Breakdown of Air due to RF fields

[James50]

I'm just trying to think of a qualititive solution to the following thought experiment:

I have an RF source of very large voltage driving a signal towards point A (say, a capacitor or a dipole antenna). The source itself is not really part of this experiment, but I just want a sinusoidal voltage across point A. Ignore any effects in the wiring/waveguides due to high voltage.

Now, assume the electric field across point A is greater than the electric breakdown of air (which is around 3 MV/m), but below the breakdown of a vacuum. Say around 1 GV/m.

Now, presume in the far field (so that R >> 2*wavelength), a pocket of air exists. Ignore any boundary effects from vacuum to air.

At both point A and point B, we expect just normal, far field radiation with E proporational to 1/distance. At point C, assume the electric field is greater than the breakdown of air, ie greater than 3 MV/m.

What happens at point C? As I see it, there are a few options (I use spark and breakdown interchangably):

1) All locations in point C have an oscillating, continuous spark, depending on the instantaneous electric field it is experiencing... First it sparks upwards, then downwards, then upwards, etc, throughout the volume of the air. This to me seems unlikely - I just can't picture it.

2) Only the very first "bit" of air that the RF hits experiences an oscillating, continuous spark, depending on the instantaneous electric field at that location. This spark acts in anti-phase to the incoming RF and therefore cancels the large RF field for the rest of the air pocket. I can't see this either - I have no reason to believe it would act in anti-phase, or even that the wavelength of radiation that the spark would emit would equal the wavelength of the incoming RF.

3) All locations in point C have, at differing times, a quasi-random sparking, that is almost uncorrelated with the electric field. Each individual spark has an interfering effect on other sparks, leading to a non-predicatable pattern of breakdowns.

4) Only the very first "bit" of air that the RF hits experiences an quasi-random sparking, which interferes with the electric field in the rest of the air (not allowing sparking)

5) Nothing happens in the air - the oscillating nature of the RF field does allow enough time for a plasma to be created and therefore there is no sparking.

6) All of the above? Something else? Is option 5) highly likely for a high frequency RF signal, but not likely for a low frequency RF signal?I suspect for a high enough frequency (whatever that may be) signal, nothing will happen, as the plasma will not be able to form before the electric field flips. For a lower frequency signal (again, whatever that may be) I really am not sure. Lightning, for example, is caused by a steady build up of static voltage, which is discharged in a single strike, but this doesn't dicharge the entire cloud/storm front, only the bit in a localised area (i.e. you get more than one strike in a storm). This would suggest, perhaps, option 3).

So, any ideas?

 

[Dr. Courtney]

I think you need to define the frequency within a specific narrow range. kHz is different from MHz is different from GHz.

 

[James50]

I avoided including frequencies, as I don't want to get bogged down in mathematics. However, let's say 3 cases:

10 Hz
100 KHz
10 GHz

 

[tech99]

I'm just trying to think of a qualititive solution to the following thought experiment:

View attachment 85922

I have an RF source of very large voltage driving a signal towards point A (say, a capacitor or a dipole antenna). The source itself is not really part of this experiment, but I just want a sinusoidal voltage across point A. Ignore any effects in the wiring/waveguides due to high voltage.

Now, assume the electric field across point A is greater than the electric breakdown of air (which is around 3 MV/m), but below the breakdown of a vacuum. Say around 1 GV/m.

Now, presume in the far field (so that R >> 2*wavelength), a pocket of air exists. Ignore any boundary effects from vacuum to air.

At both point A and point B, we expect just normal, far field radiation with E proporational to 1/distance. At point C, assume the electric field is greater than the breakdown of air, ie greater than 3 MV/m.

What happens at point C? As I see it, there are a few options (I use spark and breakdown interchangably):

1) All locations in point C have an oscillating, continuous spark, depending on the instantaneous electric field it is experiencing... First it sparks upwards, then downwards, then upwards, etc, throughout the volume of the air. This to me seems unlikely - I just can't picture it.

2) Only the very first "bit" of air that the RF hits experiences an oscillating, continuous spark, depending on the instantaneous electric field at that location. This spark acts in anti-phase to the incoming RF and therefore cancels the large RF field for the rest of the air pocket. I can't see this either - I have no reason to believe it would act in anti-phase, or even that the wavelength of radiation that the spark would emit would equal the wavelength of the incoming RF.

3) All locations in point C have, at differing times, a quasi-random sparking, that is almost uncorrelated with the electric field. Each individual spark has an interfering effect on other sparks, leading to a non-predicatable pattern of breakdowns.

4) Only the very first "bit" of air that the RF hits experiences an quasi-random sparking, which interferes with the electric field in the rest of the air (not allowing sparking)

5) Nothing happens in the air - the oscillating nature of the RF field does allow enough time for a plasma to be created and therefore there is no sparking.

6) All of the above? Something else? Is option 5) highly likely for a high frequency RF signal, but not likely for a low frequency RF signal?I suspect for a high enough frequency (whatever that may be) signal, nothing will happen, as the plasma will not be able to form before the electric field flips. For a lower frequency signal (again, whatever that may be) I really am not sure. Lightning, for example, is caused by a steady build up of static voltage, which is discharged in a single strike, but this doesn't dicharge the entire cloud/storm front, only the bit in a localised area (i.e. you get more than one strike in a storm). This would suggest, perhaps, option 3).

So, any ideas?

I think it is very much like passing a high frequency spark through air between two metal balls. Once the air breaks down, which takes picoseconds, the path remains ionised and conducting continuously. The resistance of a small spark under these conditions is very low, maybe less than 1 ohm.

 

[James50]

tech99, as I see you are only describing the situation at point A (if air were present), where there are clearly defined electrodes acting almost as a point source of free charge (once the path is ionsed). Your description, I think, only applies for a time varying E field, not a traveling wave.

At point C, in the air, where there are no such clear sources of charge or driving currents, I don't understand how your description would apply. Do the "top" and "bottom" of the airpocket act as the metal balls, with a single conductive spark between them? How does this apply when at half wavelength intervals throughout the air pocket the electric field is is pointing in the opposite directions, and is constantly changing amplitude?

 

[tech99]

tech99, as I see you are only describing the situation at point A (if air were present), where there are clearly defined electrodes acting almost as a point source of free charge (once the path is ionsed). Your description, I think, only applies for a time varying E field, not a traveling wave.

At point C, in the air, where there are no such clear sources of charge or driving currents, I don't understand how your description would apply. Do the "top" and "bottom" of the airpocket act as the metal balls, with a single conductive spark between them? How does this apply when at half wavelength intervals throughout the air pocket the electric field is is pointing in the opposite directions, and is constantly changing amplitude?

My feeling is that the air will ionise when the E-field of the first wave exceeds the breakdown gradient, and will remain in this state, being relatively slow to re-combine. It will then act as a conductor, and current will flow up and down as each wave arrives. As it is a conductor, I would expect it to reflect most of the energy. It is perhaps similar to the T-R tube used in waveguides for radar.