Arctic Silver 5 no longer king? - Page 2

You're right:
http://www.tibtech.com/conductivity.php

I could have sworn it out-performed gold, copper and silver.... I know that we used to use aluminum in our electrical wires, but had to stop because it was a fire hazard. If you over-volted an aluminum electrical wire, the entire wire, from end-to-end, would melt. It was that good of a conductor, with that low of a melting point. That's one of the reasons Aluminum is one of the most difficult metal to weld. You need to get a metal to it's melting point (or extremely close to) in order for a weld to take properly, the problem with aluminum is, if you get one spot to it's melting point, you get a huge area around it to the melting point as well - and once aluminum hits it's melting point, it's a liquid (there is no "malleable" stage for aluminum when it comes to increases in temp); it's either a solid, or a liquid, there is no transitional phase to speak of, and it conducts so well that melting one spot usually melts the rest.
This is also the reason we use copper in our lines, and not aluminum. The prices of Copper and Aluminum usually follow pretty close to one another. Aluminum Ore is more common than Copper Ore, but Aluminum Ore takes a much larger amount of energy to be turned into a useful metal than Copper Ore does. At the same time, Aluminum is infinitely recyclable because it's so reactive (the top lay of aluminum you see is nothing but tarnish. You will never see raw, untarnished, solid aluminum metal - it reacts with the surrounding air instantly. But that same tarnish keeps it from corroding further.), all you need to do is melt it to reset any of the tarnish and corrosion. This is not true with copper. We will always need to find more Copper. Aluminum will eventually fall to some of the lowest metal prices we see. Eventually.

As for AS5 (I cannot speak for ICD), what you said is true, and not, at the same time. I do not know the exact ratios of AS5, but their claim of 'using 3 different sizes and shapes of silver nanoparticles' is true. I've taken a look at a sample of AS5 under a TEM (we were bored waiting for another experiment to finish), and there are 3 distinct shapes, each with a different size, that all link up together very well. Now, we had to use a very thin sample, much thinner than would be useful on a CPU, so these arrangements may become less efficient at greater thicknesses. AS5 is silicone based as well, if memory serves, and silicone is an insulator. If I had to guess, at thinner layers, the particles are forced together (by the spreading out of the application), forcing out the silicone between them, establishing the greatest conductivity. At greater thicknesses, less of the silicone is forced out, lower the conductivity - but at the same time, lowering the capacitance. There is probably a sweet spot of thickness - thick enough to fill all the gaps between CPU and heat sink, thin enough to still be highly conductive. I'm just not sure that sweet spot is realistically achievable by human hands.


Nope.
We use gold over silver not just because it's easier to plate with than other metals (it's really no easier, or difficult, than other metals), or because it's malleable. We use it because it's stable and does not tarnish - while still being extremely conductive. Silver is difficult to tarnish, but it still happens over time as it's exposed to air. Gold on the other hand, does not. So, this complete lack of tarnish ever forming allows for the least amount of noise introduced at the connector.

Sure. I'm not familiar with the electrical conductance/capacitance of PK-1 (namely if it is or isn't). I know 2 out of the 3 of the PKs are conductive/capacitive. So that's the only concern - but not a very big one.

The only TIMs you need to worry about are phase-change metal TIMs. They're the best, but they are also the easiest to mess up. Coollaboratory makes the only Phase-change metal TIMs that I know about - Liquid Pro, Liquid Ultra and various metal pads (that melt every time the processor heats up, and resolidify when it cools down).

PK-1 should do yo just fine.