# I Kim's 1999 delayed-choice quantum eraser experiment

#### skewzme

Summary
Did this experiment prove wave function collapse is not the result of any physical interactions at the point of measurement?
As I understand the delayed-choice quantum eraser experiment, first performed by Kim in 1999, entangled photons are used to determine which-way data, and the which-way data is obtained by virtue of where the entangled particles land, as opposed to using a measuring device that may be collapsing the wave by virtue of any physical interactions.
The experiment showed wave functions remained intact when which-way information was not available, and were collapsed when which-way data was known.

My question: Did this prove wave collapse cannot be attributed to any kind of physical interaction between a measuring device and the photons?
If not, why?

Related Quantum Physics News on Phys.org

#### vanhees71

Gold Member
Wave-function collapse is a very misleading concept and in fact has never ever been observed. What has been observed is clearly stated in the scientific paper of said experiment:

#### skewzme

Wave-function collapse is a very misleading concept and in fact has never ever been observed. What has been observed is clearly stated in the scientific paper of said experiment:

For clarity, I should have said "particles were observed". That was the context I intended.

#### vanhees71

Gold Member
Have you read the paper I linked in #2? At the arXiv you can download it (legally) for free. What's you specific question? Since in the entire paper nothing about any whatever collapse is mentioned, and rightfully so, because nothing has collapsed in this experiment (neither in any other experiment), I don't know, what you want to know about the experiment.

Another, even more clear experiment is the delayed-choice quantum eraser experiment by Walborn et al. There it's clear that you really can erase the which-way information and getting an interference pattern as long after the experiment as you want. You only have to choose, whether you want to read out of the measurement protocol the locations of all photons (which in principle have which-way information and thus don't show an double-slit interference pattern) or that of one of two complementary subsets, each showing an interference pattern, and for which the in-principle-available which-way information has been erased.

I think this version of the experiment makes it even more clear that there's no collapse of whatsoever involved. It's just choosing subsets of the measured photon ensemble after all registration procedures have long been finished:

https://arxiv.org/abs/quant-ph/0106078

Both experiments are of course equivalent and use the long-range correlations of polarization-entangled two-photon states, nowadays available through the use of non-linear optics, as described in the paper by Kim et al.

#### skewzme

Haven’t read it yet but plan to. The delayed choice quantum eraser experiment was first done in 1999 and is the same experiment you mention, by name anyway.
The experiment seems to demonstrate the collapse has nothing to do with measurement devices, since none are used.

#### skewzme

Have you read the paper I linked in #2? At the arXiv you can download it (legally) for free. What's you specific question? Since in the entire paper nothing about any whatever collapse is mentioned, and rightfully so, because nothing has collapsed in this experiment (neither in any other experiment), I don't know, what you want to know about the experiment.

Another, even more clear experiment is the delayed-choice quantum eraser experiment by Walborn et al. There it's clear that you really can erase the which-way information and getting an interference pattern as long after the experiment as you want. You only have to choose, whether you want to read out of the measurement protocol the locations of all photons (which in principle have which-way information and thus don't show an double-slit interference pattern) or that of one of two complementary subsets, each showing an interference pattern, and for which the in-principle-available which-way information has been erased.

I think this version of the experiment makes it even more clear that there's no collapse of whatsoever involved. It's just choosing subsets of the measured photon ensemble after all registration procedures have long been finished:

https://arxiv.org/abs/quant-ph/0106078

Both experiments are of course equivalent and use the long-range correlations of polarization-entangled two-photon states, nowadays available through the use of non-linear optics, as described in the paper by Kim et al.

#### vanhees71

Gold Member
What's about this youtube video? I'm not sure whether this will help you to understand quantum mechanics. The only way to understand quantum mechanics is to learn the math. There's a point in physics where any non-mathematical treatment fails. I think that's already the case in classical physics (the latest if you want to understand General Relativity), but for sure it's the case for quantum theory.

A very good introduction is the volume in Susskind's "The theoretical minimum" series on quantum mechanics:

https://www.amazon.com/dp/0465062903/?tag=pfamazon01-20

#### skewzme

I offered the video to vanhees71 to confirm whether he and i were talking about the same experiment. It seems were are, but he states no wave collapse occurs in this experiment. The video confirms it certainly does, depending on whether which way data is observed, or left unobserved.

With respect to understanding the math, that is not likely for me as that is my my forte.

Assuming you do understand the math, can you answer the initial question I posed?
That is, does this experiment confirm the disappearance of the wave pattern as NOT being the result of any measuring apparatus interference?

#### vanhees71

Gold Member
A popular-science video is not a good starting point to understand quantum mechanics. The idea of a "collapse of the state" was a concept (in my opinion a misconception) of some flavors of the Copenhagen interpretation.

The collapse hypothesis, however, contradicts fundamental properties of relativity (causality), and thus local relativistic QFT, of which QED is the paradigmatic example. Particularly when you deal with photons, it's inconsistent with the underlying theory describing all experiments with light quanta correctly.

That's why your quesion in #1 doesn't make sense to start with. You cannot empirically prove a self-contradictory misconception. What the experiment shows is that the correlations described by the entanglement of the photons created by parametric downconversion are the correct description of the outcome of the experiment. These and similar experiments, particularly those showing that the violation of Bell's inequality is as predicted by quantum theory and thus ruling out local deterministic hidden-variable interpretations of quantum theory.

#### skewzme

A popular-science video is not a good starting point to understand quantum mechanics.
Who exactly does understand quantum mechanics? While I concede my inability to have that discussion on a purely mathematical basis, it would appear discussions among those who do are still unable to reach a consensus of understanding either:
If as you say my quesion in #1 doesn't make sense to start with because it is based on a self-contradictory misconception, allow me to propose the question in a slightly different way. Does the experiment demonstrate that the appearance, or lack thereof, of a wave-like pattern is based on knowledge of the outcome, as opposed to any physical interactions between the "particles" and a measuring device ?
If my rephrased question is still based on my own misconceptions, I would appreciate any attempt you care to offer to explain why in laymen's terms & concepts.
Thank you.

#### kurt101

The collapse hypothesis, however, contradicts fundamental properties of relativity (causality), and thus local relativistic QFT, of which QED is the paradigmatic example. Particularly when you deal with photons, it's inconsistent with the underlying theory describing all experiments with light quanta correctly.
I am no expert on QM, but I believe you are wrong to say the collapse hypothesis is wrong and contradicts either relativity or QFT. I have not seen any experimental evidence to support your assertion. What I suspect is that you are stating your interpretative belief of QM and not anything based on fact. Of course, I would hope that someone else on this forum who is more knowledgeable would speak up if this was the case. Based on this reason alone, I have doubt, but last time someone else said something similar to what you said and I asked for evidence to back up their claim, they never responded.

And so we are clear, my understanding of collapse with entangled photons is when you measure one of the photons, that interaction instantly (FTL) affects the other photon's state.

Given that you can't know the state of either of the entangled photons until one of them interacts, how can a collapse hypothesis possibly contradict relativity?

How does QFT/QED refute the collapse hypothesis? QFT/QED speaks to the probability. The collapse hypothesis speaks to what is actually happening. These seem to have no contradiction with each other.

#### skewzme

my understanding of collapse with entangled photons is when you measure one of the photons, that interaction instantly (FTL) affects the other photon's state.
This is true. But as I understand the purpose of Kim's 1999 experiment, it wasn't to test whether entangled photons affect their partner's state. Entangled particles were used to as the method of determining which way data in order to eliminate detectors as a possible reason for the collapse, through physical interactions. Someone correct me if I am wrong.

#### Cthugha

This is true. But as I understand the purpose of Kim's 1999 experiment, it wasn't to test whether entangled photons affect their partner's state. Entangled particles were used to as the method of determining which way data in order to eliminate detectors as a possible reason for the collapse, through physical interactions. Someone correct me if I am wrong.
That is incorrect. If you have a look at the references in the paper and go back to the 1982 delayed choice proposal by Scully and Drühl and even further back in time, you will notice that the main motivation for this study was a question that goes back to the very early days of quantum mechanics (and was more or less already solved at that time). The question was whether uncertainty or complementarity was more fundamental. Back in the early days of QM, some people argued that there was no way to measure the which-way information at the double slit without significantly perturbing the particle in question. So people were thinking about ways to get this information without perturbing the particle and came up with the idea of using entanglement. However, by that time, the problem was already more or less sorted out and the concept of complementarity was already outdated.

And no, the experiment does not demonstrate that the appearance, or lack thereof, of a wave-like pattern is based on knowledge of the outcome. Essentially you get lots of photons that will form different interference patterns in the arm containing the scanning detector. These detection events are correlated with detection events in the other arm of the experiment, where the entangled partner will be detected. As there are filters here, you will always only detect some subensemble of all photons arriving here. By doing coincidence counting, you will therefore be able to perform some kind of non-local filtering process of the photons on the side with the scanning detector, which allows you to also pick some subensembles of the detection events here. Now, depending on the subensemble you picked in the entangled partner arm, you may or may not be able to find an interference pattern in the coincidence counts. Obviously, it will be the subensembles, which do not contain any which-way information, that will yield the interference patterns. However, this will not depend on any knowledge of the experimenter. The experiment is essentially just the non-local version of the standard double slit experiment done with an incoherent source. An incoherent light source will not give you an interference pattern (or rather dozens of them, which add up to no pattern at all). Now, if you put a narrow pinhole before the double slit, the remaining beam has enhanced spatial coherence and you will see a nice double slit interference pattern. Now, the DCQE basically uses an entangled source of incoherent light and instead of placing the pinhole in front of the double slit, you place it in the arm of the entangled partner. If you now do coincidence counting, this will give you a subset of photons on the double slit side that will have the same characteristics as those that would have passed through the narrow pinhole and you will get the interference pattern. However, it will appear ONLY in coincidence counting because it is always just this filtered subset that shows this interference pattern. You do not change the past or do any of the strange stuff proposed by those gazillions of questionable youtube videos. Please do not cite them. When I give lectures on the DCQE, I explicitly mention the video you posted as an example for really bad pop-sci stuff that describes things incorrectly. This is not a valid source for anything.

#### skewzme

That is incorrect. If you have a look at the references in the paper and go back to the 1982 delayed choice proposal by Scully and Drühl and even further back in time, you will notice that the main motivation for this study was a question that goes back to the very early days of quantum mechanics (and was more or less already solved at that time). The question was whether uncertainty or complementarity was more fundamental. Back in the early days of QM, some people argued that there was no way to measure the which-way information at the double slit without significantly perturbing the particle in question. So people were thinking about ways to get this information without perturbing the particle and came up with the idea of using entanglement. However, by that time, the problem was already more or less sorted out and the concept of complementarity was already outdated.

And no, the experiment does not demonstrate that the appearance, or lack thereof, of a wave-like pattern is based on knowledge of the outcome. Essentially you get lots of photons that will form different interference patterns in the arm containing the scanning detector. These detection events are correlated with detection events in the other arm of the experiment, where the entangled partner will be detected. As there are filters here, you will always only detect some subensemble of all photons arriving here. By doing coincidence counting, you will therefore be able to perform some kind of non-local filtering process of the photons on the side with the scanning detector, which allows you to also pick some subensembles of the detection events here. Now, depending on the subensemble you picked in the entangled partner arm, you may or may not be able to find an interference pattern in the coincidence counts. Obviously, it will be the subensembles, which do not contain any which-way information, that will yield the interference patterns. However, this will not depend on any knowledge of the experimenter. The experiment is essentially just the non-local version of the standard double slit experiment done with an incoherent source. An incoherent light source will not give you an interference pattern (or rather dozens of them, which add up to no pattern at all). Now, if you put a narrow pinhole before the double slit, the remaining beam has enhanced spatial coherence and you will see a nice double slit interference pattern. Now, the DCQE basically uses an entangled source of incoherent light and instead of placing the pinhole in front of the double slit, you place it in the arm of the entangled partner. If you now do coincidence counting, this will give you a subset of photons on the double slit side that will have the same characteristics as those that would have passed through the narrow pinhole and you will get the interference pattern. However, it will appear ONLY in coincidence counting because it is always just this filtered subset that shows this interference pattern. You do not change the past or do any of the strange stuff proposed by those gazillions of questionable youtube videos. Please do not cite them. When I give lectures on the DCQE, I explicitly mention the video you posted as an example for really bad pop-sci stuff that describes things incorrectly. This is not a valid source for anything.
And because it was a PBS video, I assumed it was credible. Silly me. I truly appreciate that explanation. I need to ponder it more, And I may have some additional questions for you. Thank you very much

#### PeterDonis

Mentor
And because it was a PBS video, I assumed it was credible.
Pop science videos should never be assumed to be good sources for actually learning science. There are good reasons for PF's policy on acceptable sources.

#### skewzme

Well, you’ll have to define pop videos. It didn’t come off my Saturday morning lineup

#### vanhees71

Gold Member
Who exactly does understand quantum mechanics? While I concede my inability to have that discussion on a purely mathematical basis, it would appear discussions among those who do are still unable to reach a consensus of understanding either:
If as you say my quesion in #1 doesn't make sense to start with because it is based on a self-contradictory misconception, allow me to propose the question in a slightly different way. Does the experiment demonstrate that the appearance, or lack thereof, of a wave-like pattern is based on knowledge of the outcome, as opposed to any physical interactions between the "particles" and a measuring device ?
If my rephrased question is still based on my own misconceptions, I would appreciate any attempt you care to offer to explain why in laymen's terms & concepts.
Thank you.
Quantum mechanics is understood by every physicist with a university degree (at least that should be the case since passing the exams on several quantum-theory related lectures is necessary to obtain that degree).

Of course, if you ask for opinions on the socalled "interpretation of quantum theory" which deals with philosophical rather than scientific issues you get a plethora of opinions, and I guess on this level there are as many "interpretations" of the "meaning" of quantum theory as there are physicists (maybe sometimes even more since on such sunsolid ground your opionion easily fluctuates between various possibilities, and I'm not excluding myself from this sin).

Now, as far as the scientific issues are concerned, there's no such uncertainty at all. The natural sciences are about objective observable and quantifiable phenomena of nature, and this nice experiment is no exception. What's done is very clearly described in the scientific paper: entangled two-photon states are prepared using two berefringent BBO crystals at two places $A$ and $B$. One photon of each pair is registered by a detector $D_0$ at an adjustable position $x_0$ and the other photon of the pair is registered in one of 4 detectors $D_1,\cdots,D_4$. Now due to the entanglement of the photons and the setup with beam splitters you have the following

If there is a coincident registration of photon 1 in $D_0$ and

(a) photon 2 in $D_1$ or $D_2$ there's no way to know which path photon 1 has taken (i.e., whether photon 1 came from region A or B of the BBO crystal).

Now consider ensembles of such prepared photon pairs: Then registering all those photons 1 as a function of $x_0$, whose partner photon 2 has hit $D_1$ gives an interference pattern.

Registering all photons 1 as a function of $x_0$, whose partner photon 2 has hit $D_2$ gives the same interference pattern with some shift of the fringes, which is well understood from the unitarity of the transfer matrices describing the workings of the beam splitters.

(b) and photon 2 in $D_3$ or $D_4$. Then due to the entanglement of the photons in each pair there's which-way information about photon 1.

Thus registering all those photons 1 as a function of $x_0$, whose partner photon 2 has hit $D_3$ leads to no double-slit inteference pattern, because then it's known that both photon 1 and photon 2 must have come from the BBO at position B, i.e., there's clear which-way information and thus there's only a single-slit interference pattern.

Thus registering all those photons 1 as a function of $x_0$, whose partner photon 2 has hit $D_4$ also leads to no double-slit inteference pattern, because then it's known that both photon 1 and photon 2 must have come from the BBO at position A, i.e., there's clear which-way information and thus there's only a single-slit interference pattern.

The choice of whether you want to have two-slit interference of which-way information is delayed by the fact that the detectors $D_1,\ldots,D_4$ are much farther away from the BBO than detector $D_0$, i.e., photon 1 is always much earlier detected than photon 2, but then choosing to consider only those photons 1 as a function of position of $D_0$, whose partner photon 2 has been detected by $D_1$ (or $D_2$) leads to the loss of which-way information about both photon 1 and 2 and thus the appearance of a (shifted) two-slit interference pattern, while looking only at photons 1, whose partner photon has been registered at $D_3$ (or $D_4$) implies that we know which paths both photon 1 and photon 2 have taken, and thus there's no double-slit interference pattern (but only a single-slit interference pattern) left.

Note also that just registering all photons 1 leads to a completely flat distribution.

As is shown in the paper, the experiment agrees with this predictions of QED. The only interpretation needed here is the unanimous meaning of the quantum formalism for real-world experiments as is known as the minimal statistical interpretation of the states (aka Born's rule).

#### PeterDonis

Mentor
you’ll have to define pop videos
If it's not a video of a professor giving an actual lecture at an actual for-credit course at an actual university, it's a pop science video. And even if it is, it might be tricky to use as a source by itself because there could be a lot of context missing and because what a professor says in a class is not subject to peer review the way textbooks and peer-reviewed papers are. That's why the PF policy on acceptable sources doesn't mention videos at all; it talks about textbooks and peer-reviewed papers.

#### skewzme

Quantum mechanics is understood by every physicist with a university degree (at least that should be the case since passing the exams on several quantum-theory related lectures is necessary to obtain that degree).
Can you then help put Richard Feynmans famous quote in context?
“If you think you understand quantum mechanics, you don't understand quantum mechanics". .
Is it that we have come to understand it since he made that statement, or are you taking issue with his opinion ?

#### Lord Jestocost

Gold Member
2018 Award
“To understand” is regarding quantum mechanics a doubtful verb. In chapter 1 of “The Feynman Lectures on Physics, Volume III”, Richard Feynman puts what he means in the following way (when introducing the double slit experiment):

“Because atomic behavior is so unlike ordinary experience, it is very difficult to get used to, and it appears peculiar and mysterious to everyone—both to the novice and to the experienced physicist. Even the experts do not understand it the way they would like to, and it is perfectly reasonable that they should not, because all of direct, human experience and of human intuition applies to large objects. We know how large objects will act, but things on a small scale just do not act that way. So we have to learn about them in a sort of abstract or imaginative fashion and not by connection with our direct experience.

In this chapter we shall tackle immediately the basic element of the mysterious behavior in its most strange form. We choose to examine a phenomenon which is impossible, absolutely impossible, to explain in any classical way, and which has in it the heart of quantum mechanics. In reality, it contains the only mystery. We cannot make the mystery go away by “explaining” how it works. We will just tell you how it works. In telling you how it works we will have told you about the basic peculiarities of all quantum mechanics.” [Italic in original]

#### skewzme

Perhaps it appears peculiar and mysterious to everyone—both to the novice and to the experienced physicist, because there is still some significant principle or element missing.
Being able to predict outcomes doesn't mean it's entirely understood. I can predict outcomes involving the mechanics of my car. If I step on the accelerator I understand it will accelerate. But that is not to say I understand the inner workings that make it accelerate when I step on the accelerator.
I say this cautiously, as a lay person. But sometimes it sounds to me that experienced physicists may be making that mistake in concluding they "understand" it, simply because certain outcomes can be predicted mathmatically.
And not to ruff any feathers, but I also sense in some academic circles a feeling they are above questioning, or their intellectual prowess too high to be challenged by mere mortals. That confidence may be blinding them from some otherwise obvious considerations. Thanks

#### weirdoguy

And not to ruff any feathers, but I also sense in some academic circles a feeling they are above questioning, or their intellectual prowess too high to be challenged by mere mortals. That confidence may be blinding them from some otherwise obvious considerations.
It's funny that every single person who says such things has no proper knowledge of actual physics.

#### PeterDonis

Mentor
Who exactly does understand quantum mechanics?
"Understanding" as you're using the term here (as shown by subsequent posts) is not a physics term; we could go round and round forever arguing about what it means and get nowhere. There's no point in doing that.

I also sense in some academic circles a feeling they are above questioning, or their intellectual prowess too high to be challenged by mere mortals. That confidence may be blinding them from some otherwise obvious considerations.
You have certainly given no evidence whatever of that in this thread. All you have shown is that you are unwilling to make the effort to read actual valid sources (textbooks or peer-reviewed papers) regarding the questions you have asked. If you are not willing to do that, we here at PF cannot help you.

#### Nugatory

Mentor
But sometimes it sounds to me that experienced physicists may be making that mistake in concluding they "understand" it, simply because certain outcomes can be predicted mathematically.
Scientists do indeed assert that they understand because they can predict mathematically, but that's not a mistake because that's all that science has ever claimed. If there's a mistake, it's in the popular expectation that there's more to scientific understanding than that.

This might be easier to see if we consider a more familiar example: Newton's law of gravitation. It allows us to make accurate mathematical predictions about an amazing variety of phenomena: the motion of the planets, the shape of the earth, the trajectories of dropped and thrown objects, whether objects will float or sink, why water runs downhill and bubbles float up, .... It is just astounding that $F=Gm_1m_2/r^2$ allows us to calculate the behavior of so many things. But does it "explain" anything in the sense that you're expecting of quantum mechanics? No. It describes the attraction between two nearby masses, but tells us nothing about what is "really going on". What is the mechanism by which this force is generated? What is the one mass doing to the other that leads to a force? The force is proportional to the product of the masses and the inverse square of the distances and that's reflected in the equation - but the equation is only that way because we've observed that the force behaves that way, not because we have some deeper understanding that tells us why that's the right equation.

The difference between Newtonian gravitation and QM is that this state of affairs bothers us in one case and not the other. That's partly because we can form a mental model of gravitational attraction (I have a common-sense understanding of heavy objects and I can imagine invisible springs stretched between them - never mind that a spring doesn't have the right inverse-square behavior) but not QM, and partly because the predictions of gravity accord with our preconceptions of how things "ought to" behave.

Last edited:

### Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving