Why is entanglement necessary for understanding quantum mechanics?

[DrChinese]

All the equation needs is two polarizer angles. I don't see how can that be anything but "prior condition".

As I indicated in an early post in this thread, photons can be entangled AFTER they are measured. Clearly prior condition is not possible in this light.

 

[Jabbu]

That's the definition of correlated.

After photon A goes through polarizer A, how can photon correlation make photon B go through polarizer B with 100% chance? What does photon correlation have to do with polarizers?

Why is photon entanglement not part of the equation?

 

[Jabbu]

The result cannot be determined before make the measurements.

Set polarizer A to -30, polarizer B to +30 degrees, and by doing so you will predetermine the outcome to be 25% correlation, every time. What result and what measurements are you talking about?

 

[Jabbu]

As I indicated in an early post in this thread, photons can be entangled AFTER they are measured. Clearly prior condition is not possible in this light.

Photons can be measured without being absorbed/destroyed? Where did you read that?

 

[microsansfil]

Set polarizer A to -30, polarizer B to +30 degrees, and by doing so you will predetermine the outcome to be 25% correlation, every time. What result and what measurements are you talking about?

Made a bet on your salary with this way of thinking, in this quantum context, and you will understand what I'm talking about.

Patrick

 

[stevendaryl]

After photon A goes through polarizer A, how can photon correlation make photon B go through polarizer B with 100% chance? What does photon correlation have to do with polarizers?

Why is photon entanglement not part of the equation?

Even though polarizing filters in QM work similarly to the way they work classically, the details seem very different.

Classically, if you have an electromagnetic field falling on a filter, you can think of what happens like this: Let \vec{E} be the electromagnetic field of the light. You can write this as a superposition of two different fields: \vec{E}_{||} which is parallel to the filter, and \vec{E}_{\bot}, which is perpendicular to the filter. The perpendicular component is absorbed, so the part that passes through is just \vec{E}_{||}

Now, if you drop the intensity low enough that you start seeing individual photons, then things start looking very different. A photon is not partly absorbed. It's either absorbed completely, or it passes through unchanged. So it seems as if every photon is either polarized perpendicular to the filter, or is polarized parallel to the filter.

What's special about entanglement is that the twins photons have the same polarization state. So if Alice's photon passes through her filter, which is at angle 30°, say, then it's as if the photon was always polarized at angle 30°. And Bob's corresponding photon acts as if it were always polarized at angle 30°. So if Bob's filter is at 30°, his photon will also pass.

I say "as if", because the photons did not have a definite polarization state before they were detected.

 

[atyy]

Can you point any document on the internet where we can see samples of actual experimental data?

Here's some actual data:

http://arxiv.org/abs/quant-ph/9810080
Violation of Bell's inequality under strict Einstein locality conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, Anton Zeilinger

 

[Jabbu]

What's special about entanglement is that the twins photons have the same polarization state. So if Alice's photon passes through her filter, which is at angle 30°, say, then it's as if the photon was always polarized at angle 30°. And Bob's corresponding photon acts as if it were always polarized at angle 30°. So if Bob's filter is at 30°, his photon will also pass.

What if photon A doesn't pass, what is then preventing photon B to go through polarizer B?

 

[Nugatory]

After photon A goes through polarizer A, how can photon correlation make photon B go through polarizer B with 100% chance?

it cannot and it does not. What the correlation says is that if photon A goes through polarizer A, then we know something about the probability that photon B will or will not go through the polarizer.

We can directly measure this, by watching what happens to a large number of entangled photon pairs. That's an experiment, and it's been done by Alain Aspect and many others.

Or we can calculate what should happen in such an experiment. We can do this using the Rules of Quantum Mechanics (in which the entanglement appears explicitly in the wave function for the system) and get one prediction for the correlation, or we can get a different prediction using classical rules in which there is no entanglement. One calculation matches the experimental result, and the other does not, so we know that which one is right.

But after all of this, we still don't know how this correlation came to be. We just know that quantum mechanics predicted it correctly and classical mechanics did not. Bell's theorem further tells us that any calculation that does not assume that the setting at A can influence the measurement at B cannot make the correct prediction - but it says nothing about the nature and mechanism of the influence.

 

[Jabbu]

it cannot and it does not. What the correlation says is that if photon A goes through polarizer A, then we know something about the probability that photon B will or will not go through the polarizer.

I think stevendaryl provided very sensible answer, as far as QM answers go anyway, but it's a little bit incomplete. We are talking about the case where polarizers are aligned at the same angle, like this:

theta_A = +30, theta_B = +30
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...so if distance to polarizer A is shorter than to polarizer B, then photon B must always do the same thing photon A did.

 

[billschnieder]

Experiments are far less spectacular than Alice and Bob adventures. I don't see why make up stories when we can describe actual experiments. In the experiment there is a photon A and polarizer A on one side, and on the other side there is a photon B and polarizer B. Photon A will try to pass through polarizer A, and photon B will try to pass through polarizer B. If both manage to pass or if both fail we record '1', it's a match, and if one goes through but not the other we record '0', it's a mismatch. This is repeated with 10,000 more photons, the number of matches and mismatches are compared and then somehow interpreted to imply all kinds of crazy stuff.

I'm not impressed. The result is so very indirect and only vaguely related to what is being inferred from it. There is Malus's law in classical physics which can calculate probability for a photon to pass through a polarizer. Can it be demonstrated the outcome of the experiment in not predetermined by the angles set on the polarizers and Malus's law before the experiment even begins?

I should highlight that in actual experiments it is impossible to know that a photon did not go through, let alone that both of them did not go through. The typical setup is usually quite different than the (0,1) values being discussed. Typically you have a beam-splitter with two arms at each station. One of them labelled +1 and other labelled -1. a (+1,+1) or (-1,-1) result is a match and a (+1, -1) or (-1, +1) result for each pair is a mismatch. The way correlatiosn (actually expectation values) are calculated in the experiment is also quite different from the equations being discussed. They are calculated as <AB>, ie the average of a product of results on both sides for the given pair of angular settings. It is this <AB> value that matches the QM expectation value.

But it is even worse than that. Experiment do not give you pairs. Rather, you have a random series of time-tagged +1/-1 results at Alice, and another random series of time-tagged +1/-1 results at Bob corresponding to when the various detectors clicked. Then after the experiment you try to find pairs of clicks close enough in time which you *assume* belong to the same pair. Any unpaired value is discarded. There are no one sided (or does not pass through) values in the calculation.

 

[Nugatory]

I think stevendaryl provided very sensible answer, as far as QM answers go anyway, but it's a little bit incomplete. We are talking about the case where polarizers are aligned at the same angle, like this:

theta_A = +30, theta_B = +30
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...so if distance to polarizer A is shorter than to polarizer B, then photon B must always do the same thing photon A did.

Yes, and that's consistent with what I said. If the photon goes through polarizer A, then we know something about the probability that photon B will go through its polarizer. In this particular case, we know that the probability is 100%.

 

[Jabbu]

Yes, and that's consistent with what I said. If the photon goes through polarizer A, then we know something about the probability that photon B will go through its polarizer. In this particular case, we know that the probability is 100%.

In that case the same question for you: what if photon A doesn't pass, what is then preventing photon B to go through polarizer B? And if those photons are really unpolarized or "undefined" upon emission, then why don't we simply make them polarized first and then see what is really going on and how it actually works?

 

[DrChinese]

Here's some actual data:

http://arxiv.org/abs/quant-ph/9810080
Violation of Bell's inequality under strict Einstein locality conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, Anton Zeilinger

I gave that reference already, but I don't believe Jabbu is reading those. Also gave the Dehlinger reference on Bell tests which is one of the best tutorials.

http://arxiv.org/abs/quant-ph/0205171

 

[Jabbu]

Here's some actual data:

http://arxiv.org/abs/quant-ph/9810080
Violation of Bell's inequality under strict Einstein locality conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, Anton Zeilinger

Are you referring to the graph? I'm talking about raw binary streams data. Apparently there used to be some raw data from the Innsbruck experiment of 1998, available at:

http://www.quantum.univie.ac.at/research/bellexp/data.html

... but it's not there anymore, and that's the best I could find. I don't understand why are all those papers published without actual data obtained, I thought that's supposed to be obligatory.

 

[Jabbu]

I gave that reference already, but I don't believe Jabbu is reading those. Also gave the Dehlinger reference on Bell tests which is one of the best tutorials.

http://arxiv.org/abs/quant-ph/0205171

Does any of those papers contain or link to actual raw data samples?

 

[DrChinese]

Photons can be measured without being absorbed/destroyed? Where did you read that?

I don't think you followed my statement. They can be entangled AFTER they are observed. They no longer exist. That is because time ordering is very different than you might expect. I gave this reference previously as well.

http://arxiv.org/abs/quant-ph/0201134

And in fact photons can be entangled that have never co-existed or even been in one another's time cone:

http://arxiv.org/abs/0809.3991
http://arxiv.org/abs/1209.4191

It is more correct to consider the entire context of an experiment, from setup to detection, when determining predictions. Ie future setting are relevant to the statistical prediction. This violates normal everyday views on things, but is fully in keeping with QM.

 

[DrChinese]

Does any of those papers contain or link to actual raw data samples?

No, the raw data is very difficult to decipher and is not in a format that can be readily analyzed. It would almost be easier to do the experiment yourself. Of course, there is no real reason to see the raw data unless you simply don't believe the results. The experiments are described in plenty of detail for those who are interested.

Please keep in mind that these experiments involve a lot of elements you will need to understand first. For once, billschnieder has said something correct when he states that "typical setup is usually quite different than the (0,1) values being discussed".

However, I would strongly urge you to ignore ANY other comment he makes UNTIL you understand Bell tests better. He has a very non-standard agenda and will lead you to a place where it will be impossible for anyone to assist you.

 

[DrChinese]

I don't understand why are all those papers published without actual data obtained, I thought that's supposed to be obligatory.

I have never seen raw data published and I have probably read 1000+ papers.

 

[stevendaryl]

I should highlight that in actual experiments it is impossible to know that a photon did not go through, let alone that both of them did not go through. The typical setup is usually quite different than the (0,1) values being discussed. Typically you have a beam-splitter with two arms at each station. One of them labelled +1 and other labelled -1. a (+1,+1) or (-1,-1) result is a match and a (+1, -1) or (-1, +1) result for each pair is a mismatch. The way correlatiosn (actually expectation values) are calculated in the experiment is also quite different from the equations being discussed. They are calculated as <AB>, ie the average of a product of results on both sides for the given pair of angular settings. It is this <AB> value that matches the QM expectation value.

But it is even worse than that. Experiment do not give you pairs. Rather, you have a random series of time-tagged +1/-1 results at Alice, and another random series of time-tagged +1/-1 results at Bob corresponding to when the various detectors clicked. Then after the experiment you try to find pairs of clicks close enough in time which you *assume* belong to the same pair. Any unpaired value is discarded. There are no one sided (or does not pass through) values in the calculation.

That's a very good point. I think there have been attempts to explain EPR using local means by exploiting the differences between actual experiments and the idealization presented in most theoretical discussions of Bell's Inequality. For example, I think that someone named "Dereiter" or something like that? The idea is that Alice and Bob don't necessarily always measure the corresponding photons. If you assume that the likelihood of getting a mismatch is correlated with their filter settings, then maybe it's possible to reproduce the QM predictions for EPR.

I'm not sure if all such loopholes have been closed by experiments, but it's a little puzzling to think that errors in interpreting data would just happen to reproduce the predictions of QM.

 

[DrChinese]

And if those photons are really unpolarized or "undefined" upon emission, then why don't we simply make them polarized first and then see what is really going on and how it actually works?

Once a non-reversible measurement is performed (ie it is polarized), it is no longer entangled. Please keep in mind that the Heisenberg Uncertainty Principle (HUP) is at work, even with entangled particle pairs. You cannot beat the HUP! So when you observe Alice, you obtain identical information about Bob (and vice versa). That is the limit.

 

[DrChinese]

That's a very good point. I think there have been attempts to explain EPR using local means by exploiting the differences between actual experiments and the idealization presented in most theoretical discussions of Bell's Inequality. For example, I think that someone named "Dereiter" or something like that? The idea is that Alice and Bob don't necessarily always measure the corresponding photons. If you assume that the likelihood of getting a mismatch is correlated with their filter settings, then maybe it's possible to reproduce the QM predictions for EPR.

I'm not sure if all such loopholes have been closed by experiments, but it's a little puzzling to think that errors in interpreting data would just happen to reproduce the predictions of QM.

De Raedt et al is the team you are referring to. Very complex stuff and far out of the league of this discussion. That's another thread. If we try to go down that path here, I will seek to close this thread.

Not trying to shut down discussion but this thread is already nearing its useful limit with over 200 posts now. Jabbu has barely moved, and should be doing more homework before posting additional comments.

Jabbu, I hope you are reading this and listening carefully.

 

[Jabbu]

No, the raw data is very difficult to decipher and is not in a format that can be readily analyzed. It would almost be easier to do the experiment yourself. Of course, there is no real reason to see the raw data unless you simply don't believe the results. The experiments are described in plenty of detail for those who are interested.

According to how you said correlation is experimentally calculated, the raw data must be in this format:

theta_A = +30, theta_B = +30
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

That's the only actual information. What does it mean depends on which interpretation you like the most, but some interpretations make more sense than other. The question is which one makes the most, and hopefully, total sense.

 

[DrChinese]

According to how you said correlation is experimentally calculated, the raw data must be in this format:

theta_A = +30, theta_B = +30
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

That's the only actual information. What does it mean depends on which interpretation you like the most, but some interpretations make more sense than other. The question is which one makes the most, and hopefully, total sense.

Bell test data must be time-stamped and "lined up" (Alice's timestamps matched with Bob's) after the fact. It takes computer programming to analyze and make sense of. All of this is far past where the current thread sits. Until you understand the basic science, none of that will make much sense.

Yes, there are a variety of interpretations, and it is good to understand those: orthodox QM, Bohmian Mechanics, Many Worlds, and Time Symmetric types are good ones to be aware of.

 

[Jabbu]

I should highlight that in actual experiments it is impossible to know that a photon did not go through, let alone that both of them did not go through. The typical setup is usually quite different than the (0,1) values being discussed. Typically you have a beam-splitter with two arms at each station. One of them labelled +1 and other labelled -1. a (+1,+1) or (-1,-1) result is a match and a (+1, -1) or (-1, +1) result for each pair is a mismatch.

1/0 is more readable and easier to type than +1/-1.

The way correlatiosn (actually expectation values) are calculated in the experiment is also quite different from the equations being discussed. They are calculated as <AB>, ie the average of a product of results on both sides for the given pair of angular settings. It is this <AB> value that matches the QM expectation value.

How do you apply that to this sequence:

A: 1 0 1 0 1 0 1 0 1 0 1 1 0 0 0 0 1 1 0 1
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...what's correlation percentage?

But it is even worse than that. Experiment do not give you pairs. Rather, you have a random series of time-tagged +1/-1 results at Alice, and another random series of time-tagged +1/-1 results at Bob corresponding to when the various detectors clicked. Then after the experiment you try to find pairs of clicks close enough in time which you *assume* belong to the same pair. Any unpaired value is discarded. There are no one sided (or does not pass through) values in the calculation.

I'll keep that in mind.

 

[Jabbu]

I don't think you followed my statement. They can be entangled AFTER they are observed. They no longer exist. That is because time ordering is very different than you might expect.

If they no longer exist how do you confirm they are entangled now and were not entangled before?

Once a non-reversible measurement is performed (ie it is polarized), it is no longer entangled.

They can be entangled from the beginning if we polarize them both at the same angle. Has no one attempted the experiment with constant photon polarization?

 

[Jabbu]

It takes computer programming to analyze and make sense of.

I'm counting on it, I've written programs to analyze binary sequences before.

 

[billschnieder]

How do you apply that to this sequence:

A: 1 0 1 0 1 0 1 0 1 0 1 1 0 0 0 0 1 1 0 1
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...what's correlation percentage?

Like I said earlier, there are no 0 events in actual experiments, because only detected photons count towards the expectation value. So the sequence above seems easier to type and discuss but bears no resemblance to anything that an experiment produces, at least any experiment that has ever been performed. If you think about it, that sequence is pretty incredible. To get a sequence like that, you need to know that two photons were emitted at the same time but neither was detected in some cases, and in others that two were emitted but only one was detected. In an actual experiment you only know what was detected. If you try to measure at the source, you destroy them and invalidate the experiment. If you detect only one, you can't be absolutely sure that another one was emitted but did not go through. So the 11 case is the only certain case for matches that you can have in any experiment. And even then you have time-tagging and matching to deal with, which complicates matters even further. If you want to see some actual experimental data, see http://people.isy.liu.se/jalar/belltiming/ which has a portion of the Weihs data.

But that was for EPR, even if you think about Malus, how do you know that a photon did not go through to be able to record a 0 data point? The expectation values are obtained experimentally using very different methods, first you measure total intensity without filter, then you insert filter without changing the source and measure total intensity again. This is how you obtain cos^2 relationship experimentally. That is why you will never see experimental data for Malus, similar to your sequence above.

It is easy to discuss thought experiments and fictitious data but when you get down to actual experiments, most of that makes no sense.

 

[DrChinese]

1. If they no longer exist how do you confirm they are entangled now and were not entangled before?

2. They can be entangled from the beginning if we polarize them both at the same angle. Has no one attempted the experiment with constant photon polarization?

Are you even reading the references I am posting? You have been going around in circles stuck with the same ideas for a long time.

1. Obviously, they violate a Bell Inequality. That is the standard for entanglement. Since they are actually made to be entangled about 50 nanoseconds AFTER they are detected, I am not sure what you are looking for. Non-entangled pairs do NOT violate a Bell inequality.

2. Have you not been listening to what everyone has been telling you? Photons with a specific polarization are not entangled. Once they are given a specific polarization, entangled photons are no longer polarization entangled.

-------------------------

Again Jabbu: I am urging you to think carefully before you post. You should not be a) repeating questions you have already asked; and b) you should not be making speculative statements such as "They can be entangled from the beginning if we polarize them both at the same angle." That is completely wrong as has been repeatedly explained.

There are many others reading (and answering on) this thread. It is normal for you to ask a question, get an answer, then review relevant material before asking another question. That is how you learn, which is what we assume you are here for. If you are here to debate the subject, this thread will likely come to a very quick close. This is not that kind of forum.

 

[DrChinese]

If you want to see some actual experimental data, see http://people.isy.liu.se/jalar/belltiming/ which has a portion of the Weihs data.

Thanks for the link, Bill.