice_experiment and https://en.wikipedia.org/wiki/Delayed_choice_quantum_eraser and from what I understand, seems the test they are running might show some causality issues, but there might be other explanations.

 

[vanhees71]

It is not that weird to understand, if you use the statistical interpretation of quantum mechanics. What is done is to choose different subensembles of a large ensemble from a measurement protocol, which either allow to know which path the photons have taken (one can say a lot of "path" concerning photons, because there's no position operator for photons in the strict sense) or to strictly not being able to know this information. Then you see the interference pattern. There is nothing mysterious with that, when the minimal statistical interpretation of quantum theory is used. Particularly, there's no retrocausality whatsoever.

 

[.Scott]

This is all very interesting, but over my head.

I'll try to do better.

To me that seems to imply the 'particle' is a wave on its way to the detector, turned into a particle to be detected at the detector, turned back into a wave to go thru the slits, interfered with itself as a wave, and back into a particle once it hit the wall.

That would be confusing. Try this: the photons are waves until any measurement is made. So, as is described in the video, the particle always goes through both detectors and the photon always goes on to follow all four paths - yellow and green, erased and unerased. Then, when a measurement is made, the photon "commits" to one of those four paths. So there is only one "collapse" affecting both yellow/green and erased/unerased in a single detection. And in no case, does the photon detection actually change anything on the screen. It only looks that way because we are tempted to think in terms of well-defined time sequences and photon paths instead of wave functions where the photon can be spread out across four paths.

How can just looking at the pattern of detector A or B cause a strip pattern? Also the video said there was a 50% it would pass thru, and 50% it would bounce off... If there is a true 50% of the photon hitting either detector, then how does it make sense that any pattern can be recovered?

In this experiment, the particles do not form a striped pattern that can be seen without the photon information. To see the stripes, you need to look at the photons from the two "erased" bins and select only particles that correspond one of those bins.

Let's see if I can reword that: After the particles pass through the detectors, they strike a "screen" and form a pattern without stripes. Let's set things up so that every time a particle hits the screen, it is numbered and is location across the screen is recorded. So we get a list, 1,-2.5; 2,4.6; 3,-0.2; ...
From that list alone, there are no stripes. But now we look at the photon detectors and record which detector was triggered for each particle. So we get 1:green, 2:B, 3:green, 4:yellow, ... Now we can select every element from our first list that corresponds to "green" in the second list and we get: 1:-2.5, 3:-0.2, ... Once we sort this list out, we see a clump of detections centered near the green slit - but no stripes. When we do that with yellow, we get the same result, except that the detections are centered around the yellow slit. But when we do it with A or B, we end up selecting detections that do form stripes.

What I am proposing is you do not look at the photon. You setup in such a way that the photon is traveling for 2 seconds bouncing off mirrors (completely undetected). During that 2 seconds the observer would see the location the particle hit the wall. At that point within 2 seconds (like the photon is still traveling) he could either slide in a beam splitter or slide it out.

Ok, so you see it hit the wall at a location that is consistent with either the yellow slit or erase path A. So if you allow the erasure to happen, the particle will be detected at A and will be part of the A stripe pattern. If you do not allow the erasure to happen, it will be detected at the yellow photon detector.

Hope this helps.

 

[juzzy]

You can build a simple version of this by putting polarisation filters on the slits at 90 deg to each other. Because you now have a way of knowing which slit the photon came through you get no interference (it acts like a particle). You measure by having another polarisation filter after the slits which will block photons from the slits depending on the orientation of the film. However, if you orient the film at 45 deg the interference returns. In other words it came through the slit as a particle then you erase that knowledge later down the line, in which case it now acts as a wave (came through both slits). There are a few ways to interpret this but they are all weird. You can use an interferometer to then demonstrate that it has nothing to do with the polarisation filter.
These guys will be setting up these experiments soon so you can see. You can follow their progress here...

 

[entropy1]

To me that seems to imply the 'particle' is a wave on its way to the detector, turned into a particle to be detected at the detector, turned back into a wave to go thru the slits, interfered with itself as a wave, and back into a particle once it hit the wall.

As the video shows, the 'particle' behaviour is the consequence of two waves (!) not interacting. These are probability waves, which represent particles.

How can just looking at the pattern of detector A or B cause a strip pattern? Also the video said there was a 50% it would pass thru, and 50% it would bounce off... If there is a true 50% of the photon hitting either detector, then how does it make sense that any pattern can be recovered?

The single photon emitted by the crystals is as individual photon (detected at one of the four detectors of the eraser) correlated with the (position of the) individual particle hitting the screen. This is a mathematical process.

What I am proposing is you do not look at the photon. You setup in such a way that the photon is traveling for 2 seconds bouncing off mirrors (completely undetected). During that 2 seconds the observer would see the location the particle hit the wall. At that point within 2 seconds (like the photon is still traveling) he could either slide in a beam splitter or slide it out.

I think you mean delaying of the time the eraser is activated to two seconds? That is perfectly fine and will yield the same results. In fact, that is the point of the whole experiment!

 

[georgir]

Scott, I like your explanation. And I'll join with more questions.
It is my impression that a special and interesting property of the interference pattern in double-slit experiments is that it goes to places where neither of the single-slit patterns go (or, more accurately, its intensity in some places is higher than the sum of the intensities of the two single slit patterns).
So what happens if we pick cases where the particle hits the screen in one such area - later we can decide to keep which-path information which should lead to low intensity, or erase it which should lead to high intensity?
EDIT: Ah, I should have thought a bit more myself... I guess this is resolved because there are two possible complementary interference patterns, so there isn't a high intensity point for both at the same time that we can pick.
EDIT 2: Ugh, but then why do we get two when in regular double slit we don't... my head hurts now.

 

[entropy1]

Scott, I like your explanation. And I'll join with more questions.
It is my impression that a special and interesting property of the interference pattern in double-slit experiments is that it goes to places where neither of the single-slit patterns go (or, more accurately, its intensity in some places is higher than the sum of the intensities of the two single slit patterns).
So what happens if we pick cases where the particle hits the screen in one such area - later we can decide to keep which-path information which should lead to low intensity, or erase it which should lead to high intensity?
EDIT: Ah, I should have thought a bit more myself... I guess this is resolved because there are two possible complementary interference patterns, so there isn't a high intensity point for both at the same time that we can pick.
EDIT 2: Ugh, but then why do we get two when in regular double slit we don't... my head hurts now.

That is easy: both interference patterns do add op to two single slit patterns! Also remember that the two interference patterns do partially overlap!