Quantum Mechanics: Understanding Particles at a Distance

[whiteboard1]

how can we know the state of particle which is far? Is the quantum mechanical result right?

 

[psmt]

The quantum mechanical result has been experimentally verified, if that's what you mean.

 

[jk22]

In fact you can only make the statistical deduction/constatation that the spin are opposite in an eprb experiment, i.e. When both subsystem are measured.
If you measure only one subsystem you can't predict a measurement thereafter on the second subsystem.
In a real experiment both effect occur and the contribution due to randomness has to be substracted from the covariance so that it is not exact but reaches 90 percent. See for example the covariance curve at the end of the article 'Violation of Bells inequalities by photons more than 10km apart' by Gisin's group in geneva.

 

[Bryan Sanctuar]

In the EPR paradox, using a singlet state, the quantum result is calculated easily after an undergrad course in quantum. Using the singlet state, the correlation between two spins as they encounter filters oriented along vector a and b is <sigma(a)sigma(b)> = -a.b

Now this -a.b is experimentally confirmed, so that means to a lot of people (not me) that the singlet state stretches out between the two particles. This, of course, is non-locality, firmly established in physics but not believe by me.

Your question is: what is the state of these separated particles? Some people believe that each EPR particle carries "half a quantum state" (That is half the entangled singlet--poppycock) I do not believe that.

If you assume spin has two axes of quantization, rather than one, then each spin carries a complete quantum state and their product accounts for the EPR correlation. This state is a superposition of the two axes of quantization and has eigenvalues of +/-root(2) rather than +/-1. If you measure that spin, you get the usual point particle spin because one orthogonal axis decohers.

 

[DrChinese]

In the EPR paradox, using a singlet state, the quantum result is calculated easily after an undergrad course in quantum. Using the singlet state, the correlation between two spins as they encounter filters oriented along vector a and b is <sigma(a)sigma(b)> = -a.b

Now this -a.b is experimentally confirmed, so that means to a lot of people (not me) that the singlet state stretches out between the two particles. This, of course, is non-locality, firmly established in physics but not believe by me.

Your question is: what is the state of these separated particles? Some people believe that each EPR particle carries "half a quantum state" (That is half the entangled singlet--poppycock) I do not believe that.

If you assume spin has two axes of quantization, rather than one, then each spin carries a complete quantum state and their product accounts for the EPR correlation. This state is a superposition of the two axes of quantization and has eigenvalues of +/-root(2) rather than +/-1. If you measure that spin, you get the usual point particle spin because one orthogonal axis decohers.

...

This is not correct, and is easily refuted by the Bell logic. You should not be posting stuff such as the above. Please check forum rules regarding personal theories which are in disagreement with established science.

 

[Bryan Sanctuar]

This is not correct, and is easily refuted by the Bell logic. You should not be posting stuff such as the above. Please check forum rules regarding personal theories which are in disagreement with established science.

As a scientist we question, especially things that do not make sense. What I said is completely backed up by objective reasoning.

I have simulated the EPR correlations, one coincidence at a time and these agree completely with experiment and quantum mechanics. The model is both local and real. It is also not a classical treatment.

What you say is that Bell's theorem easily refutes local realism, but does it? First Bell's theorem follows from believing Einstein Locality is incorrect in order to get the violation. But it is his spin assumption that is wrong, not his locality assumption.

I find that if an isolated spin has two orthogonal axes of quantization then local realism resolves the EPR paradox.

So my question is quite simple to you. If you believe in Bell's theorem, then tell me how two separated EPR pairs remain entangled over space like separations. If you can do this without using nonsense words like "quantum weirdness", then you will be able to do what no one else has done.

Non-locality makes no sense. It must, therefore, be wrong.

 

[DrChinese]

Non-locality makes no sense. It must, therefore, be wrong.

That it makes sense or not misses the point of science, which is to describe as best possible.

Either there is non-locality, or there is non-realism (or both). Bell's Theorem shows us such, and I can only assume you are not familiar with it. We have no obligation to review your model for mistakes, but by the forum rules you have an obligation NOT to post links such as above to unpublished material. So please cease, the next time you post similar to the above I will report you.

 

[Bryan Sanctuar]

That it makes sense or not misses the point of science, which is to describe as best possible.

Either there is non-locality, or there is non-realism (or both). Bell's Theorem shows us such, and I can only assume you are not familiar with it. We have no obligation to review your model for mistakes, but by the forum rules you have an obligation NOT to post links such as above to unpublished material. So please cease, the next time you post similar to the above I will report you.

I am surprised by your hostility. I am not asking you to review my model for mistakes, just have an open mind to new ideas. I can assure you that I am fully versed in Bell's work and, like quite a few others, believe that Bell's theorem is incorrect. After all Bell showed that von Neumann was incorrect, so why cannot I claim that Bell is incorrect, especially when I have objective reasons for it?

If Bell's theorem is correct, then you have to explain in a scientific way how non-locality works. I think that is an obligation scientist have. If you believe in something, then you should be able to explain it to your mother. Since you cannot, and no one can, then there must be something wrong or something we have missed.

So if you wish to report me, I think that the moderators will conclude that it is healthy to question, and I have not only questioned but proven that local realism works. I am submitting the paper to Physical Review A within the next week.

Sorry if I got your back up.

 

[Greg Bernhardt]

I am not asking you to review my model for mistakes, just have an open mind to new ideas.

I am submitting the paper to Physical Review A within the next week.

Welcome to PF Bryan! Much confusion will be cleared if you read our guidelines here
https://www.physicsforums.com/showthread.php?t=414380

Please let us know the result of your paper submission. Good luck!

 

[jk22]

Bells theorem proves a case of strong emergence in physics which is not very held there. It is when the whole is more than the sum of its parts. See wikipedia.
Douglas hofstadter also speaks about reductionism and holism in his book Goedel, Escher Bach.

For a physicist wholeness is often viewed as arising because of a lack of knowledge but maybe some global aspects are not reducible and it is then the goal of science to study those global aspects too.

 

[Dali]

how can we know the state of particle which is far? Is the quantum mechanical result right?

whiteboard1 - It's a little different in different setups, there are several EPR/Bell-type setups. A common setup is an entangled pair of spin-half particles in a singlet state. In that case one knows that the sum of the two particles spin is zero. So if one measures the spin of one of the particles to be say -1/2 one instantly knows the spin of the other one is +1/2 (along the same axis of measurement).
And, yes, all experiments so far agree with the predictions of quantum mechanics. They also violate the Bell limits, meaning that the experimental results can never be reproduced with any model that is both local and objective. (Ignoring here some "loophole-arguments" which basically says that we do something wrong in each and every of those experiments...)

I have simulated the EPR correlations, one coincidence at a time and these agree completely with experiment and quantum mechanics. The model is both local and real.

No, you have not. I have no idea what you misunderstood or do wrong in you simulations. But this is precisely what Bell showed is mathematically impossible. It has nothing to do with quantum mechanics, nor common sense. Just mathematics. He showed that any local and objective ("real") model has limits on correlations between separated measurements. If the model you simulated was both local and real, it would obey those limits. In certain setups however, quantum mechanics predicts results that violate those limits. Which is not that surprising since QM looks explicitly non-local (or non-objective depending on which interpretation you prefer). What is surprising and really intriguing is that experimental results violates those limits too. They happen to agree with the QM predictions as well (but that's not the important point!).

First Bell's theorem follows from believing Einstein Locality is incorrect in order to get the violation.

This is not correct and shows that you did not understand Bell's theorem. Bell makes two assumptions when deriving his limits; locality and some form of objectivity. But the whole point of the theorem is that when results do violate those limit, those results can never be reproduced by any local and objective theory.

Non-locality makes no sense. It must, therefore, be wrong.

That the speed of light should be constant for all observers make no sense to my intuition. Must it therefore also be wrong? No - science is all about keeping an open mind and accepting what repeatable experiments shows us.

 

[jk22]

So if one measures the spin of one of the particles to be say -1/2 one instantly knows the spin of the other one is +1/2 (along the same axis of measurement).
And, yes, all experiments so far agree with the predictions of quantum mechanics.

This is in theory true, but in experiments they don't get a 100% prediction for that case. See the violet curve at the end of : http://prl.aps.org/abstract/PRL/v81/i17/p3563_1 or http://arxiv.org/abs/quant-ph/9806043

This is due to uncorrelated photons that are counted in the experimental result, some kind of noise.

 

[jk22]

in fact there is a quantum way to calculate that 'noise' which correspond to single event and you get a prediction with 87% accuracy whereas experiment give 85 percent.

But i cannot put in this forum since it is not mainstream accepted physics.

 

[Ilja]

I can assure you that I am fully versed in Bell's work and, like quite a few others, believe that Bell's theorem is incorrect. After all Bell showed that von Neumann was incorrect, so why cannot I claim that Bell is incorrect, especially when I have objective reasons for it?

What Bell has shown is something different, namely that von Neumann has made, in his proof, assumptions which have been unreasonable. In particular, there was the example of de Broglie-Bohm theory, which is a completely reasonable hidden variable theory. And this theory violates an assumption used by von Neumann.

So the theorem proven by von Neumann remains valid. It is not wrong, but irrelevant, because the assumptions made by von Neuman have been unreasonably strong.

Similarly, there is no chance that Bell's theorem is incorrect. There is only a very weak chance that there appears a reasonable theory which violates one of the assumptions made by Bell, but appear, nonetheless, "local" (that means Einstein-causal) and "realistic" (with some modified but reasonable definition of realism).

A claim "Bell's theorem is wrong" is as nonsensical as "the theorem of Pythagoras is wrong".

If Bell's theorem is correct, then you have to explain in a scientific way how non-locality works.

Wrong. This would be an interesting question, but there is no such obligation.

 

[jk22]

Btw if we look at the experimental result concerning EPRB, I could find 3 experiments measuring CHSH, the values are :

Aspect : 2.69, error 0.05
Gisin : 2.38 error .09
Wineland : 2.25 error .05

we see that all 3 a greater than 2, hence implying a kind of non-locality according to Bell. However considering the experimental errors there is no overlap of the results possible. Does this mean that the error is in fact bigger than measured, or that we cannot compare in that way different experiments ?

 

[DrChinese]

Btw if we look at the experimental result concerning EPRB, I could find 3 experiments measuring CHSH, the values are :

Aspect : 2.69, error 0.05
Gisin : 2.38 error .09
Wineland : 2.25 error .05

we see that all 3 a greater than 2, hence implying a kind of non-locality according to Bell. However considering the experimental errors there is no overlap of the results possible. Does this mean that the error is in fact bigger than measured, or that we cannot compare in that way different experiments ?

Good question. As these experiments have very different efficiencies, they are not directly comparable in the sense you suggest. What they are designed to do is test the local realistic prediction of a max of 2. I believe they all also give a value for of the QM expectation in some fashion. I believe 2.82 is about the max ideal, and less than ideal is always lower.

It is worthy to mention that although the local realistic max is 2, that is not the prediction for all separable product state models. Some would be even lower. There really is no local realistic prediction per se because there are no viable local realistic models left on the table.

 

[craigi]

Btw if we look at the experimental result concerning EPRB, I could find 3 experiments measuring CHSH, the values are :

Aspect : 2.69, error 0.05
Gisin : 2.38 error .09
Wineland : 2.25 error .05

we see that all 3 a greater than 2, hence implying a kind of non-locality according to Bell. However considering the experimental errors there is no overlap of the results possible. Does this mean that the error is in fact bigger than measured, or that we cannot compare in that way different experiments ?

If those figures are correct and they are indeed measuring the same quantity, it is very likely that one or more of those measurements involved some significant, unaccounted for, systematic error.

The Gisin and Wineland results do cross within one standard deviation. It is the Aspect result that looks like the outlier, but in copying each other's techniques, the other 2 could easily have copied each other's systematic errors. Alternatively they could have independent systematic errors.

Alternatively, as Dr. Chinese suggests, they could be measuring slightly different quantities, all of which cross at the Bell threshold of 2.

 

[DrChinese]

If...it is very likely that one or more of those measurements involved some significant, unaccounted for, systematic error.

OK, take a deep breath here. These are top experimental teams (Wineland won the Nobel last year, for example). There is no reason to speculate like this. Please stick to the science.

 

[craigi]

OK, take a deep breath here. These are top experimental teams (Wineland won the Nobel last year, for example). There is no reason to speculate like this. Please stick to the science.

Sure, I'm not suggesting that they aren't experts. I'm just looking at those errors and thinking 5 standard deviations is a lot. Wouldn't you agree?

Systematic errors do end up in experiments. Do you remember the faster than light neutrino experimets recently?

 

[DrChinese]

Sure, I'm not suggesting that they aren't experts. I'm just looking at those errors and thinking 5 standard deviations is a lot. Wouldn't you agree?

Systematic errors do end up in experiments. Do you remember the faster than light neutrino experimets recently?

You mean, the neutrino experiment that could not be replicated?

As I said, the EPR experiments are not identical for a lot of reasons. There are hundreds of ways to test local realism, and the boundary of 2 is somewhat artificial. By convention, most experiments are designed to make it that 2 is the boundary. On the other hand, any pair analyzed which is not entangled at the time (due to inefficiency) gives a result closer to 2. So there are a lot of variables at play.

And therefore the results of Wineland's experiment will not match the result of Zeilinger, and yet both rule out local realism.

 

[craigi]

You mean, the neutrino experiment that could not be replicated?

As I said, the EPR experiments are not identical for a lot of reasons. There are hundreds of ways to test local realism, and the boundary of 2 is somewhat artificial. By convention, most experiments are designed to make it that 2 is the boundary. On the other hand, any pair analyzed which is not entangled at the time (due to inefficiency) gives a result closer to 2. So there are a lot of variables at play.

And therefore the results of Wineland's experiment will not match the result of Zeilinger, and yet both rule out local realism.

If I recall correctly they reproduced the FTL neutrino but about a month later found a loose connection which introduced the time delay.

As you suggest, they're probably measuring slightly different quantities in the Bell tests. Though I'd never rule out systematic errors until different experiments give the same results for the same quantity. That said, I don't doubt that local realism has been excluded.

 

[StevieTNZ]

So the question is:

If a Bell inequality (or in this matter CHSH inequality) test results in >2, then local realism is ruled out.

But what about verifying the QM predictions? Surely all three results don't verify QM predictions as they differ so significantly. What is the QM prediction for entanglement experiments, such as those experiments whose results were easier stated?

 

[jk22]

The difference between the QM prediction 2.82 and the average experimental result 2.44 (so 13%) you mean ?

Well if this is not due to experimental aspects, such as a BBO crystal that do not produce all the time perfectly correlated photons due to imperfection for example,photons that are measured only at one side, then it's the turn to the theory too to explain this discrepancy.

 

[DrChinese]

So the question is:

If a Bell inequality (or in this matter CHSH inequality) test results in >2, then local realism is ruled out.

But what about verifying the QM predictions? Surely all three results don't verify QM predictions as they differ so significantly. What is the QM prediction for entanglement experiments, such as those experiments whose results were easier stated?

Each experiment discusses the QM expectation, which is calculated according to their setup. There is no specific disagreement between experiments. Read any two, and if you are unsure of this point, let's discuss. Just give me the names of them.

 

[StevieTNZ]

Each experiment discusses the QM expectation, which is calculated according to their setup. There is no specific disagreement between experiments. Read any two, and if you are unsure of this point, let's discuss. Just give me the names of them.

So what I gather is a QM prediction is calculated in accord with their set-up, and each three measurement results posted earlier fall within each of those calculated predictions?

 

[DrChinese]

So what I gather is a QM prediction is calculated in accord with their set-up, and each three measurement results posted earlier fall within each of those calculated predictions?

From Wineland et al:

Result: 2.25 +/- .03

"If we take into account the imperfections of our experiment (imperfect state fidelity, manipulations, and detection), this value agrees with the prediction of quantum mechanics.

The result above was obtained using the outcomes of every experiment, so that no fair-sampling hypothesis is required. In this case, the issue of detection efficiency is replaced by detection accuracy. The dominant cause of inaccuracy in our state detection comes from the bright state becoming dark because of optical pumping effects. For example, imperfect circular polarization of the detection light allows an ion in the |↓right fence state to be pumped to |↑right fence, resulting in fewer collected photons from a bright ion. Because of such errors, a bright ion is misidentified 2% of the time as being dark. This imperfect detection accuracy decreases the magnitude of the measured correlations. We estimate that our Bell's signal would be 2.37 with perfect detection accuracy."

http://www.nature.com/nature/journal/v409/n6822/full/409791a0.html

So the result excludes local realism (upper limit of 2) but agrees reasonably well with QM (upper limit of 2.37).

 

[jk22]

From Wineland et al:

Result: 2.25 +/- .03

"If we take into account the imperfections of our experiment (imperfect state fidelity, manipulations, and detection), this value agrees with the prediction of quantum mechanics.

This agree with QM in the sense that it's bigger than 2 (hence non-local), but the QM prediction is 2.82 ($$2\sqrt{2}$$) for that measurement. I don't know if it's worth to notice that or to be so picky about that exact value.

Following Gisin (http://arxiv.org/abs/quant-ph/9806043) this discrepancy is due to single events.

Note that it's possible to get the theoretical value of $$\frac{7}{4}\sqrt{2}\approx 2.47$$ when taking into account single events in the covariance.

This is done by calculating the covariance (which is used in the inequality) :

$$Cov(A,B)=\underbrace{\langle A\otimes B\rangle}_{\textrm{double events}}-\underbrace{\langle A\rangle\langle B\rangle}_{\textrm{single events}}$$

The average of A towards the singlet state is not straightforward, since A is of dimension 2 and the state of dimension 4. Then is obtained by projecting on the eigenvectors of A and for B any unit vector, the same for B, then the product is averaged over this unit vector.

 

[DrChinese]

This agree with QM in the sense that it's bigger than 2 (hence non-local), but the QM prediction is 2.82 ($$2\sqrt{2}$$) for that measurement.

I appreciate that you see it as 2.82, but Wineland's team has a different value for their setup. As I say, each setup has different issues and you should not compare one directly to the other. Although you are determined to do that anyway...

They measure lithium ions, and many experiments measure photons. Choice of angle settings can be a factor, although I don't think that is an issue here. Obviously Wineland, Zeilinger and Gisin are all well aware of each other's work.

 

[StevieTNZ]

Its just when people say that they performed an entanglement experiment and got results in agreement with predictions, the ones listed above are no-where near the QM prediction.

 

[DrChinese]

Its just when people say that they performed an entanglement experiment and got results in agreement with predictions, the ones listed above are no-where near the QM prediction.

Probably a big conspiracy.

I am not really sure what I need to say to explain that the experiments are not identical, and don't have identical QM expectation values. The primary similarity is that 2 is the local realistic upper bound in all 3, which is simply convention.

 

[Nugatory]

Its just when people say that they performed an entanglement experiment and got results in agreement with predictions, the ones listed above are no-where near the QM prediction.

Which people? If you don't go back to the original publication, you're hearing a conclusion while seeing neither the original results nor the steps by which the results are said to support the conclusion.

 

[jk22]

I appreciate that you see it as 2.82, but Wineland's team has a different value for their setup. As I say, each setup has different issues and you should not compare one directly to the other. Although you are determined to do that anyway...

My question then is how do they calculate the value for their setup ? Reading Bell's original work shows 2 things : local realism limits CHSH to 2, and QM gives 2.82 as prediction.

I found no trace of a theoretical calculation to obtain 2.37 in that experiment.

The thing that disturbs me is that 2.82 corresponds to prediction with certainty along the same axes, following EPR, whereas 2.37 does not imply such a fact.

 

[StevieTNZ]

Yes, I'm confused as well with this. I'm desperately trying to understand this matter myself.

 

[jk22]

The result 2.82 permits a prediction with certainty, whereas 2.37, since the covariance is also a cosine curve, doesn't.

Why bother about that prediction capacity : this is to find back in 1935's EPR paper :

EPR set up a criterion (without any proof) that if you can predict with certainty there should exist elements of phys. reality.

Bell quantum calculation permits such a prediction with certainty, so in the sense of EPR there should exist elements, but those are not the Bell's element lambda.

The latter point was proven by Wineland's experiment for example since CHSH is bigger than 2.

But this experimental result goes further : we cannot predict with certainty, hence following EPR there is no reason to believe there are elements of reality, the criterion for existence is not anymore fulfilled.

 

[DrChinese]

My question then is how do they calculate the value for their setup ? Reading Bell's original work shows 2 things : local realism limits CHSH to 2, and QM gives 2.82 as prediction.

I found no trace of a theoretical calculation to obtain 2.37 in that experiment.

Not quite correct. 2 is the upper limit for any local realistic (LR) theory, as you say. But 2.82 is the upper limit for QM, not the prediction. So a measured value of 2.3 rules out LR and validates QM.

Any pair which reaches the detectors but has experienced decoherence on the polarization basis serves to eat away at the actual rate and cause it to be lower. This can be estimated in advance, as calibration is occurring. For example, in Weihs et al, they obtained 97% visibility during that phase. Therefore they expected a result of about 2.82 * 97%, or 2.74. The experimental value was 2.73 +/- .02, in good agreement. See after their (1).

http://arxiv.org/abs/quant-ph/9810080

I realize it bothers you that some authors do not explain specifically how they arrived at their expected values, but this is more a matter of editing than anything else. The actual information to arrive at 2.37 is in the Wineland article for anyone who is interested.

Hint: (.88 - (2*.02)) * 2.82 and I will let you find and figure that out for yourself.