NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/sci/physics/12027

Text File | 1992-07-29 | 5.1 KB | 125 lines

Newsgroups: sci.physics
Path: sparky!uunet!sun-barr!ames!riacs!danforth
From: [e-mail] (Douglas G. Danforth)
Subject: Re: Bell dethroned
Message-ID: <[e-mail]>
Sender: [e-mail]
Organization: RIACS, NASA Ames Research Center
References: <[e-mail]> <[e-mail]>
Date: Thu, 30 Jul 92 06:21:03 GMT
Lines: 113


]From sci.physics Wed Jul 29 22:44:18 1992
]From: [e-mail] (Benjamin Weiner)
]Date: 29 Jul 92 23:10:16 GMT
]Newsgroups: sci.physics
]Subject: Re: Bell dethroned
]
]I imagine none of the sci.physics "gurus" have taken a crack at this
]because it's an old idea, the post is confusing and cries out for
]graphics,

I agree. One can build two embedded plastic cylinders using overhead
transparencies that clarifies the model.

]          and they all have real work to do.  I, on the other hand,
]only have a qualifying exam to study for !)
]
]The assumptions leading to the Bell inequalities do include perfect
]detector efficiency.  However, since real detectors are, well, real
]detectors, useful derivations of the inequalities include a detector
]efficiency factor.  (This is, as I recall, especially important
]because you need to do coincidence counting to make sure you only
]count photons that are genuinely paired, and inefficient detectors
]will decrease your coincidences.)  This is all quite well understood.

Yes.

]
]It has also been understood for a LONG time that inefficient detection
]opens the door for claims that, essentially, you're missing the
]photons that make the difference.  The first Bell-inequality
]experiments (Clauser and Horne?  Clauser, Holt, Shimony, and Horne)
]had considerably lower detection efficiency than Aspect.
]The assumption that somehow the missing photons would tip the scales
]was generally disregarded as paranoid.  (Aspect's advance was to
]switch detector settings during photon time-of-flight; some people on
]this group don't think he did it right, but it is immaterial for this
]discussion.  You need the time-of-flight switch to rule out speed-of-
]light wavefn collapse, but not to rule out generic hidden variables.)
]
]Regardless of the niceties, one has to assume some kind of no-
]mechanism, angle-dependent variation of detector efficiency, in order
]to reconcile hidden-variables with QM in this experiment.  Danforth's
]particular choice is this "crown model."  My opinion is, sure, if you
]can assume variable efficiency

The model does not use variable efficiency.  Each detector has a
fixed efficiency equal to 81.8%.

]                               you can make the results do whatever
]you want.  Why should the detectors be variably efficient?  Shouldn't
]this happen with any old photons, not just correlated photons?

Yes.

]Shouldn't, then, one be able to test this by putting a detector in
]front of a known source, and just rotating the damn detector?

Test what.  The Crown model predicts that rotating a single detector
does not change the counts observed at that detector.

]
]How can anyone possibly presume to DERIVE the efficiency of a
]detector?

Because the model specifies a specific pattern of response in
the hidden variable space.  The effeciency "derived" is the maximum
efficiency allowed by the model that will still give the quantum results.
One can lengthen the base of each crown and lower the efficiency arbitarily
to zero.  The quantum results are retain as the efficiency is decreased.

]           Do you suppose Aspect's detectors magically had the same
]efficiency as, much earlier, Clauser's?

No.  There is a range of efficiencies possible that still admit
realistic local theories for this two particle correlation experiment.

]                                          Some of the experiments used
]spin-1/2 atoms, I believe.  One wonders if the atoms and their
]detectors could be following the very same rules.  An argument that
]relies on conspiracy between the detector and the detectee is no
]argument at all.

There is no conspiracy in the crown model.

]
]Finally, aficionados of EPR-tricks should read Greenberger, Horne,
]Shimony, and Zeilinger's paper (Am J Phys 58, 1131) along with
]Mermin's treatments (Am J Phys 58, 731; PRL 65, 3373) which show that
]it is possible to construct a simple system of 3 particles of spin 1/2
]and make certain measurements such that QM predicts a -1 correlation
]and hidden-variables predict a +1 correlation.

I can't speak to this issue.

]                                                That is, one could
]never trick up hidden variables to reproduce QM in this case.  I
]rather doubt anyone will be able to devise an experiment that
]reproduces this system, but I could be wrong.
]
]Ben Weiner
]

I would like to mention that I had a lovely talk this afternoon with an
experimentalist who has done EPR tests (I don't have permission to mention
who).  He liked the Crown model and suggested I publish.

To summarize:  Experiments of sufficient detector efficiency have not
yet been performed (to my knowledge) (and this includes Aspect's 1982)
to rule out local realistic models.


- Douglas Danforth