Abstract
Various absolute steady-state methods of measuring thermal conductivity can be made to yield, verifiably, experimental results accurate to within ± 2% from 100 to 1200 K, and for conductivity values from 0.1 to 5 W/cmK. This statement is substantiated by a detailed analysis of the results obtained by eight different systems in three laboratories on ten metals: Cu, Ag, Au, Pd, Pt, W, Cr, Armco Iron, and the alloys Nb-10 W and Inconel 702. The experimental systems by which these results were obtained are also briefly described.
The achievement of such accuracy, however, requires that careful attention be paid to the design and operation of the experimental apparatus. In general, the thermal conductivity is computed from three parameters: the power developed in a specific heater; the temperature difference due to that heater; and a geometrical factor relating the two through the geometry of the heat flow pattern. The determination of each parameter presents problems of its own; these we outline, and present solutions to them which are often a matter of judicious compromises. However, because the individual problems are not equally severe for the different methods here discussed, the close agreement of the results that we show provides a gratifying reassurance that the experimental approaches and theoretical analysis, including estimates of error, are basically sound.