What do the different colors of plasma in spark gaps mean?

[js2020]

TL;DR Summary

I am interested in understanding what the different colors of plasma mean

Hi,

I am looking at videos of spark gaps and see that there are different colors of plasma. I really should've been a physicist because this may be fundamental, but I'm wondering what the different colors mean? I have seen purple and white. My guess is that white is a higher energy discharge compared to the purple. I would really like to discuss this with someone because unfortunately I don't know anyone else interested in this.

I will continue to do my own research but I would love some wisdom from those who know.

 

[Borek]

Google for "color of electrical sparks", that will give you a good starting point.

 

[DaveE]

I guess you mean a spark through air?

It really is more about physics or chemistry than electronics. Most visible light is generated when the electrons in an atom are in an "excited state" which means they have absorbed and temporarily stored extra energy. These electrons will then give off some of that energy as light when they relax to a less excited configuration. It is the electric current (and voltage) that provides this energy to the atoms.

The really cool part about this is that the colors (frequency) of the light is constrained to very specific values for the different elements and the different electron energy levels. So the plasma light that you see from Neon looks red/orange while the light from Argon looks blue/green. These are the typical colors you see in "neon signs", which is also a plasma discharge. These colors are a powerful tool for identifying atoms and molecules whether they are in a test tube or the atmosphere of another planet. This measurement technique is called "spectroscopy" if you want to learn more about it.

The problem with an air discharge is that there is a mixture of gases (Nitrogen, Oxygen, Argon mostly) all of these will give different colors and their electrons can be excited to different energy levels, which will also change their color. So, I think the white that you are seeing is likely a mixture of several different colored photons. Sorry, it's not really an answer to your question, it's kind of a mess with an arc through air.

This subject can become quite complex if you really delve into the details. This wasn't really understood until physicists developed quantum mechanics in the early 1900's.

 

[js2020]

Hi DaveE,

Yes I think you're right, I should probably move this out of the electronics forum and over to physics as it may be a better fit there.

To explain a little more though, I have my electronics that I am applying a high voltage to. When I start to see corona, it is a bluish purple color. I know this is the ionization of the air as your described. When I increase the voltage past a certain threshold, I see flash over between two parts. The flash over is a lot more intense and is white instead of bluish purple. My assumption was that the higher voltage creates a lot higher energy when the flash over happens which leads to the higher intensity and white light...or combination of multiple colors.

What I am really hoping to understand is things like what the color means as far as the type of plasma if that's a thing. For instance, I know that I can detect photon emission using a silicon photo multiplier (SiPM) diode. It's only effective over a certain range of wavelengths so determining what wavelengths are generated from what would be helpful in determining if that method of detection. Is adequate. Additionally, I can capture the corona and flash over using a camera. But again, I should know what the camera can capture to determine if its capturing everything of interest if if I'm missing things I need to see.

 

[Borek]

Try to break down the light from discharge using either a diffraction grating (even old CD can work) or a prism. With some luck it will tell you whether the color you observe has a continuous or a discrete spectra (most likely a bit of both). These discrete signals can be assigned to specific electron transitions of gases present in the air.

There are specialized spectroscopes that can be used but they are expensive, so some DIY approach can be better (at least for some qualitative observations).

 

[js2020]

Try to break down the light from discharge using either a diffraction grating (even old CD can work) or a prism. With some luck it will tell you whether the color you observe has a continuous or a discrete spectra (most likely a bit of both). These discrete signals can be assigned to specific electron transitions of gases present in the air.

There are specialized spectroscopes that can be used but they are expensive, so some DIY approach can be better (at least for some qualitative observations).

Thanks for the suggestion on the prism. I'll try that approach.

 

[JimmyinTEXAS]

@js2020:
I worked in the semiconductor industry for many years and have seen different gas plasmas in vacuum chambers. You didn't mention the voltage differential you are working with but it does sound substantial. A tesla coil? As DaveE mentioned earlier the gas mixture has a lot to do with what you see. If you could build a vacuum chamber you could introduce nitrogen or argon at lower pressures and have better visuals.

 

[Ntip]

Hi JimmyinTexas,

js2020:
You didn't mention the voltage differential you are working with but it does sound substantial. A tesla coil?

The differential voltage is around 25-35 kV. This is actually for a medium voltage SiC based power converter so there are many components in the assembly (power supplies, controller, gate driver, sensors, etc). I have corona on several components that I'm trying to eliminate. I was hoping to learn a little more about the discharges based on the color if possible, although it's not necessary for just eliminating it.

Unfortunately I don't have vacuum chamber that would support the size and voltage of my converter or it's individual components.

 

[JimmyinTEXAS]

Hi JimmyinTexas,
The differential voltage is around 25-35 kV. This is actually for a medium voltage SiC based power converter so there are many components in the assembly (power supplies, controller, gate driver, sensors, etc). I have corona on several components that I'm trying to eliminate. I was hoping to learn a little more about the discharges based on the color if possible, although it's not necessary for just eliminating it.

Unfortunately I don't have vacuum chamber that would support the size and voltage of my converter or it's individual components.

Good luck, keep us posted.