Quantum Eraser and Its Implications

[ThomasT]

My question is whether it isn't possible that the quantum eraser experiment, rather than telling us anything about what happens when information is erased, actually reflects something unexplained and fundamental about the nature of the results of the double-slit experiment.

The net effect of erasing info is that you then have less info. Which would seem to reveal less, not more, than the case where you have more info.

What's unexplained about the double-slit experiment, the essential conundrum, is 1) if what's going through the slits is a wavefront (or sequence of wavefronts, ie., a wave train), then why the extremely localized detections, and 2) if what's going through the slits is a particle (or sequence of particles), then why the interference pattern?

Since there's no way, currently (maybe ever), to answer this question, the conundrum is spoken of in terms of wave-particle duality. A nice expression of our ignorance regarding what's actually the case wrt quantum-level phenomena.

So, yes, there's something fundamental and unexplained (perhaps unexplainable) about the nature of the results of the double-slit experiment -- and any experiment which entails individual particle detection and interference patterns is a "reflection" of this conundrum.

 

[Ken G]

The bottomline of the Scarcelli and Shih paper Jon cited

"for entangled photons it is misleading and incorrect to interpret the physical phenomena in terms of independent photons. On the contrary the concept of “biphoton” wavepacket has to be introduced to understand the non-local spatio-temporal correlations of such kind of states. Based on such a concept, a complete equivalence between two-photon Fourier optics and classical Fourier optics can be established if the classical electric field is replaced with the two-photon probability amplitude. The physical interpretation of the eraser that is so puzzling in terms of individual photons’ behavior is seen as a straightforward application of two-photon imaging systems if the nonlocal character of the biphoton is taken into account by using Klyshko’s picture."

Can you say again what that paper is? I was on a thread on here some months back where quite a few self-styled quantum physics experts told me I was nuts to suggest that delayed choice experiments had a classical analog, and indeed the only thing that made the experiment strange was the attempt to connect it with the concept of discrete particles behaving independently of each other-- a notion never encountered in classical wave mechanics.

In other words, classical wave mechanics has no difficulty with DCQE, because it doesn't need to support a particle concept there. Perhaps much, if not all, of the difficulties in interpretation DCQE stem from over-intepreting the concept of a local particle. So I'm not wild about the time-symmetric interpretation's tendency to imagine physical effects traveling backward in time to the origin, then forward along entangled paths, because it's too literal a description of a process that could easily be framed as simply us "changing our story" about what happened in some past event.

Now, as you have entangled photons, the light itself is not phase stable and therefore incoherent, so you do not get a double slit or Mach-Zehnder-interference pattern by looking at any of the two sides alone (the phase differences are different for each repeated emission of a photon pair). What is however well defined is loosely speaking the relative phase of the entangled biphoton state. The phase difference differs from photon pair to photon pair, but it is of course the same for both photons forming the pair. Now if one photon is detected at D2, that gives you some information about the phase difference for the photon pair examined right now. The phase difference will surely not have a value that will cause detections at D1 and it is more likely that it is a phase difference which is connected with a high detection probability at D2. Now that you have some information about the phase difference, you also get some information about the most probable detection positions of the entangled partner on the double slit side of the experiment.

This is the clearest description of "what is really happening" in the DSQE experiment that I"ve ever seen.

 

[ThomasT]

... if the information is previously observed, it can never be erased. Erasure works by not observing the previous information, so by not destroying various coherences, which means the information was never extracted, so it is still "in" the experiment (and has therefore not been "erased"). Hence "erasure" is quite a misleading term ...

Good point, imo.

I think we need to ask, "just what is time symmetric, the reality, or our way of interpreting the reality?" I would say it is the latter ...

Another good point, imo. As well as other good points which I won't reproduce here.

So, do so-called quantum erasure experiments inform wrt the reality underlying instrumental behavior? Or, are we still left with the fundamental conundrum illustrated by quantum double-slit experiments?

 

[Cthugha]

So if a single photon (half photon??) enters the final BS, which has a 50/50 output, why don't these patterns look like this?

Assuming the setup as shown in DevilsAvocado's post, you always have two fields arriving at the final BS. One from the red path and one from the blue path. As you cannot say that the photon that will be detected later has taken one of these paths, you need to take both of them into account which then gives the interference effect mentioned earlier. If you just send single photons down one of these paths, you will instead get a pattern like the last one you posted.

Can you say again what that paper is? I was on a thread on here some months back where quite a few self-styled quantum physics experts told me I was nuts to suggest that delayed choice experiments had a classical analog, and indeed the only thing that made the experiment strange was the attempt to connect it with the concept of discrete particles behaving independently of each other-- a notion never encountered in classical wave mechanics.

G. Scarcelli et a., "Random delayed-choice quantum eraser via two-photon imaging", The European Physical Journal D - Atomic, Molecular, Optical and Plasma Physics
Volume 44, Number 1, 167-173 (2007). You can also find it on ArXiv.

However, one should not take the classical analogy too far. The two-photon probability amplitude can be quite a non-classical entity.

To me, this means that if one were to look only in the data for D1/D2 (no path info) we will also see the interference pattern there, and in the data sole for D3/D4 (path info) the interference pattern is lost.

D1 and D2 are bucket detectors. If you just look at them, all you get is some constant count rate which will obviously cannot give any interference pattern. You really need the additional information obtained from coincidence counting at various positions of D0 to get some kind of pattern.

 

[DevilsAvocado]

D1 and D2 are bucket detectors. If you just look at them, all you get is some constant count rate which will obviously cannot give any interference pattern. You really need the additional information obtained from coincidence counting at various positions of D0 to get some kind of pattern.

Of course!

I totally missed the |x in the picture, sorry...

So what you get is a number of 'blind' (no position info) detections in D2 (only talking about D0+D2 now), and corresponding entangled twin detections in D0, where the position on the x-axis for D0 is stored. For both D0 and D2 we also store the time tag in the coincidence counter.

This enables us to later filter out those photons in D0 which corresponds to D2 and this information forms the interference pattern in graph R02.

My question remains: How on Earth could we expect to get any interference pattern out of the data in DO?? Either it should be there all the time, or nothing at all, right?? Why is there even a 'seed' for anything that later could be filtered out into an interference pattern?? And this data at D0 is recorded 8ns earlier than that of D2?? We could easily extend this to seconds, hours...

I don’t get it...


P.S. Wouldn’t we be able to tell 'which path' from the phase differences in the set up in 'my' picture in post #27?

 

[DevilsAvocado]

self-styled quantum physics experts

That’s me!

()

In other words, classical wave mechanics has no difficulty with DCQE

Now you’re dreaming again Ken G.

 

[San K]

Of course!

And this data at D0 is recorded 8ns earlier than that of D2?? We could easily extend this to seconds, hours...

I don’t get it... P.S. Wouldn’t we be able to tell 'which path' from the phase differences in the set up in 'my' picture in post #27?

hi devilsavocado

thanks for editing my other post.

Not sure what you are asking, however if I guessed correctly (as to what you are asking) then the below information might help.

When the photon at Do is recorded then the probabilities of its entangled twin photon (hitting specific locations/detector also gets fixed (the wave function collapses and the idler photon has a definite state).

Thus we can probabilistic-ally say:
given that the photon at Do is (recorded)at position x,y ...the probability of its twin photon arriving at a detector d2 is z etc.

On a separate note: can we say that an avocado is much higher on the morality hierarchy and closer to god/heaven than an advocate?...:)

 

[DevilsAvocado]

hi devilsavocado

thanks for editing my other post.

You’re welcome, hope it helped!

Not sure what you are asking, however if I guessed correctly (as to what you are asking) then the below information might help.

When the photon at Do is recorded then the probabilities of its entangled twin photon (hitting specific locations/detector also gets fixed (the wave function collapses and the idler photon has a definite state).

I see a problem right there; the idler twin photons don’t hit any "specific locations" (on the x-axis). It’s only a count of photon hits and timing, and depending on path you could tell/or not tell which slit it went thru. That’s it.

Thus we can probabilistic-ally say:
given that the photon at Do is (recorded)at position x,y ...the probability of its twin photon arriving at a detector d2 is z etc.

I don’t get it. There is no "z" for D2, only registration of the hit + time.

My confusion is due to the fact that no interference pattern is ever seen in the total pattern of signal photons at D0, meaning that if you "throw away" the other idler detectors you got nothing but 'noise' at D0.

This picture shows data for D0+D2 and D0+D3:

Now imagine you would add the data for D1 & D4 also, and remove any colors and 'guiding curves' = there’s your "noise" at D0.

So my question is: I understand that we could get a mix of interference/non-interference pattern in D0, what I don’t understand is how this could be set before any of the idler detectors has recorded anything at all. In this paper the time difference is 8ns, but this could easily be extended into 'absurdum'...

Unless I’m missing something substantial – something 'weird' is happening here.

On a separate note: can we say that an avocado is much higher on the morality hierarchy and closer to god/heaven than an advocate?...:)

Definitely, case closed! ()

 

[Ken G]

So, do so-called quantum erasure experiments inform wrt the reality underlying instrumental behavior? Or, are we still left with the fundamental conundrum illustrated by quantum double-slit experiments?

I think we are left with the fundamental conundrum, but we get insights into how to make it not a conundrum: by working on our expectations rather than our physics. In some sense, the discovery of DCQE moves us one step farther away from understanding the double slit-- it is a more sophisticated way to explore the double-slit behavior, but it results in even more sophisticated questions that we cannot answer, rather than answering the original ones! But in another way it moves us closer to not being concerned about our lack of answers, because it actually teaches us something about what an answer is, specifically, it teaches us the difference between an answer to a question that an experiment can give meaning to, and an answer to a question that we imagine has meaning but probably doesn't (because we can't find an experiment to answer it, without changing the question we are answering).

I'd say the main lesson is that what we think happened in some experiment depends on how analyze and test what happened, because many roads can lead to the same destination-- many types of empirical augmentation can be relevant to the same originally stripped-down empirical investigation, but they all might be consistent with something different happening because the original stripped-down version doesn't distinguish them.

Where all this is most relevant is in regard to the "next theory" after quantum mechanics. The key question is, is there really something wrong with quantum mechanics that needs fixing, or is there something wrong with our expectations for physics that need fixing? If we work on the latter hard enough, we might be able to get quantum mechanics to seem like a "perfect theory," in that it does everything we can expect a physics theory to do. But that doesn't mean there aren't really problems with quantum mechanics, that might make some future generation look back on us, with their new improved theory, and say "I can't believe you were really satisfied with that state of affairs." Just as we look back on those before us.

 

[Ken G]

I understand that we could get a mix of interference/non-interference pattern in D0, what I don’t understand is how this could be set before any of the idler detectors has recorded anything at all. In this paper the time difference is 8ns, but this could easily be extended into 'absurdum'...

This is the point I'm making, that when you call something "noise", you have no idea what "information" went into it. One man's noise is another man's information, the only difference is how they are slicing that information. You see "no interference pattern", and conclude that no interference occurred at all. But you can't conclude that-- you can only conclude that no interference occured in the net. It seems to me that the main message of DCQE is that if our experiment cannot tell us why or how we get the "noise" we get, then we cannot conclude it is "really noise", it might be layers and layers of highly structured information (including interference patterns) that we are simply not extracting in our experiment, whether it be a stripped-down double slit with which-way information, or a more sophisticated DCQE experiment with a delayed choice to extract which-way information. Other experiments (like a delayed choice to "leave in" the which-way information rather than extract it by destroying the necessary coherences) might be able to extract that information, but not by changing anything that happened in the first detection-- but simply by looking at it in a different way. There is no problem with looking at something in a different way any amount of time later-- what actually happened is never changing. (That's what I meant by the causes of WWI, an analogy that doesn't seem to resonate for you. Too classical for your taste, probably.)

But what gets very subtle is when "what actually happened" is itself a construct of our intelligence, based on what information we have about the apparatus, and what expectations we have for physics. So looking underneath the lesson of DCQE, I think we see a deeper message: what actually happened is what the detector showed, period-- our desire to deconstruct further what happened requires different apparatus to disentangle, but if the apparatus is different then something different happened-- even if it is something that must be consistent with the stripped-down version, it doesn't have to be the same thing as happened in the stripped-down version. Gone are the days when we could imagine our measuring apparatus was a "fly on the wall" to what is really happening-- instead, we find that our measuring apparatus is a defining element of what is really happening-- regardless of when in time those defining choices are made. Thus we should not talk so much about what actually happened, or changing what actually happened, we should simply talk about what we can say about what happened, and when we can say it!

 

[Joncon]

Assuming the setup as shown in DevilsAvocado's post, you always have two fields arriving at the final BS. One from the red path and one from the blue path. As you cannot say that the photon that will be detected later has taken one of these paths, you need to take both of them into account which then gives the interference effect mentioned earlier. If you just send single photons down one of these paths, you will instead get a pattern like the last one you posted.

Ah OK, I think I've got it now. So just to confirm I've understood - the idler photon "travels both paths" in the same way the signal photon does, and so interferes with itself?

 

[Cthugha]

P.S. Wouldn’t we be able to tell 'which path' from the phase differences in the set up in 'my' picture in post #27?

No. You start with an unknown phase. So you end up with a phase of unknown +x for going one way and unknown+y for the other which is basically still unknown. The difference y-x has some influence on the result, but gives no which way info.

I understand that we could get a mix of interference/non-interference pattern in D0, what I don’t understand is how this could be set before any of the idler detectors has recorded anything at all. In this paper the time difference is 8ns, but this could easily be extended into 'absurdum'...

I am not quite sure I get your question right, so forgive me if my answer is way off target. What you get at D0 is basically a superposition of many interference patterns which then add up to noise. As a maybe easier to grasp example imagine a sine wave. You have a device that gives out two identical sine waves as a signal. You switch it on and off and you get sine waves out every time, but the initial phase differs every time you switch it on. So the sine waves sometimes start at the highest point, sometimes at zero, sometimes at the lowest point, sometimes in between and so on, but both of the waves coming out at the same time are exactly equal. Now you switch it on and off many times and perform two different measurements on the two sine waves coming out:

1) For one of them you just add up every sine wave that comes out after each time you switch your device on. If you integrate long enough, all you will get is a straight line as for each sine wave coming out having its highest value at some point there will on average also be one sine wave coming out which is exactly out of phase so it has its lowest value at that very same point. The sum of all of them just gives the straight line.

2) For the other sine wave coming out, you do not measure the shape of the sine wave, but have some measuring device that just gives you the initial phase. Nothing more.

So now you can sort. The sum at 1) is a straight line, but you can filter using the information from 2. For example you could take all runs with the initial phase being zero. If you now take all switch on processes at 1) which correspond to the subset you chose at 2), you will get a sine wave back. If you pick the subset with a phase of pi, you will get a different sine wave. And so on and so forth.

The DCQE experiment of course adds some extras, for example non-local effects enter as you have entangled photons.

I see a problem right there; the idler twin photons don’t hit any "specific locations" (on the x-axis). It’s only a count of photon hits and timing, and depending on path you could tell/or not tell which slit it went thru. That’s it.

That's almost it. One important point is missing. If you cannot tell which slit it went through, this means that the photon will either arrive at D1 or D2. However, the probability is not 50/50. Depending on which position of D0 the corresponding signal photon is detected, it is either more likely that the idler will end up at D1 or at D2.

the idler photon "travels both paths" in the same way the signal photon does, and so interferes with itself?

Yes!

 

[Joncon]

Thanks Cthugha, that one's been bugging me for a while

Although this doesn't seem too much more mysterious than the standard double slit experiment now ...

 

[Cthugha]

Although this doesn't seem too much more mysterious than the standard double slit experiment now ...

Well, there are several implementations of DCQE out there. Some of them are more complicated, some can be broken down to be explained in rather simple terms. That depends. By the way in my opinion the standard double slit can already be pretty mysterious...

 

[Joncon]

By the way in my opinion the standard double slit can already be pretty mysterious...

Oh yeah, I agree. What I mean is that the DCQE (at least the one mentioned here) doesn't seem to add as much mystery as it appears to initally.

 

[DevilsAvocado]

I am not quite sure I get your question right, so forgive me if my answer is way off target.

I have a feeling it’s a great answer... if there’s any 'problem' it most probably is located between my ears...

I think I’m almost there, but I’m slightly 'blinded' by years of discussing EPR-Bell, rotating polarizers, and stuff, so I must ask you this, before going any further:

• Has the 8ns delay, mentioned in the paper, anything to do with anything?
  
  
• What kind of pattern would one get at D0 if we completely remove everything on the idler side (from the PS and on)?
  
  
• Is the entanglement just a 'tool' to get everything "in phase" like your "two identical sine waves"?

 

[DevilsAvocado]

This is the point I'm making, that when you call something "noise", you have no idea what "information" went into it. One man's noise is another man's information, the only difference is how they are slicing that information.

That sounds like politics to me...

Seriously, I think my way of expressing myself has caused some 'misunderstanding'. When I say "noise" I mean "pattern noise", or the lack of a "meaningful pattern". The actual measurements (counts/hits) of single photons are not noise to me (okay, there are of course real noise, but this is reduce by the coincidence counting). In other threads we have had 1,500+ comments on EPR-Bell and measurement loopholes, I don’t think I can take another round on this...

You see "no interference pattern", and conclude that no interference occurred at all.

No, that’s not what I’m saying. I see the interference pattern, clearly. However, I don’t understand how this works and how it got there. (Until Cthugha is about to save my soul ;)

Other experiments (like a delayed choice to "leave in" the which-way information rather than extract it by destroying the necessary coherences) might be able to extract that information, but not by changing anything that happened in the first detection-- but simply by looking at it in a different way.

I agree that we are not changing anything in already performed measurements, but to me this is not "the problem". Maybe I’m "looking at it" in the wrong way...

There is no problem with looking at something in a different way any amount of time later-- what actually happened is never changing.

Agree 100%

(That's what I meant by the causes of WWI, an analogy that doesn't seem to resonate for you. Too classical for your taste, probably.)

Well, here we disagree. To me there’s a HUGE difference between historical human events and physics, and that is repeatable empirical data that fits the theory, again and again and again and again... until someone come up with a brighter idea.

But the data never change, and Newton’s apple does not suspend itself in mid-air just because of Einstein, it will continue to fall the same way it always has.

Philosophy, psychology, economy, history, metaphysics, etc, don’t have this luxury and this makes it very different (to me).

But what gets very subtle is when "what actually happened" is itself a construct of our intelligence, based on what information we have about the apparatus, and what expectations we have for physics. So looking underneath the lesson of DCQE, I think we see a deeper message: what actually happened is what the detector showed, period

Yes, the detector is all we have to hold on to...

Gone are the days when we could imagine our measuring apparatus was a "fly on the wall" to what is really happening-- instead, we find that our measuring apparatus is a defining element of what is really happening-- regardless of when in time those defining choices are made. Thus we should not talk so much about what actually happened, or changing what actually happened, we should simply talk about what we can say about what happened, and when we can say it!

To us the apparatus must be real (even if theory eventually says otherwise on a more fundamental level). The value the apparatus shows must be real, and we must be able to agree on this value.

If it shows 120 photons, then it is 120 photons to everyone. Not with respect to this and that. This is what we got, period.

If we give this up, we have nothing, absolutely nothing, and science becomes 'philosophical gibberish' – "Please define 120!"


(I’m not saying every measurement is 'perfect', but I think we have to assume that they are 'reasonable' and as good as it gets, in case of real science.)

 

[Joncon]

I think I’m almost there, but I’m slightly 'blinded' by years of discussing EPR-Bell, rotating polarizers, and stuff, so I must ask you this, before going any further:

OK I'll take a punt, and treat it as a test until someone with more knowledge can come back with the proper answers

[*]Has the 8ns delay, mentioned in the paper, anything to do with anything?

I think this is just the bit which puts the "Delayed" in DCQE. So we send our photons through the double slit, but then can choose to "erase" the which path information 8ns after the actual detection.

[*]What kind of pattern would one get at D0 if we completely remove everything on the idler side (from the PS and on)?

I would imagine the pattern would be completely unchanged as there is still the possibility, in principle, to obtain which path information.

[*]Is the entanglement just a 'tool' to get everything "in phase" like your "two identical sine waves"?

Again, this just enables us to retrieve or "erase" which path information *after* the signal photons have been detected, which wouldn't be possible just using single photons.

 

[Ken G]

No, that’s not what I’m saying. I see the interference pattern, clearly. However, I don’t understand how this works and how it got there. (Until Cthugha is about to save my soul ;)

That's what I have been talking about, the answer to the "how it got there." The answer is, it was always there, it just wasn't discernable-- it looks like noise when the experiment is not able to separate it. It's a bit like a code-- if you see a coded message, it might look like complete gibberish, no rhyme or reason there and certainly not a message. But if you have the decoder, the message pops right out. You don't ask "where did this message come from, it was complete gibberish a moment ago-- have I done something that propagated a signal into the past and turned gibberish into a message being sent?" The time later that you decode the message is completely irrelevant, it could be 100 years later, because the message was always there. You didn't change it, you decoded it. That's what correlating the entanglements does. But you can "decode" it in several ways, based on what you choose to do with the entanglements. Do one thing, and the message is still gibberish-- you haven't extracted that information (you extracted some other information instead, perhaps some other message that now makes sense to you). Do something different with the entanglements, and it is like using a different cypher. The message pops right out, any amount of time later.

Well, here we disagree. To me there’s a HUGE difference between historical human events and physics, and that is repeatable empirical data that fits the theory, again and again and again and again... until someone come up with a brighter idea.

It's just an analogy, but I think it is a good one. The point is, if we stick to a "just the facts ma'am" approach, we have a bunch of events that led up to WWI, and we have a bunch of photons hitting detectors at various times. That's it, nothing more. But we are not satisfied, we want to seek reasons for why these events transpired, what was the "cause" of the presence or absence of patterns. Right away we are telling a story-- we've left the dry narrative of photons hitting detectors and people knocking off Archdukes, and we are saying "this led to that." It isn't history any more, it isn't physics any more-- yet we still call it history, and physics, because in fact this is what we want to know about history and physics. But our means of analysis has entered the picture-- we no longer are dealing in irrefutable empirical data, we have invoked a process of description, and it need not be unique.

That's the key point, the different things we do with the entangled pairs, long afterward, are like choosing different processes for describing what happened in the original data. No matter which process we choose, we still have to explain the same initial data, but the way we explain it can be very different. That's quantum erasure, and it's also historical analysis-- at least, that is the similar features to them. There are of course also differences!

But the data never change, and Newton’s apple does not suspend itself in mid-air just because of Einstein, it will continue to fall the same way it always has.

Ah, but the first half of your sentence has nothing to do with the second! The data never change, true, but whether or not the apple continues to fall is a description of what happened to the apple, it isn't data! Newton says the apple fell, Einstein says the apple ceased to be accelerated by the branch. A totally different story about what happened to the apple, both consistent with the data. So the data did not change when Einstein came along, but what "happened to the apple" certainly did change with Einstein! Because what happened is a construct, and changes in information, centuries later, can change that construct dramatically.

Philosophy, psychology, economy, history, metaphysics, etc, don’t have this luxury and this makes it very different (to me).

But physics is actually not so different-- it doesn't have that luxury either. All that is different is the precision that is possible, and the scale where we encounter just where that "luxury" breaks down.

To us the apparatus must be real (even if theory eventually says otherwise on a more fundamental level). The value the apparatus shows must be real, and we must be able to agree on this value.

Yes, the value the apparatus shows-- but not why it shows it. Not whether or not interference occured, not which slit the particle went through. Those are not part of the data until we make the choice to make them part of the data-- at which point our description of what happened also changes, even long after the original experiment in which the happening happened.

If it shows 120 photons, then it is 120 photons to everyone.

Certainly, but that's not "what happened". We don't say "120 photons hit a detector in this here pattern", we say "no two-slit interference". The latter is not 120 photons, it is a kind of judgement about what happened, and that's what quantum erasure shows is not a unique thing, and can change a century later without actually changing anything at all but our mode of analysis of the original happening. This is no minor point-- quantum mechanics is extending to physics the much more general rule that our descriptions of what happened are dependent on our means of establishing what happened.

 

[DevilsAvocado]

OK I'll take a punt, and treat it as a test until someone with more knowledge can come back with the proper answers

Thanks, much appreciated!

I think this is just the bit which puts the "Delayed" in DCQE. So we send our photons through the double slit, but then can choose to "erase" the which path information 8ns after the actual detection.

Okay... interesting... but if I got this right; the entanglement does not affect the outcome at D0 one bit, right? So what’s in fact is 'delayed' is the choice to measure "which path", or not, in "cloned twin beam", right?

If I understand this right, non-locality is not the crucial thing here, but a "clone copy" of the signal beam, right?

I would imagine the pattern would be completely unchanged as there is still the possibility, in principle, to obtain which path information.

This is, I think, the 'Gordian Knot' to me... Why do we get a mixture of interference/non-interference pattern in D0? What causes it? There’s no "flip-flopping gate" at the double slit, is it? I don’t get it? In a normal experiment we would get an interference pattern or no interference pattern, not both, right??

Again, this just enables us to retrieve or "erase" which path information *after* the signal photons have been detected, which wouldn't be possible just using single photons.

This I get, but as you see in Cthugha answer we could also do it with another mechanism of two "cloned waves"... I think I’m 'over-interpreting' the part of entanglement in this experiment... I don’t know...


(Ken G, get back later...)

 

[Cthugha]

Okay... interesting... but if I got this right; the entanglement does not affect the outcome at D0 one bit, right? So what’s in fact is 'delayed' is the choice to measure "which path", or not, in "cloned twin beam", right?

If I understand this right, non-locality is not the crucial thing here, but a "clone copy" of the signal beam, right?

Well, entanglement has two main effects. One is that that the two entangled particles do indeed bahave like cloned copies (or shifted copies or whatever one likes to call it) with the main point being that you can tell what the entangled partner will do if you know what the first particle does. The second point in entanglement is that these properties are not imprinted from the beginning in a hidden variable-like fashion, but the state of both is fixed when the first measurement occurs which implies nonlocality. The latter is what is tested in Bell tests.

So for the experimental outcome it is in fact only the first property which matters. There is information that can only be accessed when detecting both entangled particles and "matching up". Non-locality does not really matter in terms of the outcome, but in terms of the interpretation of the results. For example the standard DCQE experiment could be changed such that the delay between detections of signal and idler becomes large and one could perform the measurements in such a fashion that Bell inequalities are tested.

Why do we get a mixture of interference/non-interference pattern in D0? What causes it? There’s no "flip-flopping gate" at the double slit, is it? I don’t get it? In a normal experiment we would get an interference pattern or no interference pattern, not both, right??

In a normal double slit experiment the pattern you will see depends on the geometry of your experiment. If your light source is for example not exactly centered between the two slits, you will get a slightly different pattern as the distances between the source and the two slits are now different. As you move the source around, you will get different patterns. So if you now place several light sources at different positions you will now get a superposition of all of these patterns. If you have enough sources the superposition will be no pattern at all. This is foe example the same reason why the interference pattern disappears in a common double slit experiment if you place your light source too close to the slits. It will then reappear as you increase the distance between the slits and the source.

Now in the DCQE you have a similar setting (I am referring to figure one in the paper by Kim, Kulik, Shih and Scully). You have two atoms A and B placed at the slits which could emit entangled photon pairs. This process is completely random and one cannot distinguish from which atom some certain phton pair comes (unless you have detections at D3/D4). However, the two atoms are not synchronized. The phase of the light fields emitted from the two atoms is pretty much random with respect to each other and it also fluctuates randomly. This is pretty much like having a like source emit a single photon in the common double slit experiment and then moving it somewhere else, emitting another photon, etc which will add up to no interference pattern at all. However, if you note the position of the source for each photon and afterwards just pick a subset of detections corresponding to the source being at the same position, you will find some pattern. If you pick a different subset corresponding to a different position, you will find some different pattern.

In DCQE the position of the peaks in the double slit pattern will depend on the relative phase between the fields emitted by the atoms, so you do not see a single interference pattern, but a superposition of many which add up to no interference pattern. However, you can find any of these interference patterns by just picking those detections that correspond to some phase difference between the fields. That picking is now done using the other entangled partner. Depending on the relative phase that partner is more likely to be detected at D1 or D2, so you get a fringe/antifringe pattern when the coincidence counts D0/D1 or D0/D2 are considered.

 

[DevilsAvocado]

One is that that the two entangled particles do indeed bahave like cloned copies
...
So for the experimental outcome it is in fact only the first property which matters.
...
Now in the DCQE you have a similar setting (I am referring to figure one in the paper by Kim, Kulik, Shih and Scully). You have two atoms A and B placed at the slits which could emit entangled photon pairs. This process is completely random and one cannot distinguish from which atom some certain phton pair comes (unless you have detections at D3/D4). However, the two atoms are not synchronized. The phase of the light fields emitted from the two atoms is pretty much random with respect to each other and it also fluctuates randomly. This is pretty much like having a like source emit a single photon in the common double slit experiment and then moving it somewhere else, emitting another photon, etc which will add up to no interference pattern at all. However, if you note the position of the source for each photon and afterwards just pick a subset of detections corresponding to the source being at the same position, you will find some pattern. If you pick a different subset corresponding to a different position, you will find some different pattern.

In DCQE the position of the peaks in the double slit pattern will depend on the relative phase between the fields emitted by the atoms, so you do not see a single interference pattern, but a superposition of many which add up to no interference pattern. However, you can find any of these interference patterns by just picking those detections that correspond to some phase difference between the fields. That picking is now done using the other entangled partner. Depending on the relative phase that partner is more likely to be detected at D1 or D2, so you get a fringe/antifringe pattern when the coincidence counts D0/D1 or D0/D2 are considered.

Yay! I think I’ve got it! Many THANKS!

I’ve been completely lost... (as you may have noticed). All the time I have assumed that an interference pattern is created already directly after the slits, that is "picked up" by the BBO to then be (randomly) 'transformed' into two entangled pair... god I’m stupid...

The distance is obviously too short for this to happen... the 'interference' is instead created by the random and relative phase (shift) in the photons emitted from the BBO. Please, tell me I got this right??

I was thinking... is it ever possible to move the BBO to be placed before the slits? And then use a (P)BS after the slits to separate signal & idler, and thereafter do "fancy measurements" on the two entangled twins?

Or is it already a dead end when hitting the BS??

 

[San K]

You’re welcome, hope it helped!

I see a problem right there; the idler twin photons don’t hit any "specific locations" (on the x-axis). It’s only a count of photon hits and timing, and depending on path you could tell/or not tell which slit it went thru. That’s it.

Sorry replace the concept "specific location" with detector. Once
Signal photon has stuck the detector the probability of idler hitting a particular detector is calculable (based on location of signal photon on the screen)

 

[Quaman]

Hi!
I have not finished reading through the whole thread, so please excuse me if I am asking something that was already covered. Here are my questions related to a DCQE setup:
1/a) If which-path information is observed (but not recorded!) by a non-human observer at D3 and D4, will or will not the interference pattern collapse at D0 when D0 is observed by a human?
1/b) If which-path information is observed and recorded by a non-human observer at D3 and D4, will the interference collapse at D0 when observed by a human? Let's assume that no human ever sees that record. It is clear that if the record later gets observed by a human, D0 will not show interference pattern.
1/c) What if the record made by a non-human observer at D3 and D4 is irrevocably deleted before any human can observe it? Will the interference collapse at D0 when D0 is observed by a human? If yes, why? I think it should not, because the entangled pair (idler) will never be observed.
2) In the same dcqe setup if which-path information is observed by Alice at D3 and D4, but does not have the possibility to tell Bob about it (eg. dies before she can do so), will Bob at D0 see interfence? This is somewhat similar to 1/c, but the difference is that the record was made by a human in this case and that record unfortunately also got 'irrevocably erased'.
I am pretty sure no-one has ever carried out experiment no. 2), as it is apparently against federal laws :) But I expect it should give the same results as 1/c, right?

If you know that an experiment was carried out before with any of the above or similar setup, I would be really glad if you could share the link to the paper (or at least to its abstract) here :) Thank you!

 

[Cthugha]

Same answer to all questions: The presence or absence of humans has absolutely nothing to do with the outcome. The meaning of observing in physics does not require any consciousness. Any irreversible interaction is sufficient.

 

[Ken G]

The meaning of observing in physics does not require any consciousness.

And I'd like to interject one caveat that is seemingly picky but I think is actually quite important: meaning always requires consciousness. But your point, I believe, is that consciousness does not enter at the same physical level as the other elements of the experiment, the detector, the slits, etc., nor does it enter at the level of the theory, the Hilbert space or the operators, etc. So there is no "piece of the apparatus" we need to call conscious, and there is no term in the equation we might want to call the "effect of consciousness." All the same, I think quantum mechanics is one of the places where we are forced to come to terms with the fact that the "fingerprints" of our consciousness are all over what we are doing there. A universe without consciousness has no quantum mechanics in it at all, nor anything recognizable as physics, because as Bohr put it, physics is about what we can say about nature. And by "we", he means our consciousness/perceptions/intelligence, etc.. Not our measuring apparatus, which is quite mute.

. Any irreversible interaction is sufficient.

This provides me with a more concrete example of my point-- the fundamental interactions in physics are generally time reversible, so "irreversibility" is already a concept that appears, not in the physical interactions, but at the level of our mental processing of those interactions. Hence, there is no such thing as "irreversible" in the first place, without consciousness (and here I make no effort to parse between what it means to be conscious versus intelligent or able to perceive, as I don't think we have any kind of precise language to use to parse those notions).

 

[DevilsAvocado]

... A universe without consciousness has no quantum mechanics in it at all, nor anything recognizable as physics, because as Bohr put it, physics is about what we can say about nature. And by "we", he means our consciousness/perceptions/intelligence, etc.. Not our measuring apparatus, which is quite mute.

I don’t recall Bohr or the Copenhagen interpretation ever saying anything about consciousness... this sounds like the von Neumann/Wigner interpretation.

But I’m curious; exactly how do you explain the evolution of the universe and ΛCDM if consciousness has any crucial role in this evolution? For example red-shift of the CMB, in relation to "our consciousness", is something that you could describe in detail. Thanks.

I’m all ears!

 

[DevilsAvocado]

It's just an analogy, but I think it is a good one. The point is, if we stick to a "just the facts ma'am" approach, we have a bunch of events that led up to WWI, and we have a bunch of photons hitting detectors at various times. That's it, nothing more.

I don’t agree that "events that led up to WWI" could be compared to physics and the scientific model. If this analogy was ever to operate as a theory that is refutable, you need two things:

1. A mathematical description of the human mind, which we don’t have.
   
   
2. The ability to experimentally repeat "events that led up to WWI" as many times you like, which is impossible.

Ah, but the first half of your sentence has nothing to do with the second! The data never change, true, but whether or not the apple continues to fall is a description of what happened to the apple, it isn't data! Newton says the apple fell, Einstein says the apple ceased to be accelerated by the branch. A totally different story about what happened to the apple, both consistent with the data. So the data did not change when Einstein came along, but what "happened to the apple" certainly did change with Einstein! Because what happened is a construct, and changes in information, centuries later, can change that construct dramatically.

You are mixing apples and oranges. As I said Newton’s apple does not suspend itself in mid-air just because of Einstein, it will continue to fall the same way it always has, i.e. to the ground.

This is the empirical fact that no new theory could ever change - apples will not suspend in mid-air, period.

The explanation on why and how apples fall could of course change in the future.

 

[Cthugha]

And I'd like to interject one caveat that is seemingly picky but I think is actually quite important: meaning always requires consciousness. But your point, I believe, is that consciousness does not enter at the same physical level as the other elements of the experiment, the detector, the slits, etc., nor does it enter at the level of the theory, the Hilbert space or the operators, etc. So there is no "piece of the apparatus" we need to call conscious, and there is no term in the equation we might want to call the "effect of consciousness."

I have already stated elsewhere (was it inside this thread? I do not know) that one cannot rule out the influence of a human looking at some experiment with absolute certainty as there is no possibility to test this experimentally. You cannot find out what happens if you never find out what happens. However, this also makes it a non-scientific question for exactly the same reason. It is interesting from a philosophical point of view, though.

This provides me with a more concrete example of my point-- the fundamental interactions in physics are generally time reversible, so "irreversibility" is already a concept that appears, not in the physical interactions, but at the level of our mental processing of those interactions.

I doubt that. Irreversible interactions occur in every measurement. Whenever some superposition of states ends up in an eigenstate. Whenever entropy changes. In the context of my post you quoted I just wanted to point out that there is some difference between processes like inserting a wave plate in a light beam and rotating its polarization on the one hand and processes like absorbing a photon at one certain position. The first is reversible and does not constitute a measurement. The second is usually irreversible, collapses a wave function and constitutes a measurement.

Irreversibility shows up in entropy and such things as the arrow of time. While it is true that many underlying fundamental processes may be reversible, there is nevertheless a statistical prevalence for things to happen in a certain way (see Feynman's famous broken cup example) in the sense of statistical mechanics. I do not think that mental processing is a necessity for that concept from a physics point of view. Philosophical discussions are of course a different topic, bit from my point of view rather distracting when discussing real experiments.

 

[Maui]

But I’m curious; exactly how do you explain the evolution of the universe and ΛCDM if consciousness has any crucial role in this evolution? For example red-shift of the CMB, in relation to "our consciousness", is something that you could describe in detail. Thanks.

I’m all ears!

Something tells me a non-realist would not consider the CMB and the Big Bang the ultimate explanation of existence and reality. I am not sure they believe in reality at all(and that's not the craziest I've seen in the interpretations war)