The arctic silver will make a difference, but maybe you should use the cheap stuff down at Radio Smack. $.50 isn't terribly inconvenient If you find that it's not getting the results you wanted, arctic silver ($5) or arctic silver II($9) can be ordered off the net from a bunch of different places.
BTW, it's a little square of pink thermal pad that comes on the heatsink. You just peel-and-paste it to the cpu.
They do not work together the cheap stuf is probably about the same as the pink pad as it will get the job done but if you want to overclock or are unsatisfied with the temps then get the artic sivler.
Nope, the pink pad counts as one interface, and the gel another interface. Scrape off the pink pad, or don't use any gel.
Any interface will be the limiting factor of your cooling potential. That is, if you have to 'join' two pieces together (preferably, with a gel of some sort), the conductivity of the gel is going to be your major problem. Aluminium conducts heat around 30 times better than arctic silver. About the only time it pays to have more than one interface is if you are cooling with a REALLY large differential, like water/freon/peltier.
that WOULD be true if the surfaces were perfectly flat, but the artic silver bridges the tiny air pockets (air is not a very good heat conductor) with very conductive silver. Thats why it pays to get the artic silver, it will usually help big time with the average ive seen around 4deg celcius.
It's still not solid silver. Pure solid aluminium gets around 237W/mK (that's watts per meter-Kelvin), not the oxide cream. I'm not sure if heatsink manufacturers put out the specifications for the type of metal alloy they use. Hopefully, it's not an alloy.
The thermal conductivity of AS is like 9W/mk, whereas aluminium oxide transfers around 2W/mk. I have no idea what the rate is on that un-wholey silicone. There's a real good chance that a couple of drops of urea might conduct better.