[curiouscat]

Are there any good heuristics to predict the dielectric strength of gases? In particular SF6 is a well known gaseous insulator. A replacement is proposed by way of hexafluoroacetone.

I was wondering how would hexafluoroethane or hexafluoropropane perform? Perhaps lab data is available on the simpler molecules but for the rest any way to obtain a prediction.
 

[Enthalpy]

Alas, I don't have my books here... Some qualitative elements:

• Users want the insulating gas to stay gaseous in cold weather, and even under pressure, since pressure*distance determines the insulating voltage. So CF₄ is a well-known insulator, longer C_xF_y are not common.
• Do C_xF_y decompose to graphite and CF₄ in a spark? SF₆ and CF₄ recompose spontaneously - perfluoroacetone maybe too.
• Electronegative atoms are desired: fluorine, oxygen. They quench the already formed spark, allegedly as electron getters. For instance dry nitrogen is used for its high ionization potential but usually not alone, since little CF₄ improves it much.
• But fluorinated molecules are ozone killers, so designers want to get rid of them, not only SF₆. Hexafluoroacetone is a temporary solution. This is a stone in the high-voltage garden.
• The future is probably to generalize vacuum insulation. It's more efficient than gases, contains no halogens, and is already used.

Call to researchers: Mankind has too little knowledge about the electric breakdown of vacuum and needs more. We don't even know the breakdown mechanisms, and models are inaccurate.

[curiouscat]

Hexafluoroacetone is a temporary solution. This is a stone in the high-voltage garden.[/li][/list]

Thanks @Enthalpy!

Can you expound on what you mean by that last bit? Thanks!

[Enthalpy]

Hexafluoroacetone is toxic. Nothing tragic for chemists, but for high-voltage workers it's an issue.
It's hygroscopic, and if hydrated it shouldn't stay a good insulator, at least intuitively - so the procedures must be complicated.
It liquefies at -27°C so it can't be used under pressure in most climates. This limits its insulation capabiity.

Perfluoroalkanes would avoid the humidity absorption and toxicity, but I suppose they don't rebuild spontaneously after a spark. SF₆ and N₂ do, maybe perfluoroacetone too because it contains enough oxidizing elements.

One worry common to all is the greenhouse effect - not necessarily the ozone depletion, for which perfluoroalkanes are considered acceptable, my mistake. This webpage lists the Global Warming Potential (in CO₂ equivalent) of several gases
http://unfccc.int/ghg_data/items/3825.php
and perfluoroalkanes are only 2-3 times less bad than SF₆ which is a fantastic insulator.

As a gut feeling, we'll replace SF₆ by vacuum everywhere possible, and elsewhere by N₂ which isn't fantastic but accepts a high pressure.