Understanding Unpolarized Current & Light

[sokrates]

This came out of a discussion on spin transfer torque effect, and I realized that I do not fully understand "unpolarized" current (or light for that matter)

When you say you have unpolarized current (at least in spin electronics community) people assume that you have a current that is NOT polarized in any spin direction...

And ordinarily, if you pass an unpolarized current through a magnet your current becomes (ideally) fully spin-polarized along the direction of the magnet...

My question is this: How do you think about unpolarized current? Is it really that on average half of the electrons are pointing 'up' (say relative to the magnet's direction) and half of them are pointing 'down' ?

OR is it that every electron's spin is pointing in some funny direction and they get polarized once they get through the magnet?!

These two actually make a lot of difference for subtle reasons, so I wanted your input on this.

What is really unpolarized charge current? (or very similarly unpolarized light?)

Thanks,

 

[kote]

My question is this: How do you think about unpolarized current? Is it really that on average half of the electrons are pointing 'up' (say relative to the magnet's direction) and half of them are pointing 'down' ?

There is no spatiotemporal way to think about spin that matches reality without requiring nonlocal signaling between entangled particles.

You could think of force being quantized and any electron more up than it is down will feel the same "up" force and vice versa. But this is wrong. Spin can't be defined geometrically without action at a distance, which means it can't be defined at all in relativistic space-time.

This is the "emancipation from the demand for visualization" that Bohr refers to. You can't think about spin in a coherent way.

Someone correct me if I've missed something here.

 

[Naty1]

This regards semiconductors,right?? try Wikipedia as a start:
http://en.wikipedia.org/wiki/Spintronics

which discusses orientation of electron spin...say by manipulating electron spin via magnetic fields...

Light polarization reflects regards the direction of oscillation of light waves. Most sources produce uncorrelated electromagnetic waves...unpolarized light...Try Wikipedia for a brief discussion at http://en.wikipedia.org/wiki/Polarized_light

Or is your question more subtle??

 

[Naty1]

I see kote posted while I wrote...yes as he says
I think spin has no classical equivalent, but I thought it was the source of a magnetic dipole moment...

yes, that's what Wikipedia suggests: http://en.wikipedia.org/wiki/Electron_spin

am unsure just what your issue is...check the wiki sources for a quick read and return with specifics...

in general polarized means "aligned" somehow or other and unpolarized reflects a random aligment of a physical feature...

 

[sokrates]

There is no spatiotemporal way to think about spin that matches reality without requiring nonlocal signaling between entangled particles.

You could think of force being quantized and any electron more up than it is down will feel the same "up" force and vice versa. But this is wrong. Spin can't be defined geometrically without action at a distance, which means it can't be defined at all in relativistic space-time.

This is the "emancipation from the demand for visualization" that Bohr refers to. You can't think about spin in a coherent way.

Someone correct me if I've missed something here.

Well, I don't know about Bohr's philosophy, but let me tell you this, spin, as well as spin-currents, as well as spin-torque, are VERY real, we use directions to imagine spin-currents almost daily, LLG equation is solved based on these, people are making M-RAMs out of these etc.. etc.. etc...

There's a huge field out there working on these things, visualizing spin. So I didn't understand anything about your technical theoretical argument which says nothing to me. What action at a distance are you talking about? I described some very basic things...

 

[sokrates]

This regards semiconductors,right?? try Wikipedia as a start:
http://en.wikipedia.org/wiki/Spintronics

which discusses orientation of electron spin...say by manipulating electron spin via magnetic fields...

Light polarization reflects regards the direction of oscillation of light waves. Most sources produce uncorrelated electromagnetic waves...unpolarized light...Try Wikipedia for a brief discussion at http://en.wikipedia.org/wiki/Polarized_light

Or is your question more subtle??

I am one of the people editing the Spintronics article in Wiki... Thank you for pointing that out, but I am no beginner in this subject, maybe I missed the point in my initial question.

Electron current could have a well-defined spin direction *and YES, you can VISUALIZE this, and there's nothing wrong with that.

Magnets have a direction and that comes from the electron spin direction!

Spin has a very very close analogy to the polarization of light (see Datta-Das paper 1990 APL for instance) that's why I brought it up...

 

[sokrates]

I see kote posted while I wrote...yes as he says
I think spin has no classical equivalent, but I thought it was the source of a magnetic dipole moment...

yes, that's what Wikipedia suggests: http://en.wikipedia.org/wiki/Electron_spin

am unsure just what your issue is...check the wiki sources for a quick read and return with specifics...

in general polarized means "aligned" somehow or other and unpolarized reflects a random aligment of a physical feature...

My question is as specific as it could get. What is unpolarized current? {in terms of spin direction of individual electrons that constitute that current!}

Half up, half down (if so, with respect to what basis) and if not, is it the collection of a randomly oriented spins...??

If that's not clear to you, you must check with the basics yourself.

 

[Cthugha]

What is really unpolarized charge current? (or very similarly unpolarized light?)

Let me give an answer on the unpolarized light part of the question. I am more familiar with that branch of physics (I should know a lot more about spintronics than I actually do :/ ).

From a classical point of view unpolarized light one usually thinks of a transverse light field, which can be decomposed into two linearly polarized modes. If the second order moments of these modes are invariant under any geometrical rotation and any differential phase shift (of course also under combinations of both), the light is considered unpolarized.

From a quantum point of view, the definition is similar, but more severe. Any unpolarized state must be invariant under any geometrical rotation around its direction of propagation and any differential phase shift. Although the definition looks extremely similar to the classical case, in this case not only the second order moments, but the moments of all orders must be invariant. This definition has first been mentioned by Prakash and Chandra (H. Prakash and N. Chandra, Phys. Rev. A 4, 796 (1971).) and has been clarified by Agarwal (G. S. Agarwal, Lett. Nuovo Cimento 1, 53 (1971).). More modern definitions have been given in Phys. Rev. A 53, 2727 - 2735 (1996). Another approach can be found on Arxiv in case you do not have access to peer-reviewed magazines: http://arxiv.org/abs/quant-ph/0007099.

I suppose an analogous definition can be given for unpolarized charge current.

 

[kote]

My question is as specific as it could get. What is unpolarized current? {in terms of spin direction of individual electrons that constitute that current!}

Half up, half down (if so, with respect to what basis) and if not, is it the collection of a randomly oriented spins...??

If that's not clear to you, you must check with the basics yourself.

My answer was to this question exactly. If either of these definitions were accurate, classical probability distributions would match QM predictions for EPRB type Stern-Gerlach experiments. I can't find any good links for a description of these, but it all stems from Bell. Neither option you give is possible. No spatiotemporal description is possible. You can get away with pretending that spin is different things in different situations, but there is no one answer that works consistently.

http://en.wikipedia.org/wiki/Bell's_theorem

 

[sokrates]

Let me give an answer on the unpolarized light part of the question. I am more familiar with that branch of physics (I should know a lot more about spintronics than I actually do :/ ).

From a classical point of view unpolarized light one usually thinks of a transverse light field, which can be decomposed into two linearly polarized modes. If the second order moments of these modes are invariant under any geometrical rotation and any differential phase shift (of course also under combinations of both), the light is considered unpolarized.

From a quantum point of view, the definition is similar, but more severe. Any unpolarized state must be invariant under any geometrical rotation around its direction of propagation and any differential phase shift. Although the definition looks extremely similar to the classical case, in this case not only the second order moments, but the moments of all orders must be invariant. This definition has first been mentioned by Prakash and Chandra (H. Prakash and N. Chandra, Phys. Rev. A 4, 796 (1971).) and has been clarified by Agarwal (G. S. Agarwal, Lett. Nuovo Cimento 1, 53 (1971).). More modern definitions have been given in Phys. Rev. A 53, 2727 - 2735 (1996). Another approach can be found on Arxiv in case you do not have access to peer-reviewed magazines: http://arxiv.org/abs/quant-ph/0007099.

I suppose an analogous definition can be given for unpolarized charge current.

Yeah, with "minor" differences, polarization of light is very similar to spin...So you probably latently know a whole lot about spintronics :) Jones calculus (I think used to analyze polarization) and spinors that are used to describe spin almost look exactly same. There are spintronic proposals in the field where they draw analogies to "electro-optic modulators" using the same equations for polarization...

Rotation invariance gave me a lot of hint, and I'll definitely follow up with the articles you pointed out.. And it made it clear that there's no very simple answer to what I asked -- I think I just got developed too much classical intuition while working with the macro-spin models and LLG dynamics of magnets :)
Thanks for the responses...

 

[sokrates]

My answer was to this question exactly. If either of these definitions were accurate, classical probability distributions would match QM predictions for EPRB type Stern-Gerlach experiments. I can't find any good links for a description of these, but it all stems from Bell. Neither option you give is possible. No spatiotemporal description is possible. You can get away with pretending that spin is different things in different situations, but there is no one answer that works consistently.

http://en.wikipedia.org/wiki/Bell's_theorem

OK. You are probably correct, this might essentially be the reason why we have to first POLARIZE the electrons before we can do any kind of spin-torque reversal.

See, there's also an engineering aspect of the problem in contrast to theory, and I must somehow connect the two without violating physics and without denying real experiments.

Thanks for the brainstorming