On Correcting Thermophysical Property Data for Thermal Expansion Effects

Abstract

Most researchers do not correct thermophysical property data published in the literature for thermal expansion effects. Even values reported by the National Bureau of Standards for standard reference materials (SRM’s) are not corrected for dimensional changes. Users of the data and, in fact, many researchers assume that corrections have been made or that they are negligible compared to experimental errors. However, the corrections may be several times larger than they are assumed to be because customary thinking concerning dimensional correction is based on the units of the thermophysical properties. The present paper shows that there is not necessarily a one-to-one relation between the expansion correction and the power to which length enters into the property units. In the case of thermal conductivity “λ” for example, the units are often expressed as W cm^-1 K^-1 and one might think that λ_corrected=(1+Δ L/L₀)^-1 λ_uncorrected. Actually the correction term may be (1+Δ L/L₀)^-n where n=0, 1, 2 or 3 depending upon the experimental technique used — or the correction term may even be based on the expansion of a material other than the sample. Similarly the hemispherical total emittance (a dimensionless quantity) may require a (1+Δ L/L₀)^-2 correction term. Thus the units of the property involved do not necessarily provide a good guide to the correction term.

Basically one must correct each experimentally measured value for dimensional changes rather than merely using property units and correcting the final value. Also in the case of anisotropic materials, one must use the appropriate values of expansion for the directions in question. In this case λ is W cm cm^-2 K^-1 and not W cm^-1 K^-1.