Is there a lower limit to the separation between capacitor plates?

1. Oct 19, 2011

theman2000

Let's say I build a parallel plate capacitor using two metal plates with perfectly smooth surfaces. They are separated by some small distance d, with only vacuum between them. I then apply some AC voltage across the plates. Is there some effect which causes loss in the capacitor (equivalent series resistance) that increases as the plates are moved closer and closer together (as d decreases)? More practically, we're talking about d on the order of ~10 microns or so. At some point will the loss become so high that the capacitor will become more of a resistor than a capacitor?

2. Oct 19, 2011

Low-Q

The loss is firstly in the metal itself, by its resistance. A capacitor does not have loss in terms or electric resistance between the plates. However, if you bring those plates close enough, very little voltage is required to let the electrons start to jump from one plate to the other. Then you will have an electron flow between the plates which can be seen as a loss pretty much as in a resistor as not all electric current is flowing through the generator, but also through the capacitor. If you heat the plates you will increase the electron flow - just like inside a radio tube, but also reduce the electric resistance.

3. Oct 19, 2011

nasu

There is energy lost in the dielectric, due to dielectric relaxation and/or residual conduction. It is measured by the angle of loss (or dissipation angle). It may be represented or modeled by a resistor in series with the ideal capacitor. The angle of loss is defined by
tan(delta)=R/X where R is this equivalent resistance and X is the reactance of the capacitor.
However, I expect that the factor that determines the minimum distance (for a given nominal voltage) is the dielectric strength or breakdown field.
For air is about 10^6 V/m so this will be about 1V/micron. However for this layers the strength may change. For sub-micron dielectric layers tunneling conduction may occur.