What would be the effect on capacitance of capacitor if we replace insulator with semiconductor?

[legend killer]

Hello everyone,
I would like to know what would be the effect on capacitance if we replace insulator with semiconductor. Just curious to know . What you guyz say?

 

[DerStrom8]

A doped semiconductor or a pure semiconductor?

 

[alec_t]

You'd have a varactor

 

[steven7890]

semiconductor is partially conductive material, so the capacitor will have high leakage current between its two plates/leads, that is my very first thought of it.....

 

[DerStrom8]

semiconductor is partially conductive material, so the capacitor will have high leakage current between its two plates/leads, that is my very first thought of it.....

Not quite. Saying a "semiconductor" is a "partially conductive material" is not exactly accurate. Its conductivity is determined by the dopants used.

Alec is right--if you have a P-N junction (P-doped semiconductor joined to an N-doped semiconductor", then you will get a varactor, which allows you to vary the capacitance based on a DC reverse bias voltage.

 

[legend killer]

A doped semiconductor or a pure semiconductor?

Actually I didnt think about it, but that is again interesting question. I would love to discuss both. As steven said, even I thought that capacitor will have leakage current and as a result the capacitance will decrease. But capacitance of capacitor is dependent on the relative permittivity of the material. If insulator SiO2 used whose relative permittivity is 3.9, and if replaced by Si whose relative permittivity is approx 11, then the numbers suggest that the capacitance will increase

 

[DerStrom8]

There would definitely be a lot of leakage. But remember that there is a difference between permittivity and conductivity.

 

[legend killer]

There would definitely be a lot of leakage. But remember that there is a difference between permittivity and conductivity.

Ohh yeaa, that didnt came to my mind. However the equation of capacitance C = εA/d doesnt involve conductivity of the material. Thats starnge

 

[DerStrom8]

Ohh yeaa, that didnt came to my mind. However the equation of capacitance C = εA/d doesnt involve conductivity of the material. Thats starnge

Right, but someone mentioned earlier that semiconductors were partially conductive. I was just pointing out that that statement isn't entirely accurate

 

[legend killer]

Right, but someone mentioned earlier that semiconductors were partially conductive. I was just pointing out that that statement isn't entirely accurate

In my opinion the capacitance will decrease whether it is intrinsic or extrinsic, there will definitely be some charge leakage. What do u say?

 

[DerStrom8]

In my opinion the capacitance will decrease whether it is intrinsic or extrinsic, there will definitely be some charge leakage. What do u say?

I say it's highly dependent

 

[steveB]

Ohh yeaa, that didnt came to my mind. However the equation of capacitance C = εA/d doesnt involve conductivity of the material. Thats starnge

For AC signals, you can represent the permittivity as a complex number, as follows, where σ is the conductivity and ω is the angular frequency, with j=sqrt(-1).

ε'=ε-jσ/ω

In general, the imaginary part of the permittivity represents losses of one type or another. Capacitors are rarely well described by a simple constant permittivity (except perhaps vacuum and air dielectric), and have frequency dependent real and imaginary parts to the permittivity. With enough effort you could analyze a particular semiconductor and develop a low frequency model that describes the behavior. My guess is you could find something already worked out in the literature too. Often you see such treatments when dealing with biological materials, which rarely are well approximated as pure insulators or pure conductors.

 

[crutschow]

Leakage doesn't change the capacitance, it just makes a leaky capacitor (resistance in parallel with the capacitance) which won't hold the charge as long.

 

[steveB]

While I would not use the wording "leakage changes capacitance", I would also not say "leakage doesn't change capacitance". The situation is a bit complicated and it's not necessary to assign cause and effect. Real capacitors made with material dielectrics will have a very complicated frequency dependence to the capacitance. As I mentioned, the conductivity and other losses show up as the imaginary portion of the permittivity (if you do a sinusoidal analysis). The end result is a very complicated real and imaginary part, and these parts are related to each other through the Kramers-Kronig relations (a type of Fourier analysis). Hence, if you know the imaginary part of the permittivity (the losses), you can (in principle) know the real part. Also, the imaginary part of the capacitance changes the phase relation of the capacitor, and modifies the ideal 90 degree shift you expect between current and voltage.

https://en.wikipedia.org/wiki/Kramers–Kronig_relations

**broken link removed**

https://en.wikipedia.org/wiki/Permittivity

 

[Ratchit]

Hello everyone,
I would like to know what would be the effect on capacitance if we replace insulator with semiconductor. Just curious to know . What you guyz say?

A semiconductor slab between the plates of a capacitor instead of a good dielectric insulator? You would have a leaky capacitor. The capacitance would depend on the permittivity and physical dimensions of the capacitor, like always. Leakage does not change the capacitance value, but it will affect the ability of the capacitor to store energy and also degrade most circuits. The conductivity of a semiconductor depends on the semiconductor material, temperature, and the doping concentration. A varactor is not relevant because we are not discussing junction diodes.

Ratch

 

[DerStrom8]

A varactor is not relevant because we are not discussing junction diodes.

That's why I asked if the semiconductor was doped. If it was doped in such a way as to create a P-N junction, then a varactor would most definitely be relevant.

 

[Ratchit]

That's why I asked if the semiconductor was doped. If it was doped in such a way as to create a P-N junction, then a varactor would most definitely be relevant.

I believe the OP was referring to a semiconductor being doped for a homogeneous extrinsic result to be used as a dielectric slab, not a semiconductor being manufactured into a PN junction.

Ratch

 

[DerStrom8]

I believe the OP was referring to a semiconductor being doped for a homogeneous extrinsic result, not a semiconductors being manufactured into a PN junction.

Ratch

That was my original assumption as well, but it's always good to cover all the bases

I'm not disagreeing with you, just expanding.