Thread: perfect dielectric

We have observed that once a superconductor reaches it's transient temperature , both the E field inside drops and the B field is expelled from the conductor.Different alloys and metals have different temperatures when they reach superconducting phase.

Now I wonder , is there any material that under certain pressures or temperatures or other factors could become a perfect dielectric , in other words a material that would expell any electric field near it and not let any E field through itself?

anyone please?

thank you

Just to let you know, if you don't get a reply it's likely because no one that has read your post has a good answer for you. (On the plus side, it also means that no one thought your question was just wrong.)

Hi. Perfect dielectrics doesn´t exists.

well a superconductor would be a perfect dielectric as it would expell the E field from it, so they exist.

Originally Posted by Crazymechanic

well a superconductor would be a perfect dielectric as it would expell the E field from it, so they exist.

I've hesitated in answering partly because your definition of "perfect" seemed to embed criteria that are foreign to me. And it remains that we have very different notions of what "perfect" would mean for a dielectric.

One level of ideality would be that the dielectric be completely free of dissipation. If one expresses the dielectric constant as the sum of a real and imaginary part, dissipation is zero if the imaginary part is zero.

Another requirement would be for the dielectric to be an insulator (otherwise we'd call it a conductor). If you accept that requirement, then a superconductor cannot be a perfect dielectric, because it's not a dielectric. Merely expelling an interior electric field does not give us an insulator, for a superconductor supports conduction with zero electric field (think of the field-theoretical cognate of Ohm's law: J=sigma*E, where J is the current density in A/m^2, sigma is the conductivity of the material in S/m, and E is the electric field in V/m).

Perhaps you could clarify what you mean by the term perfect as it applies here. Then it would be easier to have a discussion.

Originally Posted by tk421

One level of ideality would be that the dielectric be completely free of dissipation. If one expresses the dielectric constant as the sum of a real and imaginary part, dissipation is zero if the imaginary part is zero.

Well said tk. I will just add for Crazymechanic that this microscopicaly means that you have no absorption nowhere which is impossible in any material.

Well , yes a superconductor is a conductor so that eliminates it being a dielectric as it would not insulate current.
tk421 I was thinking about a material that once placed between the two plates of a capacitor for example would be able to block the E field from the plates completely so that no charge would reside on them.

But another confusion is that vacuum has the lowest permittivity or in other words lets the E field travel through freely.Other materials with higher permittivity weaken the E field so the charges per unit area on the plates increase to compensate for the weaker field right?
But if this is true then as we go up in relative permittivity that would imply the charge per unit area to also increase , so then does a perfect dielectric or since nothing is perfect near perfect dielectric would mean a enourmous permittivity and hence huge capacitance (charge per unit area)?

Also a thought i have been thinking for a while , does any such material exists which could be able to change its dielectric permittivity and so the capacitance of a capacitor would change ?

Originally Posted by Crazymechanic

Well , yes a superconductor is a conductor so that eliminates it being a dielectric as it would not insulate current.
tk421 I was thinking about a material that once placed between the two plates of a capacitor for example would be able to block the E field from the plates completely so that no charge would reside on them.

But another confusion is that vacuum has the lowest permittivity or in other words lets the E field travel through freely.Other materials with higher permittivity weaken the E field so the charges per unit area on the plates increase to compensate for the weaker field right?
But if this is true then as we go up in relative permittivity that would imply the charge per unit area to also increase , so then does a perfect dielectric or since nothing is perfect near perfect dielectric would mean a enourmous permittivity and hence huge capacitance (charge per unit area)?

Also a thought i have been thinking for a while , does any such material exists which could be able to change its dielectric permittivity and so the capacitance of a capacitor would change ?

I think it would be better to avoid using the word perfect altogether, as you haven't explicitly defined what you mean. There's no universally acknowledged definition that applies here, as I pointed out earlier. What constitutes perfection might vary with context and with the eye of the beholder.

Here it seems that you mean a dielectric with infinite permittivity. No such material exists, but there are certainly materials whose permittivity exceeds that of free space by many orders of magnitude. Also, the permittiivity of nearly all materials depends on field strength, allowing changes in small-signal capacitance. This behaviour is sometimes exploited to make voltage-controlled capacitors (although semiconductor junctions are used much more often).

As for your confusion, I am confused by why you are confused (aside from a small circularity in your chain of reasoning), for your analysis is actually correct. The increase in capacitance by a dielectric is due precisely to the mechanism you outline. The reduction in field by the dielectric allows an increase in the amount of displaced charge per applied volt. Since capacitance is the ratio of charge to voltage, capacitance increases. If your definition of perfection is infinite permittivity, then an infinitely permissive (?) dielectric would permit the displacement of infinite charge for an infinitesimal applied voltage.

You see tk421, I'm actually thinking about something like a charge pump, a device that could be something of an analogue to the transistor, only more rugged and much more welcoming for high voltage.
If I have say a many KV aboe ground single wire conductor and then I have a transformer , one side of the winding is grounded while the other say is connected via switch to the incomming high voltage DC wire , Now as I would repeatedly switch the switch on and off induced flux would rise in the transformer and it would work,
What if i could replace the switch with a capacitor whose dielectric permittivity can be varied by external or whatever means on purpose.
It should create the same fluctuations in the transformer because charges would rush towards and drain away from the capacitor if the permittivity would be changed of the dielectric , atleast according to theory ?
In a standalone capacitor which isnt attached to anyting by changing the permittivity the voltage would change since there would be no way for the charges to go but in a cap that is attached to a circuit the charges should flow right ?

I guess the question then is how dos one vary the dielectric etc...

Originally Posted by Crazymechanic

You see tk421, I'm actually thinking about something like a charge pump, a device that could be something of an analogue to the transistor, only more rugged and much more welcoming for high voltage.

Ah, ok. So you are thinking of a high-voltage variable capacitance, whose value can be controlled from zero to infinity. Alas, there aren't even fixed capacitors with infinite capacitance, so that by itself rather puts a damper on things; the lack of significant control of the dielectric constant just makes things worse. One would do much better with a semiconductor device, such as a transistor. The same mechanisms ultimately limit the breakdown voltages of both capacitors and transistors, so there's no fundamental advantage to a dielectric-based technology in that regard.

There are, by the way, small-signal versions of the sort of device you are thinking of. They're called parametric amplifiers, and they function by periodically varying a capacitance (usually that of a semiconductor diode) to mediate the transfer of power from a source to a sink. These once found wide application in radioastronomy as low-noise amplifiers (some of the earliest "high" resolution measurements of the CMB were made with these), before the vastly superior maser came into prominence.

There also exists a dual of a capacitive amplifier -- the magnetic amplifier. These are extremely rugged, and have been used for many decades in applications such as servomechanism control amplifiers for artillery, etc.

I wasn't thinking for infinite capacitance or permittivity or whatnot as such infinities dont exist in the real world , I was just thinking about the mechanism.
You are right about the breakdown voltages that affect both semiconductors like transistors and capacitors and spark gaps and all other devices.
I wasn't focused so much on the breakdown voltages as of the very way a transistor works, aka by a controllable current path, but to my mind the weak point here is the very fact that the physical current has to run through the " semiconducting" material , i mean the applications in microelectronics and CPU's and most household devices are good enough for them as the voltages and currents are manageable.
But in high voltage high current power transmission and other specific applications it gets kinda different.
I like the capacitor more because instead of providing a physical path for the current to run through it uses the attraction and repulsion of charges to achieve the same charge/current flow in wire. So if we just want to make oscillations from a DC source by having a varyable capacitance the only thing that really feels the " pressure" of the current is the wire itself , and as long as i know wires are pretty rugged as compared to semiconductors etc.

In other words if a device that I describe could work in my mentioned applications it could do better and probably longer than semiconductors, what do you think?
The problem ofcourse is , how do we vary a capacitance by doing something with the dielectric.

Could this be what you're looking for?:

Superinsulator

Unless I'm missing something, the OP was correct in such that superconductors do have an infinite dielectric constant (inside the London penetration limit). However, the term "perfect dielectric" is actually used in science circles, and it means 0 permittivity. Recall, permittivity is defined as epsilion = D/E. In the case of the superconductor, E is always zero no matter what the electric displacement is, and in the case of the "perfect dielectric", the electric displacement is always zero no matter what the electric field is. Another way of putting it is that if one were to make a capacitor with the insulative layer made of our "perfect dielectric", it would have a ideal capacitor response. As so far as I know, no such "perfect dielectrics" exist.

Edit: After further thought, I'm not sure I buy my argument. Anyone more knowledgeable want to explain where I went wrong? I know the semiconductor stuff is correct, however.

On the subject of semiconductors: You are getting stuck on the idea that current flows through a "semiconducting" material in a transistor, which is partially true. A semiconductor is a semiconductor because it has a wide bandgap, and there are very few charges in the conduction band available for transport, whereas a conductor has no bandgap (at least, close to the fermi level), so there are filled states going all the way to the Fermi level, meaning that there are a multitude of free charges for transport. The "magic" of transistors is that they use either a electric field (FET) or current (BJT) to promote carriers into the conduction band of the semiconductor, meaning it becomes essentially a good conductor. The great thing is, this is tunable, not an on-off thing, meaning we can use that tunable region to produce gain. If you want to improve transistors, the two things that are holding the semiconductor industry back are low gain from MOSFETS, and tunneling in small size transistors (they are making them 12nm from source to drain now, most people think it is possible to get to ~8nm or so, then tunneling will be so bad that there will be more noise than signal).

Originally Posted by ajarjour

However, the term "perfect dielectric" is actually used in science circles, and it means 0 permittivity.

Yes, as I mentioned earlier, the adjective "perfect" is unhelpful, as perfection is very much in the eye of the beholder. To some, a perfect dielectric is merely a lossless dielectric, independent of the dielectric constant. To others, perfection corresponds to an infinite constant (allowing the creation of infinitesimally small capacitors with infinitely large capacitance). It is thus far better to say "lossless, infinite dielectric constant" or "lossless, zero dielectric constant" as the case may be. In the case of the OP, he means a lossless dielectric whose constant can be changed from (near) zero to (near) infinity under the control of another physical variable. Sadly, as I've said, I'm unaware of any dielectric material with anything close to the characteristics required to make a robust high-voltage ac switch that would be superior to a transistor (or other semiconductor device). The dielectrics with the greatest control range happen to be made of semiconductors, and they function by varying the effective "plate" separation, rather than by varying the dielectric constant.

The magnetic amplifier is the dual of what the OP is asking for. It is extremely rugged, and it depends on operation between low fields, where permeability is large, and high fields, where permeability is low. It is wholly analogous to what the OP is thinking of.

But , I didn't intended to think about some kind of an abstract perfect capacitor or perfect dielectric , nor did i search for a dielectric that could vary from 0 to infinity in its permitivity, as I know such materials cannot exist.
All I am searching for is the very mechanism and how this phenomenon works or do any materials that exhibit atleast partly this phenomenon exist.A small change in capacitance varying the dielctric is good enough.Varactor diodes do the job but their change in capacitanc is too small to be usable for inducing fields in a transformer for reasonable loads.

As I said i;m not after a perfect dielectric , I;m just after something that could work like 10% of one.I mentioned the word perfect because I was explaining the imagined workings of such a device and in many cases people in physics describe an ideal case of something just for the overall picture but when it comes down to a real world device or scenario the abstract words get interchanged for more realistic ones.

The quesion then becomes , is there any material that could vary it's permittivity considerably by any means ?
Maybe a material whose permittivity changes as the E field gets stronger and then it resets back to it's original state , it would turn out something like a self oscillating capacitor

by the way one of you folks mentioned a metal or aka a conductor between the two conducting plates of a capcitor , what would happen in thios scenario because the middle conductor wuld get charge polarized from the two plates yet the E field couldn't penetrate the metal like it does a dielectric so what would that mena for the charges on the plates , would such a capacitor have a large capacitance? Also once charged and the middle conductor polarized what would happen if one were to drain the middle conductor and make it neutral , what would happen to the charges on the plates ?

thank you

Originally Posted by Crazymechanic

But , I didn't intended to think about some kind of an abstract perfect capacitor or perfect dielectric

Then it is unfortunate that you chose a thread title that was guaranteed to produce confusion as to the subject under discussion. It is also unfortunate that your writing continues to be confusing, as in:

A small change in capacitance varying the dielctric is good enough.

followed immedately by

Varactor diodes do the job but their change in capacitanc is too small

And then later by

;m just after something that could work like 10% of one.

A varactor, which you have rejected above, provides changes in capacitance that are much greater than 10%. Ordinary diodes will generally provide a 2:1 capacitance range. Specially designed ("hyperabrupt") varactors can exceed 10:1.

I hope that you can appreciate that your writing is more than a little confusing at times.

I mentioned the word perfect because I was explaining the imagined workings of such a device and in many cases people in physics describe an ideal case of something just for the overall picture but when it comes down to a real world device or scenario the abstract words get interchanged for more realistic ones.

Which is precisely why I pointed out almost immediately after your initial post that the word "perfect" means different things to different people, and therefore ought to be avoided altogether. Just say what you actually mean.

The quesion then becomes , is there any material that could vary it's permittivity considerably by any means ?

I've given you the answer at least twice: "Yes, but..."

Maybe a material whose permittivity changes as the E field gets stronger and then it resets back to it's original state , it would turn out something like a self oscillating capacitor

A dielectric material works by dipole motion. Once the applied field is large enough to align essentially all of the dipoles, the capacitance drops. This saturation occurs for all practical dielectrics. If the dielectric is low loss, then the dipoles will reset just fine (there will be no hysteresis). However, this saturation of dipole moment occurs for field strengths that are typically near the breakdown field of the dielectric material.

by the way one of you folks mentioned a metal or aka a conductor between the two conducting plates of a capcitor , what would happen in thios scenario because the middle conductor wuld get charge polarized from the two plates yet the E field couldn't penetrate the metal like it does a dielectric so what would that mena for the charges on the plates , would such a capacitor have a large capacitance? Also once charged and the middle conductor polarized what would happen if one were to drain the middle conductor and make it neutral , what would happen to the charges on the plates ?

thank you

Inserting a plate between two other plates merely produces two capacitances in series. The capacitance measured between the two outer plates remains unchanged if the third plate is left insulated.

If you earth the middle plate, then the capacitance between the two outer plates drops to zero.

so we charge the outer plates of our middle plate cap it now has a certain capacitance, then we earth the middle plate , the capacitance as you said drops to zero , then we remove the earth connection what happens to the capacitance , does it goes back to the previos state before the earth connection was made to the middle plate ?
Assuming a constant DC source connected to the outer plates of this capacitor.

Excuse me for my confusing writing style sometimes I just find hard to express myself more clearly.

Originally Posted by Crazymechanic

so we charge the outer plates of our middle plate cap it now has a certain capacitance, then we earth the middle plate , the capacitance as you said drops to zero , then we remove the earth connection what happens to the capacitance , does it goes back to the previos state before the earth connection was made to the middle plate ?
Assuming a constant DC source connected to the outer plates of this capacitor.

Excuse me for my confusing writing style sometimes I just find hard to express myself more clearly.

No problem -- I was just trying to explain why you might not be getting answers to the questions that are really at the core of what you're curious about.

In the latest example above, we must take care once again to use terminology precisely. I fear that you might be using "capacitance" and "charge" somewhat interchangeably because you are taking care to specify that a constant DC source is connected to the outer plates. As far as capacitance is concerned, the presence or absence of such a source is completely irrelevant. The capacitance is independent of the charge that resides on the plates. All that matters is the proportionality between charge and voltage (that proportionality being, in fact, the capacitance).

With the middle plate earthed, the outer plates do not "see" each other; the earthed plate acts as a Faraday shield, and prevents changes in voltage of one plate from influencing the field of the other. Hence, the capacitance between the two outer plates is zero.

If the middle plate is insulated from any fixed potential, then the field between the two outer plates is the same as when the middle plate is absent altogether. Hence, the capacitance between the two outer plates is the same as any other parallel-plate capacitor of the same area and spacing; the middle plate acts as if it were not there.

Notice that the foregoing explanation doesn't depend on a DC voltage being applied to the outer plates, say.

well I kinda intuitively know that charge and capacitance are not one and the same. In my own words I would explain that charge is simple the excess amount of electrons or the lack of them depending on which plate we look at in a paralell plate capacitor.
But the more charge per given area the higher the capacitance and also the voltage , because a higher charge concentration per given volume results in higher voltage but also more charges are physically on the plates so the capacitance is also higher. I hope im getting this right , the way one value changes the other etc.
To gain more charges while keeping voltage fixed would imply we need to make a dielectric which has a higher permittivity so that there would be more charges per given area to establish the needed E field flux through the dielectric right?

As for the metal plate in the middle , thanks , I kinda already thought that if the plate is insulated or in other words in mid air just between the two charged outer plates hen it acts as a mirror for the charges and hence can be considered non existing in the frame at which we are looking at.
So basically if I would make a parallel plate cap with a third metal foil inbetween then the only factor determining permittivity and hence the capacitance per given voltage , would be the dielectric that insulates the middle foil from the outer plates ?

So technically if we would have this middle plate capacitor with the third lead coming out attached to the middle plate we would have in any other aspect an oridnary capacitor just that if the middle plate would get grounded time after time the capacitance of our cap would go from a fixed value to zero and back as we would drain the middle plate respectively?
The question is how the charges flow in such scenario ? if we connect the middle plate to ground the charges flow from the middle plate to ground until the plate becomes neutral , but how do the charges on the two outer plates flow and where ? Ii guess it also depends on whether the outer plates are charged and disconnected or charged and connected to a dc source as in this case it would matter probably.

I guess i have bored tk421 to death and all the other ones arent interested in this topic i guess.