Motion in insulating fluid under high voltage

I observed the following phenomenon in an experiment (I'm not a student of physics, just an amateur) and was hoping for an explanation.

A metal pan is electrically grounded and a layer of insulating fluid (vegetable oil) is placed in the pan. If you like, assume the pan is infinite in diameter. An electrode (a needle) is placed above the pan, pointing downward at the oil. A high voltage (-12kV) is applied to the electrode. The electrode is close enough to the pan to generate an ion wind, but not close enough to overcome the breakdown voltage of the air and oil, so there is no arc. The oil is repelled from the electrode (perhaps because it acquires a negative charge from the ion wind) and also, more remarkably, the surface of the oil begins to undulate.

My theory to explain the undulation is that the negatively charged oil on the surface is repelled from the electrode until it touches the pan, where it loses its charge and therefore is permitted to rise to the surface again, and the process repeats. So the undulations are something like convection cells, only with electric charge instead of heat.

Is this explanation correct? Is there a name for this phenomenon? Would a similar phenomenon create leakage current in capacitors with liquid dielectrics?

-

Your initial explanation is close to what happens in reality. Let us call things by their names; the discharge you initiate by biasing the needle is called a Corona discharge (if the pressure was much lower it becomes a glow discharge). The configuration you built is called a dielectric barrier discharge, for short a DBD.

In a corona discharge streamers form to connect the anode to the cathode. The streamer is basically a channel of partially ionized gas that is capable of carrying current between the electrodes. The streamer starts to from by local ionization of gas molecules in the vicinity of the needle. So a plasma cloud forms there, at the other edge of the plasma cloud (the edge away from the needle) a strong electric field is generated which accelerates the local electrons there giving them the necessary energy to ionize gas molecules. The energetic electrons create a plasma cloud there (which is connected to the plasma cloud next to the needle from one side), the process continues until the streamer connects the needle to oil surface. This phenomenon of generating streamer is called ionization wave propagation.

When the streamer edge arrives at oil surface, negatively charged particles (electrons and negative O2 ions) are deposited on the surface, creating a local negatively charged spot. At this particular spot the pressure of oil increases, which from Navier-Stokes equations cause the fluid to move in the direction of decreasing pressure.

This particular phenomenon has been studied experimentally in the late 90s. The motion of oil is called Electrohydrodynamic surface waves (EHD). Have a look at this paper where they study a very similar configuration to yours.

-
Thanks, I will take a look at that paper. I don't think that what you said about streamers describes my setup. There were no streamers between the point electrode and the oil surface, and there was no plasma cloud around the oil surface. The formation of ions was confined to a small stable region around the point electrode, as in pac.iupac.org/publications/pac/pdf/1985/pdf/5709x1353.pdf, p.2, but with the polarities reversed. Does this affect your answer? – Trevor Wilson Sep 25 '13 at 15:43
Plasma in the general sense is a mixture of ions and electrons, it doesn't necessarily need to be glowing at an intensity you can observe with your naked eyes. Researchers working in the atmospheric pressure range use special cameras to observe the glow of the plasma at that pressure. The glow is observed at the tip of the streamer moving at timescale of nanoseconds. I don't know what makes you think it is not the streamer scenario, can you tell? The difference in polarity is not a big deal. Instead of ions moving toward the oil electrons move. – Gotaquestion Sep 25 '13 at 16:57
To be precise, changing the polarit makes some differences in the mechanisms of ionization and transport of charged species. But in the general layout it is not different – Gotaquestion Sep 25 '13 at 17:12
Okay, perhaps all the air in the vicinity of the experiment satisfies the definition of "plasma" in some weak sense because it contains ions. However, I think ions are only being formed near the point electrode. The reason I say there were no streamers is that I know what streamers look like and I didn't see them. It appeared to be a unipolar corona. (Surely you acknowledge the existence of such a thing?) – Trevor Wilson Sep 25 '13 at 17:28
Also, I didn't say that the difference in polarity was a big deal, but actually negative and positive coronas do behave differently as explained in en.wikipedia.org/wiki/Corona_discharge#Negative_coronas and also in my earlier link. – Trevor Wilson Sep 25 '13 at 17:29