Liquid Metal Research Experience (Score 1)

As part of some research work I was doing for the Spallation Neutron Source (http://www.sns.gov/) here in Oak Ridge, I ran across a lot of information about the nature of liquid metal cooling systems. I will try and list them for you all: 1) Magnetically driven pumps are inefficient, on the order of about 1% total pump efficiency 2) Liquid Metals (especially Mercury) have cavitation problems with use traditional centrifugal pumps. 3) Liquid metals are not as terribly corrosive as you all have suggested. Mercury and GIT (a eutectic alloy of Gallium, Indium and Tin) will operate for a very long time in Stainless steel. 5) A GIT alloy is probably the best bet for a design utilizing a liquid metal. If it were to be allowed to come in contact with the chip (I.E. reduced overall thermal resistance) it would form an oxide layer with excellent heat transfer capabilities. I would Love to see some people do some research into this problem as it seems to me to be very interesting. I am in the middle to using GIT to enhance the function of a sensor I am working on and already it has yielded heat transfer far superior to any type of ceramic or oxide mixture.