In DCQE you can get which way and still get interference?

[San K]

in DCQE you can get which way and still get interference?

for example when the s-photon strikes the detector ...let's assume setup is such that which-way is known

now we send the idler via erasure (after registration of s) and later when we sub-sample via co-incidence counting we get interference pattern.

thus essentially we have got which-way and interference pattern?

what is missing in this?

 

[vanhees71]

I don't know what DCQE is, but if you refer to the usual quantum-eraser setup, you never have which-way information and an interference pattern at the same time. If you chose a sub ensemble from the idler photon whose polarization state refers to a $\pi/4$ rotated direction compared to the polarization direction chosen to mark the which-way information, you don't know through which slit each of the photons has come. So you lose the which-way infromation completely, which is the "erasure" here. I you use another relative angle, you get an ensemble which is biased toward more particles coming through one slit thant the other, but at the same time the contrast of the interference pattern is less pronounced. This must be so since the which-way information, i.e., the transversal position of the photon at the slits and the phase of its wave function are incompatible observables.

 

[Demystifier]

Vanhees, DCQE = Delayed Choice Quantum Eraser.

 

[San K]

I don't know what DCQE is, but if you refer to the usual quantum-eraser setup, you never have which-way information and an interference pattern at the same time. If you chose a sub ensemble from the idler photon whose polarization state refers to a $\pi/4$ rotated direction compared to the polarization direction chosen to mark the which-way information, you don't know through which slit each of the photons has come. So you lose the which-way infromation completely, which is the "erasure" here. I you use another relative angle, you get an ensemble which is biased toward more particles coming through one slit thant the other, but at the same time the contrast of the interference pattern is less pronounced. This must be so since the which-way information, i.e., the transversal position of the photon at the slits and the phase of its wave function are incompatible observables.

Vanhees:

I am referring to the Delayed Choice Quantum Eraser setup where we:

Step 1 - first get which-way for s (corresponding to an no-interference pattern)
Step 2 - then after s is registered, use eraser on idler to erase which-way (to get an interference pattern)

thus in a sense:

you got which way (which although it erased later) and an interfernce pattern

i know that both cannot happen...so my question is:

when we use eraser, don't we still have the which-way info from step 1?

 

[unusualname]

Vanhees:

I am referring to the Delayed Choice Quantum Eraser setup where we:

Step 1 - first get which-way for s (corresponding to an no-interference pattern)
Step 2 - then after s is registered, use eraser on idler to erase which-way (to get an interference pattern)

thus in a sense:

you got which way (which although it erased later) and an interfernce pattern

i know that both cannot happen...so my question is:

when we use eraser, don't we still have the which-way info from step 1?

No, because we haven't measured the corresponding p-photon yet, if we measure it without erasing the which-way info then there will be no interference pattern in the the coincidence matches. If we erase the which-way info before the p-photon is detected then we will get an interference pattern in the coincidence matches.

You should stop thinking about photons traveling along like trains to their respective detectors, you should consider the experimental setup in its entirety at the time measurements are made.

 

[San K]

No, because we haven't measured the corresponding p-photon yet, if we measure it without erasing the which-way info then there will be no interference pattern in the the coincidence matches. If we erase the which-way info before the p-photon is detected then we will get an interference pattern in the coincidence matches.

You should stop thinking about photons traveling along like trains to their respective detectors, you should consider the experimental setup in its entirety at the time measurements are made.

I agree with what you wrote above... usually unusual name...:)

However there is still an issue to be resolved/understood.

The s-photon has struck the detector, it has registered its position (though we don't know which one it is, because we cannot filter it out of the noise till we compare with p) before p has registered its.

The position of s has been registered...it either lies on the interference pattern or not

Now how can that s-position "change" to match with p that strikes later? Or if the position does not change then what changes?

 

[unusualname]

I agree with what you wrote above... usually unusual name...:)

However there is still an issue to be resolved/understood.

The s-photon has struck the detector, it has registered its position (though we don't know which one it is, because we cannot filter it out of the noise till we compare with p) before p has registered its.

The position of s has been registered...it either lies on the interference pattern or not

Now how can that s-position "change" to match with p that strikes later? Or if the position does not change then what changes?

Nothing changes, the s-photon knows whether the p-photon has which way information or not when the s-photon hits its detector. Its bizarre I know, but the easiest way to understand it is to assume the wavefunction is non-local (in fact this is one case where the naive de Broglie Bohm pilot wave analogy is helpful). So the s-photon knows the entire experimental setup even if the eraser is placed very distant so it takes a long time for the p-photon to pass through it.

An interesting question, which has not been experimentally investigated afaik, is what about if we change the experimental setup after the s-photons has been detected (but before the p-photons have been detected), eg by removing the eraser. This requires a long distance experiment which is not currently feasible, since we'd require several hundred spdc photon pairs to be generated to get a recognisable interference pattern in the coincidence matches, and the first one will travel a huge distance by the time the last one is generated.

Experiments have been carried out between Canary Islands, and there are plans to use a satellite, so it may become feasible soon.

 

[San K]

Nothing changes, the s-photon knows whether the p-photon has which way information or not when the s-photon hits its detector. Its bizarre I know, but the easiest way to understand it is to assume the wavefunction is non-local (in fact this is one case where the naive de Broglie Bohm pilot wave analogy is helpful). So the s-photon knows the entire experimental setup even if the eraser is placed very distant so it takes a long time for the p-photon to pass through it.

An interesting question, which has not been experimentally investigated afaik, is what about if we change the experimental setup after the s-photons has been detected (but before the p-photons have been detected), eg by removing the eraser. This requires a long distance experiment which is not currently feasible, since we'd require several hundred spdc photon pairs to be generated to get a recognisable interference pattern in the coincidence matches, and the first one will travel a huge distance by the time the last one is generated.

Experiments have been carried out between Canary Islands, and there are plans to use a satellite, so it may become feasible soon.

the eraser has to be moved just a few micro inches to be in, or off, the path...not sure why that would not be possible...i mean you have about 8 ns...

 

[unusualname]

the eraser has to be moved just a few micro inches to be in, or off, the path...not sure why that would not be possible...i mean you have about 8 ns...

It would be impossible to get the timing right if you tried to do it for each SPDC emission, since the emission rate is sporadic and not predictable.

This is not like the Aspect experiment where you have a randomly moving deflector, in this case you have to be 100% sure that the eraser is in or out of the path of the incident p-photons. The only way to do this afaics would be to have the eraser + p-photon detector placed extremely distant wrt the s-photon detector, so distant that hundreds of spdc pairs will have been emitted and their s-photons all detected before a single p-photon can even reach the eraser.

btw, I don't think the outcome of the experiment is in doubt (there will still be an interference pattern in the coincidence counts even though the eraser might be removed before a single p-photon reaches it)

 

[San K]

It would be impossible to get the timing right if you tried to do it for each SPDC emission, since the emission rate is sporadic and not predictable.

This is not like the Aspect experiment where you have a randomly moving deflector, in this case you have to be 100% sure that the eraser is in or out of the path of the incident p-photons. The only way to do this afaics would be to have the eraser + p-photon detector placed extremely distant wrt the s-photon detector, so distant that hundreds of spdc pairs will have been emitted and their s-photons all detected before a single p-photon can even reach the eraser.

btw, I don't think the outcome of the experiment is in doubt (there will still be an interference pattern in the coincidence counts even though the eraser might be removed before a single p-photon reaches it)

ok, thanks for the info.

agreed, the outcome is not in doubt.


the pattern will match/correlate with whatever is done with the p-photon at the time/instance of interaction?

 

[unusualname]

ok, thanks for the info.

agreed, the outcome is not in doubt.the pattern will match/correlate with whatever is done with the p-photon at the time/instance of interaction?

Well, I think it's the exact and entire experimental setup at the time of the s-photon detection that determines what will be observed in coincidence counts. Once the s-photon has been detected the entanglement with the p-photon is broken, so the p-photon can do what it likes as long as we don't disturb its path so much that a coincidence match (with timing offset) can't be done with the corresponding s-photons.

In principle this is regardless of whether the eraser and p-photon detector is placed in another galaxy or a few meters away. But obviously it is difficult to do coincidence matching for photons traveling to another galaxy since time of travel might not be accurately predictable.

 

[San K]

Well, I think it's the exact and entire experimental setup at the time of the s-photon detection that determines what will be observed in coincidence counts. Once the s-photon has been detected the entanglement with the p-photon is broken, so the p-photon can do what it likes as long as we don't disturb its path so much that a coincidence match (with timing offset) can't be done with the corresponding s-photons.

In principle this is regardless of whether the eraser and p-photon detector is placed in another galaxy or a few meters away. But obviously it is difficult to do coincidence matching for photons traveling to another galaxy since time of travel might not be accurately predictable.

i used to think that way too..couple of weeks back...this won't explain completely because the pattern by s matches with p...

unless you go further and say...when s is detected the behavior of p is now "probabilistically" constrained/determined/predictable

 

[unusualname]

i used to think that way too..couple of weeks back...this won't explain completely because the pattern by s matches with p...

But have you experimental results to show this?

No you don't, because it is an experiment that has not been done.

unless you go further and say...when s is detected the behavior of p is now probabilistically constrained/determined/predictable

The behaviour of p is now probabilistically independent of s.

 

[San K]

But have you experimental results to show this?

No you don't, because it is an experiment that has not been done.



The behaviour of p is now probabilistically independent of s.

The DCQE (the kim-yoon paper) discussed this with the results and also other DCQE experiments/papers

the paper shows that the pattern of s will match with what is done to p

 

[unusualname]

The DCQE (the kim-yoon paper) discussed this with the results and also other DCQE experiments/papers

Yes they do, the p pattern matches the s pattern because the experimental setup is static during all the measurements.

This is trivial.

 

[San K]

Yes they do, the p pattern matches the s pattern because the experimental setup is static during all the measurements.

This is trivial.

how do you explain such a trivial thing of s pattern matching p even though both encountered "opposite/complimentary setups"? prior to striking the detectors Ds and Dp

 

[unusualname]

how do you explain such a trivial thing of s pattern matching p even though both encountered "opposite/complimentary setups"?

eh, where? What experiment do you mean? You haven't really understood my argument about the non-locality of the wavefunction. All the published experiments are quite clear in explaining the setup of the apparatus. If you think there is a problem it is because you are still thinking classically and worrying about photons traveling like "trains" through the apparatus.

You want an explanation that doesn't exist, nature is not classical, ok?

 

[San K]

Yes they do, the p pattern matches the s pattern because the experimental setup is static during all the measurements.

This is trivial.

In that case s and p are not probabilistically independent...since the pattern matches.

 

[unusualname]

In that case s and p are not probabilistically independent...since the pattern matches.

Because the p-photon's path was "decided" at the time the corresponding s-photon was measured, and we don't do anything to change this, that doesn't mean you can't change the path of the p-photon, but you need a coincidence match so mustn't interfere with its path so much that the coincidence timing is not possible.

But remember that I am talking about changing things in real-time after the s-photons have been detected, not a static experimental setup, and this experiment has never been performed.

With a static setup, even if the eraser and p-photon detector are placed in another galaxy it is easy to analyse the DCQE by assuming non-locality of the wave-function.

 

[San K]

Because the p-photon's path was "decided" at the time the corresponding s-photon was measured, and we don't do anything to change this, that doesn't mean you can't change the path

the above is what I am trying to get at...

is your opinion/take that:

the path of p is decided when s is measured?

that's what I am thinking however not all the experts on this forum would agree with that.

 

[zonde]

in DCQE you can get which way and still get interference?

for example when the s-photon strikes the detector ...let's assume setup is such that which-way is known

now we send the idler via erasure (after registration of s) and later when we sub-sample via co-incidence counting we get interference pattern.

thus essentially we have got which-way and interference pattern?

what is missing in this?

It seems that you are missing that idler photons after eraser are split in two subsamples where each subsample shows interference patter in coincidences but put together they show no interference.

 

[unusualname]

the above is what I am trying to get at...

is your opinion/take that:

the path of p is decided when s is measured?

that's what I am thinking however not all the experts on this forum would agree with that.

I think people interpret what is happening in different ways, many prefer to just discuss the measurement results which are correctly predicted by QM rather than attempt any further analysis. You can get into all kinds of muddles trying to think about what is "happening" in QM. I'm confident that a fundamentally probabilistic interpretation is the correct one, but others have their own ideas and can disagree.

 

[San K]

It seems that you are missing that idler photons after eraser are split in two subsamples where each subsample shows interference patter in coincidences but put together they show no interference.

interesting, i like your reply.

so these two sub-samples, correspond to each of the two slits? or something else?

 

[San K]

I think people interpret what is happening in different ways, many prefer to just discuss the measurement results which are correctly predicted by QM rather than attempt any further analysis. You can get into all kinds of muddles trying to think about what is "happening" in QM. I'm confident that a fundamentally probabilistic interpretation is the correct one, but others have their own ideas and can disagree.

no one is disagreeing with the idea that quantum position is fundamentally probabilistic.

the discussion is about how to explain DCQE

 

[unusualname]

no one is disagreeing with the idea that quantum position is fundamentally probabilistic.

the discussion is about how to explain DCQE

It is "explained" by adopting anyone of the many interpretations of QM, all of them "explain" the result of the experiment.

 

[zonde]

interesting, i like your reply.

so these two sub-samples, correspond to each of the two slits? or something else?

Something else. They are photons detected by D1 and photons detected by D2.

Look here:
http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm"
There at the end of the page there are two graphs - Fig. 3 and Fig. 4
These are coincidences between D0/D1 and D0/D2.
You can see that where there are peaks in one graph in other graph there are troughs.