Uncertainty principle.

[jason.p]

I seem to remember recently reading somewhere that someone had found a way to do quantum observations in which the procedure doesn't alter the outcome. A sort of indirect approach. I'm just reading Brian Greene's Fabric of the Cosmos (again!) and just got to the "uncertainty bit" and wonder if this alters, in any way, the basis of understanding of the uncertainty principle. Can anyone point me to a reference (whether it's actually there or not )

Thanks

Jason

[mytola]

Hi Jason,

I'm having trouble seeing how this could be possible, so if you find a link, please post it

Cheers

[jason.p]

I imagine it is doing what Einstein, Podolsky, Rosen (EPR) proposed as a way around the uncertainty principle (because basically they didn't like it). I can't remember where I read it, but it struck me that it could alter the basic understanding.

[acey]

Measurement doesn't always need to "alter" the observable. Measurement will put the system in an eigenstate of the operator that corresponds to the observable being measured, and subsequent measurement of the same observable can then yield the same result. For instance, arrange a system in which the spin of an electron in the x-direction is "up"; then subsequent measurement of the spin in the x-direction again yields "up". Information is however completely lost with regard to spin in the y- or z-directions, and any attempt to measure those will mean that one can no longer say that the spin in the x-direction is still "up". That's the uncertainty principle. The principle should be thought of as a restriction on what can simultaneously be known about two distinct observables (e.g. position and momentum, or spin in orthogonal directions) rather than knowledge of a single observable.

Einstein, Podolsky and Rosen said: suppose a particle splits, such that spin is conserved, and if one half has spin "up" (in any given direction) then the other half must have spin "down". We measure spin-x of one particle and find "up", thus concluding that the other must be spin-x "down". But we could instead have measured spin-y or spin-z. According to EPR, there must therefore be a "real" value of spin-x, spin-y and spin-z, and the uncertainty principle merely expresses a restriction on what we can know of these values. The alternative, they say, is that when spin-x is measured, a signal must propagate instantaneously to the other particle, telling it which value to take for spin-x. This "spooky action at a distance", they said, would violate special relativity.

Bell looked at the probabilities that would arise for certain states measured in this way, according to quantum theory, thus deriving the "Bell inequalities", which were verified in an experiment (using photons rather than electrons) by Aspect. Thus EPR was considered to have been disproved, though it leaves the apparent "action at a distance", called entanglement, which cannot however be used to transmit information, and thus does not violate special relativity.

http://en.wikipedia....iki/EPR_paradox

http://en.wikipedia....ll_inequalities

http://en.wikipedia....um_entanglement

http://en.wikipedia....ki/Alain_Aspect

http://en.wikipedia.org/wiki/Measurement_problem

[jason.p]

Measurement doesn't always need to "alter" the observable. Measurement will put the system in an eigenstate of the operator that corresponds to the observable being measured, and subsequent measurement of the same observable can then yield the same result. For instance, arrange a system in which the spin of an electron in the x-direction is "up"; then subsequent measurement of the spin in the x-direction again yields "up". Information is however completely lost with regard to spin in the y- or z-directions, and any attempt to measure those will mean that one can no longer say that the spin in the x-direction is still "up". That's the uncertainty principle. The principle should be thought of as a restriction on what can simultaneously be known about two distinct observables (e.g. position and momentum, or spin in orthogonal directions) rather than knowledge of a single observable.

Einstein, Podolsky and Rosen said: suppose a particle splits, such that spin is conserved, and if one half has spin "up" (in any given direction) then the other half must have spin "down". We measure spin-x of one particle and find "up", thus concluding that the other must be spin-x "down". But we could instead have measured spin-y or spin-z. According to EPR, there must therefore be a "real" value of spin-x, spin-y and spin-z, and the uncertainty principle merely expresses a restriction on what we can know of these values. The alternative, they say, is that when spin-x is measured, a signal must propagate instantaneously to the other particle, telling it which value to take for spin-x. This "spooky action at a distance", they said, would violate special relativity.

Bell looked at the probabilities that would arise for certain states measured in this way, according to quantum theory, thus deriving the "Bell inequalities", which were verified in an experiment (using photons rather than electrons) by Aspect. Thus EPR was considered to have been disproved, though it leaves the apparent "action at a distance", called entanglement, which cannot however be used to transmit information, and thus does not violate special relativity.

http://en.wikipedia....iki/EPR_paradox

http://en.wikipedia....ll_inequalities

http://en.wikipedia....um_entanglement

http://en.wikipedia....ki/Alain_Aspect

http://en.wikipedia....urement_problem

Thanks for that Acey. I've read up on the Bell and Alain/Aspect experiments some time ago as I found entanglement fascinating, if not mind bending!

The reference I was looking for may have been a more recent revision of those experiments. I will continue to search. The trouble with this is that I still approach QM from an intuitive level and can't get my head round "uncertainty" so I want things to be real and measurable even if I can't see or touch them. There's an awful lot to learn!

Jason

[jason.p]

Maybe this one

http://www.scientificamerican.com/article.cfm?id=heisenbergs-uncertainty-principle-is-not-dead

or:

http://www.sciencedaily.com/releases/2012/01/120116095529.htm

[nameunknown]

May have been this?

http://www.theregister.co.uk/2013/02/01/cambridge_boffins_doubt_quantum_experiments/

[acey]

Maybe this one

http://www.scientifi...ple-is-not-dead

or:

http://www.scienceda...20116095529.htm

Only time for a very quick glance, but I see that the article describes the new work as disproving Heisenberg's original "disturbance theory", i.e. the electron microscope thought experiment he presented in his first paper, which is still used in lecture courses as a way of deriving the principle. But the disturbance theory was already condemned by Bohr, on the grounds that it presupposed some "real" value for the observables which became disturbed by measurement, rather than saying (as in the Copenhagen Intepretation) that no "real" value exists prior to measurement. There are other ways of deriving the uncertainty principle that do not make use of the "disturbance" notion: the article quotes the work of Kennard and Robertson, and the more recent work that seems to have been borne out by experiment. So the upshot is that the uncertainty principle is correct, but Heisenberg's original way of proving it was wrong.

[ArmyAirForce]

I thought I understood the uncertainty principle, but now I'm not so sure!

[mytola]

It's insane how fast knowledge acquired at uni attenuates when not used, but I seem to remember that decoherence could be argued to be a severe obstacle to implementing useful quantum bits, and the decoherence times (apparent attenuation time of quantum properties) for a typical two state oscillating molecule even due to vacuum fluctuations was ~10^-18 s. Is that's what they're pushing at here?

[Nillchill]

I think this is the link you all need to read

http://www.sciencedaily.com/releases/2013/03/130303154958.htm

[jakeybob]

There are ways to work around the uncertainty principle to a certain extent.

For instance, if you observe a "quantum non-demolition" observable like momentum as opposed to position. The uncertainty principle states that deltaX * deltaP = h (give or take some factors of 2 or whatever . So, imagine you are trying to measure some quantum system - if you measure x (the position) you necessarily impart some momentum P. This means that when you take your next position measurement the object will have moved unexpectedly, due to the momentum you imparted by observing it. If you can directly sense the momentum, then all that happens is you impart a change in position. This change in position has no effect on successive measurements of the momentum.

http://en.wikipedia.org/wiki/Quantum_nondemolition_measurement

You can also do some funky things using "squeezed" light. This light has more noise in one quadrature (e.g. amplitude or phase, which are analogous to position and momentum in terms of the uncertainty principle) than "regular" light, where the noise is uncorrelated. This allows you to make e.g. extra sensitive measurements of light phase using an interferometer. This has been done with success on table-top experiments, and has been employed on a large-scale in the GEO600 gravitational wave interferometer ( http://arxiv.org/pdf/1109.2295v1.pdf ), with similar squeezing and QND techniques likely to be implemented in future detectors in order to make measurements below the Standard Quantum Limit set by the uncertainty principle. No laws are being broken here, it's more that the "standard" limit doesn't apply to special cases

[jason.p]

I think I'm rapidly getting out of my depth. Been treading water for some time, now I'm going under

[nameunknown]

hi Jason,

see also the "three polariser" experiment here: http://www.informationphilosopher.com/solutions/experiments/dirac_3-polarizers/

its a "WOW" quantum effect you can see with a few optical elements you might have lying about

P

[J_M_Franklin]

I seem to remember recently reading somewhere that someone had found a way to do quantum observations in which the procedure doesn't alter the outcome. A sort of indirect approach. I'm just reading Brian Greene's Fabric of the Cosmos (again!) and just got to the "uncertainty bit" and wonder if this alters, in any way, the basis of understanding of the uncertainty principle. Can anyone point me to a reference (whether it's actually there or not )

Thanks

Jason

I'm uncertain about this

[Tiki]

..... The trouble with this is that I still approach QM from an intuitive level and can't get my head round "uncertainty" so I want things to be real and measurable even if I can't see or touch them. ....

The double-slit experiment (eg. with electrons) compels us to dispense with the notion that electrons move in paths. In quantum mechanics there is no such concept as the path of a particle. This forms the content of what is known as the 'uncertainty principle'. No paths and threrefore classical mechanics alone cannot describe our observations.

[jason.p]

The double-slit experiment (eg. with electrons) compels us to dispense with the notion that electrons move in paths. In quantum mechanics there is no such concept as the path of a particle. This forms the content of what is known as the 'uncertainty principle'. No paths and threrefore classical mechanics alone cannot describe our observations.

This illustrates the problem I have with my "intuitive" approach. The electron, (or as in the original experiment, the photon), has travelled from the source to the slits so surely it must have had a path. I realise this may be semantics, but to me something that has travelled from point A to point B has followed a path. Of course it's probably unhelpful to think of these particles as "points" where they are in reality waves (or are they strings?)

I've read quite a bit about QM and get an idea of some of the theories but find it difficult to relate that to my view of the "real" world. Perhaps it's just the way my brain's wired

[Tiki]

..... so surely it must have had a path.

The double slit experiment shows that it must have had a combination of paths. ie. no absolute path.

[mytola]

Another way to think of it (as far as my understanding goes, which is limited and rusty) is that there isn't any electron at any definite position before you make a measurement. The system can be described by evolving superpositions of mathematical operators and states in some mathematical space that is not necessarily the same as the space-time we observe in our daily life. When the measurement is done, a probability to find a particle at a certain location can be calculated.

In some interpretations (and there are many) particles only "exist" as a legacy and because we humans need some reference tags to not go mad, but in the mathematical framework of quantum theory they're nothing more than clicks in a measuring apparatus at a local point in the space of possible states. So an extremely brilliant child, skilled in mathematics, who has never heard of a particle, can probably understand the theory without having to use the concept!