IBM's thermal paste cooling innovations detailed

It was but a few months back that IBM was tooting its own horn in regard to a "breakthrough" technology sure to revolutionize the processor cooling realm as we know it, and while miraculous claims often vanish after their bold declarations, it looks like this invention is moving forward. While we'd already heard the typical techie hub bub that goes along with such systems, IBM is now explaining a bit more about how the process will eventually work. Essentially, researchers have created a system in which "tree-like branched trenches" are placed in the copper cap, where a newly-thinned thermal paste can be applied with half the pressure of current renditions, netting a "twofold increase in cooling performance." The micrometer-sized channels basically act as an "irrigation system" to allow the toasty particles to homogeneously escape rather than building up in the self-proclaimed "magic cross" section. We know, this still isn't spelling things out in layman's terms, but if you're truly interested in knowing precisely how IBM plans on slashing the heat emitted by your future CPUs, grab your reading glasses and hit the read link.

[DWells55 @ Mar 23rd 2007 7:53AM]

Great news - not all of us can afford a fancy watercooling system and this is a very cool innovation given it works as well as they make it sound. Hopefully this means better-performing silent coolers too. Here's to hoping that we see this in the near future.

[yaksplat @ Mar 23rd 2007 8:12AM]

Heat particles......never learned about those in engineering school.

[ph43drus @ Mar 23rd 2007 8:49AM]

Where does TFA cite "heat particles"?

Thermal paste is a liquid medium filled with solid particles that have high thermal conductivity. IBM says that the particles become non-uniformly distributed during the process of pressing the heatsink down on top of the paste, due to the complicated flow of a filled paste in a constrained area.

The IBM invention just seems to be a way to pattern the bottom of the heat sink to facilitate flow, giving a more uniform distribution of the filler particles - and hence better thermal conductivity in areas that would normally be starved of particles.

Very much an empirical, engineering solution to a subtle rheology problem.

[regomodo @ Mar 27th 2007 9:02PM]

come on mate. This is engadget after all.

It seems someone a while back replaced their vitamin tablets with stupid pills

[yaksplat @ Mar 23rd 2007 8:56AM]

"The micrometer-sized channels basically act as an "irrigation system" to allow the toasty particles to homogeneously escape "

Toasty particles escaping?

[Paul @ Mar 23rd 2007 9:00AM]

I fail to see how this will cut down on the heat produced by a processor.

It seems to me that a bigger problem is getting the heat to dissipate efficiently from the heatsink, rather than a problem with the heat getting from the processor to the heatsink.

[ElvisLives @ Mar 23rd 2007 9:29AM]

It won't cut down on the heat produced by the chip. The article indicates that 40% of the thermal budget is consumed by the paste itself, i.e., 40% of the heat is prevented from reaching the heatsink. The heatsink and the thermal paste work together...

[ElvisLives @ Mar 23rd 2007 9:25AM]

Actually, the article was dumbed-down to a level below "layman's terms" . . . Try to notch it up a bit, OK Engadget? Like, engage writers who understand what they're reviewing...

[mark @ Mar 23rd 2007 12:29PM]

I agree with earlier posters, the linked article is simple and easy to understand.