Supercurrent can flow forever in a superconductor

[granpa]

it has always mystified me how a supercurrent can flow forever in a superconductor. seems like perpetual motion to me. it ocurrs to me now that the simplest explanation is that there is no actual current at all. that nothing is actually flowing in any supercurrent. that the supercurrent merely represents a new quantum state of matter. (this would also explain how insulators can be superconductors)

what experimental evidence is there that a supercurrent is the flow of actual electrons?

 

[f95toli]

Well, it is not a flow of electrons; it is a flow of Cooper pairs. It is not the same thing (and there are devices where you can tell the difference between a quasiparticle current and a the supercurrent).

The obvious "evidence" that there really is a current is that if you hook up a superconducting wire across the terminals of a current source there will be a current flowing without any voltage appearing across the terminals.
Another -perhaps better- demonstration is that the magnetic field of a superconducting coil does not change when you disconnect it from the source and put it in persistent current mode(as long as you short the terminals at the same time, this is usually done with a heat switch). The field will in theory stay the same forever; in reality there are always some losses (due to mechanisms that are not directly related to the superconductivity) but the field decays extremely slowly in a well made coil.
Anyway, the point is that the current is "real", from a practical point of view this means that most of the usual E&M formulas etc work; in many cases (but not all) the superconductor can just be treated as a lossless metal.

And in a way it IS "perpetual motion", although obviously no energy is "created"; as soon as you try to do work using the current it will decay.
In many ways it is similar to an object traveling in space; in a perfect vacuum and in the absence of a gravitational field an object will travel forever at the same velocity.

 

[granpa]

thank you for the reply.

The obvious "evidence" that there really is a current is that if you hook up a superconducting wire across the terminals of a current source there will be a current flowing without any voltage appearing across the terminals.

this is what I am looking for. I am hypothesizing that a supercurrent and a flow of electrons might be different things therefore it has to be a 'current' source not a 'supercurrent' source. for instance, a battery would work but a transformer would not. I am looking for evidence that if you force actual electrons through the superconductor that no voltage will appear across the terminals.

Another -perhaps better- demonstration is that the magnetic field of a superconducting coil does not change when you disconnect it from the source and put it in persistent current mode(as long as you short the terminals at the same time, this is usually done with a heat switch).

this is what I am hypothesizing is a new quantum state of matter. it would not therefore require any actual current (flow of electrons) just a supercurrent (which would be something completely different and wouldn't involve any actual motion of anything)

 

[granpa]

wait a minute. did you mean to say 'there are devices where you can tell the difference between a quasiparticle current and a regular current[/color]'?

if so then what sort of device? (josephson junctions?)

 

[f95toli]

this is what I am looking for. I am hypothesizing that a supercurrent and a flow of electrons might be different things therefore it has to be a 'current' source not a 'supercurrent' source. for instance, a battery would work but a transformer would not. I am looking for evidence that if you force actual electrons through the superconductor that no voltage will appear across the terminals.

I am not sure I understand what you are saying here; if you are asking if you can drive a superconducting circuit via a transformer the answer is yes (and you can also build superconducting transformers, quite a common application actually). You can also couple them using a capacitor.

Anyway, you are correct in saying that a supercurrent is not a flow of electrons. it is a flow of Cooper pairs. The current carrying "particle" in a superconductor has charge 2e; not e.

this is what I am hypothesizing is a new quantum state of matter. it would not therefore require any actual current (flow of electrons) just a supercurrent (which would be something completely different and wouldn't involve any actual motion of anything)

You have to be a careful here, there are all sorts of subtle problems that crop up if you try to imagine ANY electrical current to be a flow of "real" electrons even in normal conductors, electron transport in real materials is strictly speaking always a flow of quasiparticles; in metals that is often essentially the same thing as a free electrons but in e.g. semiconductors you also have holes. There are also higher order excitations that can carry current that are not directly related to any single particle phenomena.

It is therefore pretty difficult to create conditions where we can see actual charges "moving" but it can be done. Electron pumps is a good example where we can see single electrons being pumped from one island to another. The same thing can be done with Cooper-pairs and we then see something with charge 2e being transported instead.

 

[f95toli]

wait a minute. did you mean to say 'there are devices where you can tell the difference between a quasiparticle current and a regular current[/color]'?

if so then what sort of device? (Josephson junctions?)

No, quasiparticle currents ARE what we often refer to as "normal" current.
It is very difficult to see these effects in single Josephson junctions (they need to be very small and well insulated), it is much easier to combine two junctions into a single electron transistor (SET). If you measure a superconducting SET you can sometimes see peaks in the dynamic conductance at e and 2e, the latter disappear if you suppress the superconductivity using a magnetic field.

Maybe I should point out that the reason why you see peaks at BOTH e and 2e is that there are always residual quasiparticles even in the superconducting state; the number scales roughly as \sqrt{1-^(T/T_c)^2^}.
This is the reason why we have to cool superconductors well below Tc in order for them to behave "almost" as ideal conductors; this is especially evident at higher frequencies (a few GHz) where the losses can be quite substansial even at 0.5Tc. In a SET there are also other mechanism that can lead to pair-breaking and therefore the appearance of peaks at e.

 

[granpa]

I think that most people would agree that the electron doesn't really spin. it just behaves as though it did. I'm thinking that the supercurrrent is like that. the cooper pairs don't (necessarily) move. they just behave as though they did.

 

[ZapperZ]

I think that most people would agree that the electron doesn't really spin. it just behaves as though it did. I'm thinking that the supercurrrent is like that. the cooper pairs don't (necessarily) move. they just behave as though they did.

Have you found the average position of such long-range coherent copper pairs to deduce that they "don't move"?

Zz.

 

[sokrates]

You have to be a careful here, there are all sorts of subtle problems that crop up if you try to imagine ANY electrical current to be a flow of "real" electrons even in normal conductors, electron transport in real materials is strictly speaking always a flow of quasiparticles; in metals that is often essentially the same thing as a free electrons but in e.g. semiconductors you also have holes. There are also higher order excitations that can carry current that are not directly related to any single particle phenomena.

How does simplifying the quantum many-body problem change a fundamental fact about the electron transport?

Current flow is the flow of electrons, no matter how you analyze it.

The concept of "holes" in semiconductors is a MERE convention where counting filled electron states would be computationally much more intensive than counting empty electron states.

At the end of the day, it's all electrons. I am a theorist working on current flow, and I don't see any reason as to why I have to be careful if I imagine current flow as a flow of real electrons.

You have to remember that the "quasi-particle" concept has to do with nothing but OUR way of modeling electrons.

If you write, for instance, the full Bloch waves and a real atomistic Hamiltonian, the current is REALLY that of REAL particles.

But for the sake of simplicity if you choose a "smoothed out" Hamiltonian where an effective mass description is chosen, then, yes it's a quasi-particle current.

But can a fact of nature by altered by our understanding and analysis? It's the same electrons that are moving and causing the current.

 

[f95toli]

I am not sure what you are getting at here. I did not say that an electrical current was NOT due to a flow of electrons (I am an experimentalist, and having spent most of my career doing transport measurements is do know at least some of the fundamental stuff...), merely that when you DO use a "first principles" model for real materials one needs to keep in mind the various many-body effects etc that come into play and that can be extremely important (as in e.g a semiconductor where electrons and holes have different effective masses, even thought the latter obviously are not "real). One advantage of using concepts like quasiparticles etc is that we avoid many of these problems.
Hence, for someone trying to understand the basic physics it is probably not very useful to take a first-principles approach to current transport and think about single "real" electrons.

Also, there is another point here which is important (which is what I was trying to get across above): In many cases these "composite particles" BEHAVE as if they were real: We never see a Cooper pair behave as two individual electrons (unless we split it, but then it can't contribute to the supercurrent); in all experiments we see a charge carrier of charge 2e. The same is true for holes (just think of a Hall effect measurement of an p-doped semiconductor, the charge carrier is positive) and -more generally- higher order excitations.

Hence, while we should of course be aware that there really isn't anything in a solid except ions and electrons this does not change the fact that collective effects are in many cases just as important as the particles.

 

[sokrates]

I am not sure what you are getting at here. I did not say that an electrical current was NOT due to a flow of electrons (I am an experimentalist, and having spent most of my career doing transport measurements is do know at least some of the fundamental stuff...), merely that when you DO use a "first principles" model for real materials one needs to keep in mind the various many-body effects etc that come into play and that can be extremely important (as in e.g a semiconductor where electrons and holes have different effective masses, even thought the latter obviously are not "real). One advantage of using concepts like quasiparticles etc is that we avoid many of these problems.
Hence, for someone trying to understand the basic physics it is probably not very useful to take a first-principles approach to current transport and think about single "real" electrons.

Also, there is another point here which is important (which is what I was trying to get across above): In many cases these "composite particles" BEHAVE as if they were real: We never see a Cooper pair behave as two individual electrons (unless we split it, but then it can't contribute to the supercurrent); in all experiments we see a charge carrier of charge 2e. The same is true for holes (just think of a Hall effect measurement of an p-doped semiconductor, the charge carrier is positive) and -more generally- higher order excitations.

Hence, while we should of course be aware that there really isn't anything in a solid except ions and electrons this does not change the fact that collective effects are in many cases just as important as the particles.

I was not proposing to throw away all the useful concepts defined in terms of quasi-particles because as you have suggested, in some cases such as Cooper pairs, the concepts are indispensable.

I respect you as an experimentalist and I think you have a broad knowledge in fundamentals and I usually follow your posts.

I was simply trying to point out the "simpler" approach for the layman. Nobody of course remembers the exact details of the nuclear potentials, or the first principles when thinking about current flow, but introducing another concept such as quasi-particles is just another complication!

Even the very simple effective mass approach carries a whole lot of information in itself, and covering it in a theory class that takes a whole semester.

I think Cooper pairs deserve a special place because they really don't behave as real electrons, but most of the other "quasi-particle" concepts are merely our tools of interpretation.

So, usually, it's OKAY to say that it's the same old "real" electron (whatever that is) but with slightly different properties ( mass, statistics, etc..) to account for the second-order effects.

I was trying to suggest that a bottom-up picture is clearer and more understandable.