Crucial voltage
Yes, that crucial voltage needs to come down. I'm not sure I'd want 1000 volts floating around anywhere near my very expensive CPU!
Scientists at Indiana's Purdue University have worked out to significantly increase the efficiency of kit designed to keep toasty microprocessors cool by zapping the air above the chip with a thousand Volts. Coolers work by pulling air across a chip's upper surface to draws heat away from it. Unfortunately, interactions …
This concept of ionising the air sounded familar, so I dug around.
http://www.itnews.com.au/News/36424,boffins-build-tiny-onchip-cooling-system.aspx
or
http://www.cbronline.com/article_news.asp?guid=BD4566FF-B02E-4360-ACA8-3CA61EB90CFB
Yep... about a year ago, the University of Washington announced the same thing. And according to them, they didn't even need a fan, which is something these new boffins required for their test.
The efficiency of the cooling system could be described as the amount of heat it can clear out per unit of energy it takes to do so. I heard someone run some math (which looked as reasonable as not) suggesting the ion wind costs a few times as much power as the fan it makes more effective.
So overall they might not have changed power efficiency, just efficiency of space, weight, complexity (assuming, of course, that ion wind hoosits are less complex than other high-performance alternatives), and perhaps noise.
I say we turn instead to doping and etching the chips onto diamond and let 'em run smoking hot. ;)
They forget to mention the cost (i.e., power and effort) of removing the ozone generated as a byproduct of the ionization process. Unless of course the designers don't care about ozone in the locality of a computer using their newfangled technology.
Perhaps they could design some sort of liquid cooling system. It would be simpler, cheaper, already available off the shelf, doh!
This test was a prototype using 1000 volts across a 1cm gap.
If the requisite voltage follows an inverse square law, the hoped-for 1 micron gap would need only 10 volts.
My thanks to The Register for actually reading the research paper and providing the experimental conditions that were absent from the Purdue press release.