Query on dielectric action

[741]

Here is my present feel for the way a dielectric between plates increases capacitance:

The dielectric molecules orient themselves in the field. In so doing they reduce the net field.  If we begin with a pre-charged, isolated air-gap capacitor, then introduce a dielectric, the PD across the plates falls. But the capacitor is isolated, so the charge on the plates is constant. Then from C = Q/V, we see the capacitance has risen.

But (for a non-isolated capacitor, in a circuit): Given that the dielectric reduces the field, does that mean the force acting to push (or pull) electrons to/from the plates is actually reduced - so we would see less 'charge per volt', so Q/V = C is less... it seems contradictory to me.

Unless, that is, the dielectric itself also stores potential energy.

So - is that right - are the dielectric molecules themselves storing some of the capacitor's PE due to their 'correlated orientation', in contrast to the charge build-up on the plates?

[Marco]

So - is that right - are the dielectric molecules themselves storing some of the capacitor's PE due to their 'correlated orientation', in contrast to the charge build-up on the plates?

Generally yes, unless you do something sneaky like making the orientation permanent such as in electrets. The energy is still stored of course, but not as electrical energy of the capacitor.

It will take force to push a high relative permittivity dielectric into a charged plate capacitor.

[T3sl4co1l]

More energy is stored, yes.  The bulk version of capacitor energy is \$ e = \frac{1}{2} E^2 \epsilon \$, i.e. it goes up with dielectric constant (\$ \epsilon = \epsilon_0 \kappa \$).  That's little-e for energy density (which is also a pressure, which gives the force exerted on a charged object), big-E for electric field (magnitude), and \$\epsilon\$ for permittivity.

This is contrary to the magnetic case, \$ e = \frac{B^2}{2 \mu} \$, inversely proportional to permeability.

Tim

[In Vacuo Veritas]

Now move the plates apart.... Or, how to get voltage gain with a "passive".

[CopperCone]

fun question: what if you put a dielectric belt of infinite diameter in between the plates of a capacitor and spun it so that the dielectric in the electric field passes by like a convyer belt?

[T3sl4co1l]

The dielectric is polarized as it enters the electric field, and unpolarized as it leaves.  There shouldn't be any drag on the belt, because the force is conservative.

Drag == friction == loss == nonconservative force.  Compare, for example, a moving belt in a magnetic field: if it is an insulator, there is no induced current, and no force.  If it is conductive (but not perfectly so), there is a drag force, due to eddy currents, and the difference is dissipated as heat in the material.  If it is perfectly conductive, then the field is conservative and there may be a force, but it will be a springy (conservative) force.  Example: if the belt is perforated (so flux can pass through it), and it is positioned in the magnetic field and cooled to become superconducting, then the holes will pin the flux and the belt and magnet will remain locked by a spring force (complementary magnetic dipoles).  If the belt is moved, a force is generated proportional to the change in field; this force is returned when the belt is moved back, i.e., it is conservative, and no loss has occurred.  If the belt continues to move, of course, at some distance, the pinned flux becomes just some magnet out there, while the unmagnetized belt, moved in, repels the magnet the usual amount, so the force doesn't increase further.  It's not a linear force, but it is a conservative force.

Going back to dielectrics, if the material is polarizable but lossy (say, a leaky insulator), then some drag force will arise.  If it is not polarizable, it's literally just vacuum.  If it's polarizable and not lossy (say, a very good conductor), then there will be a springy force.

Note that relative motion through an E-field creates an apparent mixture of E and M: classic Relativity.  So that, as velocity rises, even a very good conductor will still experience induced eddy currents, and cause a drag force.  It's simply the same thing from a different reference frame.

Tim

[CopperCone]

http://te.fisica.edu.uy/fuerza_en_Capacitor.pdf

are there any practical applications?

Here is a more advanced evaluation:
http://web.pdx.edu/~hopl/EJP_1993.pdf

This one is where the force/velocity of the dielectric is controlled

and this I think explains how its conservative in more detail
https://electronics.stackexchange.com/questions/263621/adding-slab-to-a-capacitor

[741]

"polarizable and not lossy".

Hmm... That is an interesting thought. I wonder what an example would be?

You can induce charge on a conducting rod (eg with a charged comb, glass rod etc) assuming it's supended from an insulating thread, and the metal rod will swing around to point at the glass rod. In that case, only the electrons move so the whole object is 'bulk polarised'. This contrasts with a dielectric where many individual molecules align their orientation together.

If I introduce a slab of metal - obviously leaving some air gap each side!-, it's just 2 series capacitors then.

So, the metal does not store PE in the same way as an insulating dielectric - I'm a bit unsure how to think about this. A metal 'slab' reduces the capacitance, the insulator raises the capacitance.

What if I introduce a slab or ferrite - would that act like a normal dielectric?

[CopperCone]

I don't know but does anyone wanna add a capacitor inside of their drill press now?

[T3sl4co1l]

http://te.fisica.edu.uy/fuerza_en_Capacitor.pdf

are there any practical applications?

The force acting on a dielectric gives an electrostrictive effect.  I'm not sure which is stronger, the electromechanical induced motion, or actual electrostriction, or piezoelectricity, in, say, film capacitors.  But it's definitely a thing, polyester capacitors at least can buzz when driven with, say, mains AC.

Or maybe the example I experienced was just poorly constructed and the film was vibrating internally.

Ceramic capacitors definitely do, but that's piezoelectric specifically.  You can tell because one is linear and the other is quadratic.

Tim