Microdischarges in solid dielectrics

[vanesch]

Hi all,

I'm looking for information/theories/data/phenomenology about micro discharges in dielectrics (and surfaces).

I'll explain the problem: we have some conductors under a potential of ~2500V which are (of course) in contact with solid dielectrics (types of plastic, ceramics...), and we observe low-frequency (~0.01 Hz) microdischarges on them of the order of ~0.05 pC - 0.2 pC which, well, screw our measurement setup.
Now, I know some basic techniques to diminish it, like using guard rings and stuff like that, but I'd like to know more about the phenomena themselves, and I can't find any material about it.

So any pointers are welcome.

 

[ZapperZ]

Hi all,

I'm looking for information/theories/data/phenomenology about micro discharges in dielectrics (and surfaces).

I'll explain the problem: we have some conductors under a potential of ~2500V which are (of course) in contact with solid dielectrics (types of plastic, ceramics...), and we observe low-frequency (~0.01 Hz) microdischarges on them of the order of ~0.05 pC - 0.2 pC which, well, screw our measurement setup.
Now, I know some basic techniques to diminish it, like using guard rings and stuff like that, but I'd like to know more about the phenomena themselves, and I can't find any material about it.

So any pointers are welcome.

Not sure about this, but you might have the same problem that we have.

We have a traveling wave dielectric tube with coupling to copper waveguides. We detect breakdowns at the joints where the dielectric meets the copper when we have roughly 3 MW of RF power going through. We see light glow when this happens and we attribute this to secondary electron emission causing some form of gas ionization (the whole system is under 10^-7 Torr). So this is in addition to the breakdown at the joints. In high gradient studies, the "triple point" region is usually (depending on geometry), the region most susceptible to problems. So in this case, one has the dielectric-metal-vacuum point. In our simulation using Microwave Studio, this part has one of the highest field and thus susceptible to such breakdown/sparks.

Zz.

 

[vanesch]

I don't think both phenomena are related: I'm in a quiet, all-static environment, and normally, the fields are way below ionizing any gas, and the dielectrics are stressed way below their breakdown fields. In other words, I shouldn't see anything, but I do: I see small discharge pulses of a few tens of femtocoulombs, one every few minutes.

 

[berkeman]

Interesting problem. 0.01Hz is pretty suspicious, though. Any chance it's a measurement setup issue? How are you measuring these discharges? Are there any other paths that the discharges might be taking? Are there any other things that seem to be possible triggers, or at least are correlated with the discharge events? Like ESD events nearby, or HVAC thermostat trips, or AC mains noise?

The only semi-related issue that I can think of for the dielectric breakdown would be some disclinations or other breaks/pinhole paths through the dielectric. How pure is the material? What is it?

 

[ZapperZ]

I don't think both phenomena are related: I'm in a quiet, all-static environment, and normally, the fields are way below ionizing any gas, and the dielectrics are stressed way below their breakdown fields. In other words, I shouldn't see anything, but I do: I see small discharge pulses of a few tens of femtocoulombs, one every few minutes.

But you don't need to get to the breakdown threshold to get secondary emission. All you need is a few dark current electrons that are above the first secondary electron energy crossover that can induce multipactor. We have observed such phenomenon without causing any permanent damage to our dielectric, and dielectric are notorious for this because their secondary emission coefficient are generally way higher than metals.

The onset of multipactoring is enough cause either ionization or some form of luminescence for light emission, even for a short burst. This could be what you are detecting. It would help if you can monitor the field in the tube and see if there's a sudden drop in such a field when this occurs.

Zz.