There is experimental data out there somewhere that plots thermal resistance across an interface against mounting pressure for a range of materials. Unfortunately the only place I've seen this was in a couple of thermodynamics lectures, and I can't seem to hit upon the right search terms or textbooks to find a reproducible referance.
Still, I was astonished to see that for low mounting pressures, nickel plated copper - nickel plated copper has a lower resistance than copper to copper. The explanation being that electroplated nickel under pressure deforms to fill some of the voids in the joint to a greater extent than solid copper does.
How thermal interface material / solder affects this, I do not know. I imagine it decreases the difference between the two. I've mentioned this on these boards once or twice before, but without any form of reference it's hard to take me seriously. If therefore you do find any tables on thermal resistance at an interface between metals, please do post them, as it has important implications for lapping.
(To within the normal margins of error of computer watercooling though, I'd agree entirely that it's cosmetic.)
