When people start talking about HS compound, I like to bring up the guru I followed for a while on the OCing newsgroup. When he pronounced, everyone listened. He claimed that as long as the two pieces were very flat -the HS bottom and CPU slug, that is- it made no difference what compound you used. He was not interested in polishing his HSs; he honed them in order to make them absolutely flat. At that time AS original was the vogue. But since it was expensive, he refused to use it as a matter of principle. It didn't make any difference, so why should he waste good money. He used margarine. Yes, the butter substitute that is entirely grease and no filler.
When the two surfaces are truely flat, the only function of the HS compound is to fill the roughness in the surface. When the compound is that thin, the temperature drop aross it is also tiny; so tiny that anything solid or liquid gives about the same result.
In point of fact, few people know whether their HS is truly flat, how to make it flat, or even how to determine how flat it is.
CPU slugs are probably flatter than they were a couple of years ago, but at least the Celerons I fooled with were not very flat. Under those circumstances, what is in the HS compound could make a difference. In other words, the reason AS makes a differerence, when it does make a difference, is that a comparatively large amount is needed to fill the gaps.
I believe the silver containing compounds got their start where the surfaces to mated were large in area, and therefore impractical to make extremely flat. (CPU slugs are rather small.) The AS inventor adapted those compounds to CPU use. AS is supposed to be non-conductive electrically. It is probably not a great insulator like plastic, but for practical purposes you should be able to cover the whole CPU with it without causing any electrical problems. I have gotten some on the bridges at times and never worried about it. Quite a few people have tried to use AS to close Athlon bridges, and it has not worked.
I guess it must seem wierd to subject elrctrical things to water, but check what goes on:
When the ICs are soldered to a circuit board, it is coated with a slightly corrosive flux first to make the solder join efficiently. Afterwards the flux is washed off with hot water and detergent in something reminiscent of a car wash, and rinsed. The circuit boards are then kind of cooked to drive off all the water and detergent that inevitably remains under the chips. If you touch a circuit board just after this, you will burn yourself. When I worked repairing circuit boards in a factory, we sometimes got a box of completed boards that were done improperly and gave loads of errors in the automated testers. You cold blow liquid out from underneath the chips with compressed air. ICs need to be impervious to water.
Chips also need to be sealed to protect them from corrosion due to gases in the air (like oxygen), and humidity which accelerates corrosion.
A CPU should be waterproof as it comes from the factory even with the bridges burned open. Burning the bridges open with a laser should seal the plastic about as good as anything. If you cut the bridges, I think the surrounding plastic still seals off anything electrical. The only thing that gets exposed is a grounded metal grid just beneath the plastic and the ends of the wires you have cut.
I really think regular users stick with AS mainly because they are sure of what they are getting. The company appears to have proven itself. I don't personally know about ASIII, because I haven't gone through enough CPU and HS changes to use up my tube of ASII. The tube is tiny, but you only need a drop. Any more and you can't keep from slopping it all over. There is nothing miraculous or super about AS as best as I can determine. I used a very, very flat HS. When I experimented with different compounds and AS I got no difference in temps whatever, that I could attribute to the compound. I did sometimes get a heat rise on some trials, but when I retried, it would go back to normal. Apparently sometimes applications go a little wrong. A phase change pad gave results very close, provided you make sure it gets hot and stays hot a few days. The stuff is thick and oozes out very gradually. After that, it looks very thin, not much different than AS, and it is hard to say the temperature difference is over 1 degree.