When effectiveness of a circulating coolant is in question, the primary properties (of a non-chilled design) that need to be selected for are specific heat and viscosity. Conduction, in a well-designed system, is almost entirely irrelevant to our purposes.
The specific heat of water is 4.186 j/gm K, higher than any metal commonly used in cooling systems. However due to specific heat's dependence on mass, density is just as important for specific heat per unit volume.
Some values of specific heat per unit volume, obtained by multiplying density values by specific heat values (in J/cc °K).
Water = 4.186 / Viscosity = 1.0
Fluorinert (most dense formula) = 0.483 / Viscosity = 12.6
Hydrogen peroxide, 100% = 0.885 / Viscosity = 1.22
Mercury = 1.904 / Viscosity = 1.55
For viscosity values, lower is better. Fluorinert ranges from 0.4 to 14.0 cS, so one of the less dense formulas would be a better choice if it was absolutely necessary.
Water is dramatically more effective than anything else here - even mercury, which is nearly 14 times as dense.
Ferrofluids, then. Considering the specific heat per volume of some metals:
Iron = 3.502
Aluminum = 2.43
Copper = 3.427
Silver = 2.465
and then weighting that value by the amount of the material actually suspended in the fluid. Generously assuming 50% suspension, and an oil with a specific heat/vol of 1.0, we get some rather pessimistic initial estimates for ferrofluid thermal capacity (between 1.7 and 2.3 or so), without even considering such a substance's viscosity's effect on cooling capability.
Would you really want to go through all that trouble only to get reduced performance? I can't imagine any of the materials would be cheap either...
|