Does current flow through a capacitor?

[richrussell]

In other words then - direct current.

[ms660]

DC...that's another can o'worms A capacitor is connected across a DC supply, does DC current flow through the capacitor during the time it takes for the current in the circuit to become zero.

Lawrence.

[G8HQP Dave]

Originally Posted by IanBland

Not according to the Copenhagen Interpretation they don't.

Even in Copenhagen the cat doesn't disappear when we don't look; we just don't know whether it is alive or dead. I am not in Copenhagen.

Originally Posted by dseymo1

I thought that the reason for quarks' having frivolous-sounding names was as a reminder that they don't 'really' exist.

I thought the reason was as a reminder that physicists have a sense of humour too. Quarks almost certainly exist, but I guess the evidence for this may be slightly weaker than the evidence for the existence of electrons.

Originally Posted by pwdrive

What I mean by current and flow (as per the original question) is current that flows along a conductor as that's how I've always imagined to be.

If by 'current' you mean a flow of charge carriers then no, current does not flow through a capacitor. You have a restricted definition of current. This means you have to explain where the current goes to/comes from as there is certainly a current (of your definition) in the capacitor leads. The fact that it creates a problem shows that your definition of current is flawed. This is how science advances: we find that a simple idea creates a problem so the idea has to be extended.

[ms660]

Some bloke once said a quark is the sound of a posh duck.

The consensus on here and in articles on the web seems to be that the current doesn't flow through the dielectric unless the dielectric has broken down.

Lawrence.

[daviddeakin]

Yes and no. Current does not really 'flow' anywhere. It is the rate of change of charge; only a number; a scalar.
Charge is like a cloud or sphere of influence that enshrouds each electron. Charge, therefore, does penetrate into and across the capacitor, even though we say the electrons can't cross the gap themselves. Therefore, there can be a rate of change of charge within the dielectric or indeed anywhere in the circuit (and into the universe beyond, though only very weakly!). If you want to call that change 'flow' then yes, current 'flows' through the capacitor. But you have to remember that no physical 'substance' is actually travelling across the capacitor or through the wires.

[G8HQP Dave]

Originally Posted by daviddeakin

It is the rate of change of charge; only a number; a scalar.

No. Current is a vector; it has direction too. Current started as a flow of charge; now we know we have to extend it to include displacement current too. Flow is a vector.

Charge is like a cloud or sphere of influence that enshrouds each electron. Charge, therefore, does penetrate into and across the capacitor,

No. That is electric field.

Originally Posted by pwdrive

The consensus on here and in articles on the web seems to be that the current doesn't flow through the dielectric unless the dielectric has broken down.

It all depends on what you mean by current. Take the limited definition and AC circuit theory suddenly gets a lot harder as we can no longer use Kirchoff's laws.

[Radio Wrangler]

The concept of DC opens another can of worms.

Pure DC with no AC component can only be a current which started at the beginning of time and will continue to the end, without any change. All other currents have non-zero AC components.

There is an idea that any varying function can be expressed as the sum of a series of sinusoidal components of assorted frequencies, amplitudes and phases. One of those components will be a zero-Hertz or DC component. The other way to put this is that a varying thing can be expressed either in terms of variation with time, or as a set of spectral components. The thing doing the varying is the same, the only difference lies in the way of looking at it.

You aren't tied to expressing a signal as an equivalent set of summed sinewaves. Many other series of orthoganal functions can do the same trick, for example Walsh functions.

All this culminates in the phrase "Fast Bass" causing people involved in signal and information theory to have a giggle fit.

David

[kalee20]

So, with pwdrive's definition of flow, no, current doesn't flow through a capacitor.

[ms660]

"In studying the charging process of a capacitor, you must be aware that NO current flows THROUGH the capacitor. The material between the plates of the capacitor must be an insulator. However, to an observer stationed at the source or along one of the ci
rcuit conductors, the action has all the appearances of
a true flow of current, even though the insulating mate
rial between the plates of the capacitor prevents the
current from having a complete path. The current wh
ich appears to flow through a capacitor is called
DISPLACEMENT CURRENT."

I'm going with the above quote for the moment.

Lawrence

[G8HQP Dave]

It would be more accurate to say that no charge flows through a capacitor. The simple definition of current is the flow of charge. We don't have a flow of a flow, so strictly speaking no current flows anywhere in a circuit - it is the charge that flows. However, in English (both technical and everyday) we get sloppy so we know that a river is a flow of water but we also say that a river is flowing - which is tautology. If it stopped flowing it would cease to be a river!

Anyway, you choose:
1. simple definition of current, no current through a capacitor, no AC Kirchoff's laws, difficult circuit design, current appears and disappears
2. extended definition of current, current through a capacitor, AC Kirchoff's laws, easy circuit design, current always in loops

[dseymo1]

An interesting point here is that it's the flow of charge, not of charge-carriers, which contitutes the current.
Although DC 'flows' at something approaching the speed of light (in an ideal conductor), the electrons themselves travel very slowly in comparison, and may never get from one end of a length of wire to the other.
Thinking about it that way, while it's clear that no electrons pass through the dielectric, I'd be fairly happy to say that the charge on one plate 'travels' to the other (where else does it come from?).

[G8HQP Dave]

Flow of charge and flow of charge-carriers is the same thing. Charge is always carried by carriers. You need to carefully distinguish between propagation speed (of EM fields) and drift speed (of electrons). A tiny movement of electrons can create a huge charge change, because there are so many electrons.

No charge travels from one plate to the other. What happens is that the charge put onto one plate causes a redistribution of charge on the other plate.

[dseymo1]

My (quite possibly faulty) understanding is that electrons act rather like the ball-bearings of a Newton's cradle. The charge is in effect transferred through the 'chain', rather than being carried by individual electrons wandering along the wire. Suppose a Newton's cradle had at its centre a fixed tennis ball, through which no bearings could, of course, pass. The energy of the bearings would nevertheless pass through the ball to the other side of the cradle (OK, it would be pretty lossy, but you see what I mean).
Have I got the wrong end of the stick?

[Okto1984]

As I understand it current flows through a capacitor, but not individual electrons. The main problem seems to be how we define current. Most explanations I've seen appear to be correct, regardless if they say yes or no, provided we use the writer's definition of current.

To expand that a bit, this is generally how I think about capacitors (apologies if it's a bit simple, but it works for me):

Capacitors are constructed much like a sandwich. There's two plates, each connected to one of the pins, and between them is some kind of dielectric layer which may vary in composition, but is always an electrically non-conductive material. Because of this dielectric layer separating the plates there's no direct electrical connection and electrons don't flow through a capacitor in the same way they would through a wire (except for a small leakage current). Despite this, current flows. It happens because electrons repel other electrons in their vicinity, and by enriching or depleting electron concentrations in one plate which is very close to another plate, you can transfer charge via the electron repulsion alone. Push some in one end of the capacitor and they build up on the plate connected to that wire, which repel some out of the other plate and through its wire. Taking electrons out of the plate also works, as some will pull in to the other plate now that there's less nearby electrons repelling them out. Thus the charge is transferred and a current is present. However, because the current flow relies on charge being transferred between the capacitor's plates, and the maximum charge is limited by the capacitance of the capacitor, a DC current cannot be sustained forever. The capacitor will eventually become fully charged and the transfer of charge and the current it induces stops. This is how they block DC. An AC current however, assuming it never fully charges the capacitor before changing direction, can stay below this limit and effectively be transferred via charge through the capacitor largely unchanged.

So, the issue seems to be, does the conversion of current to charge and back again fit within your definition of current flowing or not? Personally, it fits mine, but I can understand how it might not fit someone else's. Everything we understand is just a model in our minds and we have to use the ones that work for us. Accepting the diversity of answers helped me find a useful way to think about it for myself.

[daviddeakin]

Originally Posted by G8HQP Dave

No. Current is a vector

No, current density is a vector. Current is a scalar. Current cannot be a vector any more than the speed of your car can be a vector.

Originally Posted by G8HQP Dave

No. That is electric field

No. Gauss' law tells us that the electric flux 'flowing' out of a Gaussian surface is equal to the charge enclosed, i.e. charge and flux are equatable, indeed, they are interchangeable. Hence charge is a sphere of influence without boundary.

[G8HQP Dave]

The fact that we can calculate one thing from another does not mean that they are the same thing. The electric flux is the effect of charge, not charge itself. From adding up the total effect we can calculate how much charge there is. The charge stays firmly on the charge carrier; its effects go off to infinity.

[ms660]

I've found a solution...."via" not "through" that model fits the current state of my brain ('scuse pun)

Lawrence.

[AC/HL]

This thread reminds me of Night School. Five to nine, lessons understood, books packed, and I'm just thinking I can get the earlier bus when someone on the back row (it always seemed to be the back row) says "But".