Bell Test Violates Local Realism w/ Loophole-Free Electron Spins 1.3km Apart

[jim mcnamara]

Comments, please:

http://www.nature.com/nature/journal/vaop/ncurrent/full/nature15759.html
Loophole-free Bell inequality violation using electron spins separated by 1.3 kilometres
B. Hensen,
H. Bernien,
A. E. Dréau,
A. Reiserer,
N. Kalb,
M. S. Blok,
J. Ruitenberg,
R. F. L. Vermeulen,
R. N. Schouten,
C. Abellán,
W. Amaya,
V. Pruneri,
M. W. Mitchell,
M. Markham,
D. J. Twitchen,
D. Elkouss,
S. Wehner,
T. H. Taminiau
& R. Hanson
Nature (2015) doi:10.1038/nature15759 Received 19 August 2015 Accepted 28 September 2015 Published online 21 October 2015

More than 50 years ago, John Bell proved that no theory of nature that obeys
locality and realism can reproduce all the predictions of quantum theory: in
any local-realist theory, the correlations between outcomes of measurements on
distant particles satisfy an inequality that can be violated if the particles
are entangled. Numerous Bell inequality tests have been reported however, all
experiments reported so far required additional assumptions to obtain a
contradiction with local realism, resulting in ‘loopholes’. Here we report a
Bell experiment that is free of any such additional assumption and thus directly
tests the principles underlying Bell’s inequality. We use an event-ready
scheme that enables the generation of robust entanglement between
distant electron spins (estimated state fidelity of 0.92 ± 0.03). Efficient spin
read-out avoids the fair-sampling assumption, while the use of fast random-basis
selection and spin read-out combined with a spatial separation of 1.3 kilometres
ensure the required locality conditions. We performed 245 trials that tested the
CHSH–Bell inequality20 S = 2 and found S = 2.42 ± 0.20 (where S quantifies the
correlation between measurement outcomes). A null-hypothesis test yields a
probability of at most P = 0.039 that a local- realist model for space-like
separated sites could produce data with a violation at least as large as we
observe, even when allowing for memory in the devices. Our data hence imply
statistically significant rejection of the local-realist null hypothesis. This
conclusion may be further consolidated in future experiments; for instance,
reaching a value of P = 0.001 would require approximately 700 trials for an
observed S = 2.4. With improvements, our experiment could be used for testing
less-conventional theories, and for implementing device-independent quantum-
secure communication and randomness certification.

 

[Doc Al]

http://arxiv.org/abs/1508.05949

 

[bhobba]

It simply confirms QM - no biggie.

All such entangled correlations are is exactly that - correlations. Nothing spooky about it.

Thanks
Bill

 

[jfizzix]

It simply confirms QM - no biggie.

All such entangled correlations are is exactly that - correlations. Nothing spooky about it.

Thanks
Bill

I'd say they're definitely just correlations, but that their implications may be what's "spooky"..

Is it not "spooky" that such correlations can not be explained by a local hidden variable model?
- i.e.,that there is no way that their correlations could be determined by the (known and unknown) circumstances of the pair's creation?

It's hard to nail down what the implications would be, but at the very least, it means that if there are underlying deterministic laws beneath quantum mechanics, then they could not have a cosmic speed limit as we know it. On the other hand, it could just as well be that there is no underlying deterministic universe beneath quantum physics. Quantum metaphysics like this is what makes Bell-inequality violations so interesting, I feel.

 

[bhobba]

Is it not "spooky" that such correlations can not be explained by a local hidden variable model?
- i.e.,that there is no way that their correlations could be determined by the (known and unknown) circumstances of the pair's creation?

I don't want to get into a semantic discussion of 'spooky' which would be counter productive. Spooky in this context usually is associated with spooky action at a distance. Ordinary QM is based on the Galilean transformations where instantaneous action at a distance is built into it at its foundations - for that not to be the case you need to go to QFT. And in QFT locality is a much more nebulous thing associated with the cluster decomposition property which doesn't apply to correlated systems - which EPR is.

My view is that locality is not a concept applicable to correlated systems so any 'spookiness' related to that is rather moot. It interesting and important that QM allows a different kind of correlation than classically - but that all it is - a different kind of correlation. I even go as far to as to say entanglement is THE thing that differentiates classical probability theory from QM:
http://arxiv.org/abs/0911.0695

It seems to be absolutely foundational to QM. But 'spooky' is not what I would use to describe a theory that's simply a different kind of probability theoy allowing a new kind of correlation.

Thanks
Bill

 

[Sturk200]

I just saw this on the front page of the New York Times and am trying to understand it. What exactly does it mean that it rules out hidden variables? And how? I wish I could understand more than like three consecutive words of their paper.

 

[bhobba]

I just saw this on the front page of the New York Times and am trying to understand it. What exactly does it mean that it rules out hidden variables? And how? I wish I could understand more than like three consecutive words of their paper.

Its Bells theorem:
http://www.drchinese.com/Bells_Theorem.htm

Bell's original paper on it is also rather good:
https://cds.cern.ch/record/142461/files/198009299.pdf

It is a big part of quantum weirdness that you have this strange kind of non-classical correlation.

My opinion is its simply that - a rather weird correlation. But some read more into it than that. That's totally legit - I simply have a view QM is weird enough without making it weirder than necessary. This however is a view of QM - sort of like an interpretation - it can't be proved right or wrong.

Thanks
Bill

 

[jfizzix]

I just saw this on the front page of the New York Times and am trying to understand it. What exactly does it mean that it rules out hidden variables? And how? I wish I could understand more than like three consecutive words of their paper.

The basic idea, is that if there were some local hidden variables (i.e., pieces of information in the shared past of both particles) that could explain away the correlations they see, then the joint measurement probabilities would have to factor in a certain way.

As a consequence of the joint probability factoring in this way, you can derive Bell inequalities, which the measurement statistics would have to obey.

Then, if you can violate those Bell inequalities, the joint probabilities cannot be factorable this way, and as a result, all the pieces of information in the shared past of both particles cannot be enough the explain the correlations we see.

What counts for locality in local hidden variables, is that these pieces of information have to be in the shared past of both particles, or rather that those pieces of information could conceivably affect both particles through actions traveling at or below the speed of light.

 

[Nugatory]

I just saw this on the front page of the New York Times and am trying to understand it. What exactly does it mean that it rules out hidden variables? And how? I wish I could understand more than like three consecutive words of their paper.

We scooped the New York Times by many weeks :)
There are several threads from a few months back when earlier reports of this experiment were making the rounds.

 

[ddd123]

Can you explain why (if) is it that if we don't require determinism then we can get locality again? Is that what consistent histories does?

 

[bhobba]

Can you explain why (if) is it that if we don't require determinism then we can get locality again? Is that what consistent histories does?

That's not what Bells theorem says.

It says can't have both properties existing independent of measurement (called Counterfactual Definiteness - although strictly speaking its a bit subtler than that - but no need to go into that) and locality. Its really got nothing, per se, to do with determinism.

Thanks
Bill

 

[stevendaryl]

Can you explain why (if) is it that if we don't require determinism then we can get locality again? Is that what consistent histories does?

Determinism is a red herring in understanding Bell's theorem. No local realistic theory can reproduce the predictions of quantum mechanics, whether or not you allow nondeterminism.

Determinism comes into play in the EPR reasoning in the following way: In a twin pair experiment, the measurement of the spin of one particle along a particular axis by Alice allows us to predict, with 100% certainty, the result of a second measurement by Bob of the twin particle along that same axis. So after the measurement of Alice's particle, a property of Bob's particle becomes definite. So there are two possibilities, according to a realistic theory: Either Bob's particle had a definite spin along that axis BEFORE Alice's measurement, or Alice's measurement changed something about Bob's particle. So the EPR experiment, together with realism, implies either determinism or nonlocality. Then Bell goes on to show that determinism is out, as well.

But determinism isn't an assumption in this argument, it's a CONCLUSION from realism and the facts of twin-particle experiments.

In the above, by "realism" I just mean, vaguely, that individual particles have properties that exist independently of whether those properties have been observed yet.

 

[haushofer]

Why is determinism out then? The confirmation of EPR still allows for the Bohmian interpretation which obeys realism and is non-local, right?

 

[haushofer]

Btw, i never understood the problems people are having with the Bohmian interpretation being non-local explicitly. Should 't this be regarded as a merit since experiments like the one in the OP show that nature is non-local (according to the definition in the original EPR-paper)?

I guess I'm missing something because the Bohmian interpretation is notoriously hard to reconcile with relativity, whereas standard QFT is not.

 

[bhobba]

Why is determinism out then? The confirmation of EPR still allows for the Bohmian interpretation which obeys realism and is non-local, right?

That's correct.

Determinism isn't really an issue - you can have it if you want.

I personally have no issues with BM. The reason I don't hold to it is, as you correctly point out, it has a preferred frame which IMHO is at odds with the central lesson of SR and the symmetries of the POR. It doesn't disprove it or anything like that, it simply doesn't gel with my world view of physics where symmetry is the fundamental thing.

Thanks
Bill

 

[ddd123]

That's correct.

Determinism isn't really an issue - you can have it if you want.

I personally have no issues with BM. The reason I don't hold to it is, as you correctly point out, it has a preferred frame which IMHO is at odds with the central lesson of SR and the symmetries of the POR. It doesn't disprove it or anything like that, it simply doesn't gel with my world view of physics where symmetry is the fundamental thing.

Thanks
Bill

Thought there were advances in making dBB Lorentz-invariant?

 

[bhobba]

Thought there were advances in making dBB Lorentz-invariant?

I thought it already can be - but Dymystifyer is our expert not me. Obeying the POR and not having a preferred frame and Lorentz invariance are different things eg LET is Lorentz invariant.

Thanks
Bill

 

[stevendaryl]

Why is determinism out then? The confirmation of EPR still allows for the Bohmian interpretation which obeys realism and is non-local, right?

I meant that EPR shows that any realistic theory must be either deterministic or nonlocal, and Bell shows that deterministic doesn't work. So any realistic theory is nonlocal (whether or not it is deterministic).

 

[haushofer]

How does one reconcile EPR then with standard QFT if it is a nonlocal effect? I'm missing something important, apparently.

 

[bhobba]

How does one reconcile EPR then with standard QFT if it is a nonlocal effect? I'm missing something important, apparently.

Locality in QFT is based on the cluster decomposition property which excludes correlated systems.

Thanks
Bill

 

[stevendaryl]

How does one reconcile EPR then with standard QFT if it is a nonlocal effect? I'm missing something important, apparently.

Roughly speaking, applying quantum mechanics has two steps to it:

1. You use Schrodinger's equation to compute amplitudes.
   
2. You use the Born rule for converting amplitudes into probabilities for measurement outcomes.

The first step is continuous evolution of the wave function, which is local (in configuration space, anyway). The second step is what causes all the head-scratching about whether QM is nonlocal, and whether there is a "collapse" of the wave function following an observation, and all that stuff.

The same two steps are involved in QFT, as well. QFT gives a recipe for computing amplitudes for processes that is not too different from QM; it just allows for an indefinite number of particles, instead of the fixed number assumed by QM. The second step, which is what happens when you make a measurement, works the same in QFT as in QM.
EPR is as much a part of QFT as it is a part of QM. As a matter of fact, EPR is really a QFT problem, if you consider the creation of the twin pair in the first place.

 

[Keln]

The media continues to portray this as faster-than-light transmission of information, which seems wrong to me. This isn't my field and I have a very slippery grasp of the basics of QM as it is, so please forgive me if I'm mistaken in my thinking.

My understanding of entanglement and measurable properties is that, when entanglement occurs, two particles will have determined properties (such as spin up on one and down on the other), but since they are unknown until measured they are "both" and we don't know which is which (like the cat thought experiment). Then when they are separated and one is measured, then you know the measured property of the other no matter how far away it is. Like taking two slips of paper and writing "1" on one of them and "2" on the other, then randomly selecting one and putting it into your pocket and traveling a mile away. The paper in your pocket is both "1" and "2" as far as is known, but it is in reality one or the other. When you read it, you know instantaneously which the paper you left at home is. How is this actual information being relayed faster than light?

Or at the quantum level, is it physically and in reality both and the state only collapses when measured and actual information is being sent faster than light?

 

[ddd123]

The media continues to portray this as faster-than-light transmission of information, which seems wrong to me. This isn't my field and I have a very slippery grasp of the basics of QM as it is, so please forgive me if I'm mistaken in my thinking.

My understanding of entanglement and measurable properties is that, when entanglement occurs, two particles will have determined properties (such as spin up on one and down on the other), but since they are unknown until measured they are "both" and we don't know which is which (like the cat thought experiment). Then when they are separated and one is measured, then you know the measured property of the other no matter how far away it is. Like taking two slips of paper and writing "1" on one of them and "2" on the other, then randomly selecting one and putting it into your pocket and traveling a mile away. The paper in your pocket is both "1" and "2" as far as is known, but it is in reality one or the other. When you read it, you know instantaneously which the paper you left at home is. How is this actual information being relayed faster than light?

Or at the quantum level, is it physically and in reality both and the state only collapses when measured and actual information is being sent faster than light?

When measuring commuting observables the slips of papers example works, but when they don't commute there can be statistical correlations that exceed those allowed by local hidden variables. Though these correlations still aren't enough for transmission of information.

 

[bhobba]

The media continues to portray this as faster-than-light transmission of information, which seems wrong to me

You are correct - its wrong.

Its simply a strange correlation. Its different to the slips of paper:
https://cds.cern.ch/record/142461/files/198009299.pdf

But it's still just a correlation.

This whole thing is simply the correlation can not be explained if you assume, like the slips of paper, it has the properties independent of observation and any influences are local. But if locality applies to correlated systems is something that's rather moot anyway. Once you exclude it then the issues disappear.

Thanks
Bill

 

[DrChinese]

Btw, i never understood the problems people are having with the Bohmian interpretation being non-local explicitly. Should 't this be regarded as a merit since experiments like the one in the OP show that nature is non-local (according to the definition in the original EPR-paper)?

I guess I'm missing something because the Bohmian interpretation is notoriously hard to reconcile with relativity, whereas standard QFT is not.

The titles and text are misleading if you are not reading the original article. This is not a new result except in the important area that multiple loopholes are closed simultaneously. This is something of a holy grail in this area, so it is a great experiment. The rejection of local realism for the separate loopholes had already been performed a number of years back.

So this is essentially the same violation of a CHSH inequality as we have always seen. It is also a delayed choice experiment wrapped in for good luck. It does NOT prove there are nonlocal forces or actions (in other words there are no major quantum interpretations ruled out, such as MWI or Time Symmetric types).

 

[stevendaryl]

My understanding of entanglement and measurable properties is that, when entanglement occurs, two particles will have determined properties (such as spin up on one and down on the other), but since they are unknown until measured they are "both" and we don't know which is which (like the cat thought experiment). Then when they are separated and one is measured, then you know the measured property of the other no matter how far away it is. Like taking two slips of paper and writing "1" on one of them and "2" on the other, then randomly selecting one and putting it into your pocket and traveling a mile away. The paper in your pocket is both "1" and "2" as far as is known, but it is in reality one or the other. When you read it, you know instantaneously which the paper you left at home is. How is this actual information being relayed faster than light?

No, QM is NOT like the slips of paper example. Einstein and his collaborators P and R (umm...Podolsky and Rosen, maybe?) believed that entangled particles could be explained in those terms, but years later, John Bell proved that they can't be.

The difference between the EPR experiment and your slips of paper example is that the experimenters can make a choice as to which measurement to perform.

To turn your slips of paper example into something that is more like EPR, let's suppose that we have a game involving three players: Alice, Bob and Carol. There are two boxes that are identical, one for Alice and one for Bob. Each box has three little compartments, one with a Red door, one with a Blue door, and one with a Green door. There is a button on the side labeled "Ready". When the Ready button is pushed, the box behaves as follows: If one of the doors is opened, the contents behind the other two doors is instantly incinerated. So after the Ready button is pushed, it's only possible to check the contents of one compartment.

The game consists of many, many rounds. Each round of the game proceeds as follows:

• Carol puts slips of paper into each compartment of each box. Each slip of paper has "1" or "2" written on it.
  
• Carol pushes the "Ready" button on each box, and gives one box to Alice, and the other box to Bob.
• Alice and Bob each choose a door to open, and then pull out and read their respective slips of paper.
  
• The other slips in the unopened compartments are incinerated (so it's not possible to know what was on them except by asking Carol, who never tells).

The statistics of this game turn out to be the following:

• No matter which compartment is opened, there is a 50% chance of getting a "1" and a 50% chance of getting a "2".
  
• If Alice and Bob both open the same compartment (both Red or both Blue or both Green), then they always get opposite results.
• If Alice and Bob open different compartments, then they get the same result 75% of the time, and get opposite results 25% of the time.

So you could imagine that the way this could be pulled off is for Carol to randomly put "1" or "0" in each compartment of Alice's box, and then put the opposite results in the corresponding compartments of Bob's box. That's easy enough. But how is she going to make the statistics right in the case where Alice and Bob open different compartments? She might try the following:

1. Either put a 1 in Alice's Red compartment , or 0, with 50/50 chance of each.
2. Put the opposite in Bob's Red compartment .
3. With 25% chance, copy the number from Alice's Red compartment into her Blue compartment . With 75% chance, put the opposite number into her Blue compartment .
4. Put the opposite in Bob's Blue compartment .
5. With 25% chance, copy the number from Alice's Blue compartment into her Green compartment . With 75% chance, put the opposite number into her Green compartment .
6. Put the opposite into Bob's Green compartment .

But now she's stuck. If Alice chooses Red and Bob chooses Blue, then there is a 75% chance that Alice result will agree with Bob's. So far, so good. If Alice chooses Blue and Bob chooses Green, there is a 75% chance that their results will agree. But if Alice chooses Red and Bob chooses Green, then there is only a 62.5% chance that their results will agree (25% * 25% + 75% * 75%), instead of 75%.

You can actually prove that there is NO way for Carol to make it work, unless she knows ahead of time which compartments they will open, or if she is allowed to change the contents of Bob's compartments after Alice makes her choice (or change Alice's contents after Bob makes his choice, if he goes first).

 

[haushofer]

Mmm, doesn't ring a bell, so here's something I was missing, apparently ;) Thanks!

 

[haushofer]

The titles and text are misleading if you are not reading the original article. This is not a new result except in the important area that multiple loopholes are closed simultaneously. This is something of a holy grail in this area, so it is a great experiment. The rejection of local realism for the separate loopholes had already been performed a number of years back.

So this is essentially the same violation of a CHSH inequality as we have always seen. It is also a delayed choice experiment wrapped in for good luck. It does NOT prove there are nonlocal forces or actions (in other words there are no major quantum interpretations ruled out, such as MWI or Time Symmetric types).

Ok, I already thought so. The buzz is about the experimental set-up, not about the conclusions. But isn't entanglement considered to be non-local in itself because 'communication' between the two states is ruled out by experiment?

That's why I thought explicit non-locality, like in Bohmian mechanics, is a good thing to have. But now I'm confused :P

 

[DrChinese]

Ok, I already thought so. The buzz is about the experimental set-up, not about the conclusions. But isn't entanglement considered to be non-local in itself because 'communication' between the two states is ruled out by experiment?

That's why I thought explicit non-locality, like in Bohmian mechanics, is a good thing to have. But now I'm confused :P

Explicit non-locality may be good, it's all in your interpretation. The non-locality of QM is often referenced as "quantum non-locality". That is because of its unusual attributes that do not map to non-locality in a more general sense. Two critical elements of that:

1. There is no signal non-locality.
2. ALL non-local effects are limited in distance by a space-time diagram tracing in which c is respected, but time direction is not. Examples are typical Bell tests and entanglement swapping. Both of these factor c into the equation. You wouldn't really expect that.

Note that there is no obvious reason why #2 appears in this equation in "traditional" non-local interpretations such as Bohmian. On the other hand, they have the mechanism ready made.

 

[ddd123]

Explicit non-locality may be good, it's all in your interpretation. The non-locality of QM is often referenced as "quantum non-locality". That is because of its unusual attributes that do not map to non-locality in a more general sense. Two critical elements of that:

1. There is no signal non-locality.
2. ALL non-local effects are limited in distance by a space-time diagram tracing in which c is respected, but time direction is not. Examples are typical Bell tests and entanglement swapping. Both of these factor c into the equation. You wouldn't really expect that.

Note that there is no obvious reason why #2 appears in this equation in "traditional" non-local interpretations such as Bohmian. On the other hand, they have the mechanism ready made.

Sorry, what is the equation you are referencing?

 

[DrChinese]

Sorry, what is the equation you are referencing?

Sorry, "equation" is probably a poor choice of words on my part. I should have said something like "there is no obvious reason why #2 occurs".

But if you were asking what the maximum limit is: You must always factor in [ velocity * |delta(time)| ] for each particle/system involved in the entanglement, and sum that to get the maximum distance of the non-local effect. For photons, velocity=c so that is how the speed of light gets in there. So for a typical Alice and Bob who measure entangled photons 1 microsecond after being created, the *maximum* distance apart the non-local effect appears to exist between Alice and Bob is [ 2 * c * 1 microsecond ] or about 600 m apart. If you measured one of them later, the distance would go up for that leg of the total.

On the other hand: there is no obvious maximum distance limit in explicit non-local interpretations, although again that is not a flaw in them. But there are no known non-local effects outside some max as I have described.

Note: Using "repeaters" (entanglement swapping) involving progressively more quantum systems, you can change the "2" in the equation to "4" or whatever you can string together. And you can increase the max distance much greater by playing tricks with the delta(time) factor too. In these type, there is a zigzag pattern in spacetime to the remote connection.

 

[simplex1]

It appears that there could be two random phenomena that stay correlated without any communication at all.

For example, if somebody runs same random number generator with the same seed on two different computers with synchronized clocks, each time one looks at computer A and see there a value same value will be displayed by computer B also both of them appear to generate random numbers if watched separately.
Is this an example of 'Spooky' entanglement?

 

[DrChinese]

It appears that there could be two random phenomena that stay correlated without any communication at all.

For example, if somebody runs same random number generator with the same seed on two different computers with synchronized clocks, each time one looks at computer A and see there a value same value will be displayed by computer B also both of them appear to generate random numbers if watched separately.
Is this an example of 'Spooky' entanglement?

No, this is classical/deterministic. Entanglement is qualitatively different. It takes Bell's Theorem to really see why. Without that, you won't see what everyone is talking about. That's what happened with EPR (which came out about 20 years before Bell).

 

[stevendaryl]

Explicit non-locality may be good, it's all in your interpretation. The non-locality of QM is often referenced as "quantum non-locality". That is because of its unusual attributes that do not map to non-locality in a more general sense. Two critical elements of that:

1. There is no signal non-locality.
2. ALL non-local effects are limited in distance by a space-time diagram tracing in which c is respected, but time direction is not. Examples are typical Bell tests and entanglement swapping. Both of these factor c into the equation. You wouldn't really expect that.

That doesn't seem like much of a limitation. For any two events $A$ and $B$, you can get from $A$ to $B$ through slower-than-light signals, if you allow signals to propagate back in time. You just go back in time from $A$ to an event $C$ in the intersection of the backwards light cones of $A$ and $B$, and then go forward in time to $B$.

 

[simplex1]

The credibility of this article: http://www.drchinese.com/Bells_Theorem.htm written by the user DrChinese and talking about Bell's Theorem is quite low. It starts with so many inexact things that it is hard to believe the rest of the text is correct.

"No physical theory of local Hidden Variables can ever reproduce all of the predictions of Quantum Mechanics.

The significance of this statement is as follows: Quantum Mechanics is the "strange" theory introduced in 1927 by Niels Bohr and Werner Heisenberg to describe the fundamental nature of basic particles: the atomic nucleus, electrons and light (photons, or electromagnetic waves). This theory was a tremendous improvement upon pre-existing theory, and yielded immediate successes. In fact, the same theory exists today as Quantum Mechanics with virtually no change (although it has been extended to explain more phenomena). The 1927 version introduced such novel concepts as: the Heisenberg Uncertainty Principle; Max Born's statistical interpretation of the wave function, including superposition; and Bohr's complementarity (wave-particle duality). In addition, it included important recent advances such as the Schoedinger wave function (1925); the Pauli exclusion principle (1923); Bohr's semi-classical model of the atom (1913); additional contributions from Louis de Broglie and Paul Dirac; and of course the early seminal work of Max Planck (1900) and Albert Einstein (1905)."

1) Louis de Broigle formulated the wave-particle duality principle, not Bohr (see: https://en.wikipedia.org/wiki/Louis_de_Broglie).
2) The basis of Quantum Mechanics appeared in September 1925 due to Werner Heisenberg, not in 1927. Schrodinger published his paper regarding the wave function in January 1926. DrChinese puts Schrodinger in 1925 before Heisenberg (1927) which is incorrect.