Effects of Interface Resistance on Measurements of Thermal Conductivity of Composites and Polymers

摘要

When testing samples of moderate thermal conductivity (λ～ 0.3-10 W/mK) using steady state methods the thermal contact (or interface) resistance R must be addressed, otherwise significant errors will result. It can be taken into account by instrumenting the sample - that is taking the temperature measurement of the sample surfaces with thermocouples placed directly on the sample. Another effective means of accounting for the thermal contact (or interface) resistance is the Two Thickness Analysis. By using at least two samples of the same material with different thicknesses Axj and Ax2 a system of two equations containing two unknown values can be solved: Q_1 = ΔT / [(Δx_1/λ +2R) S_(cal)] (1a) Q_2 = ΔT / [(Δx_2/λ +2R) S_(cal)] (1b) where ΔT is the plates' temperature difference, λ is thermal conductivity, 2R is sum of two thermal contact resistances, Q_1 and Q_2 are signals from the heat flow transducers, and S_(cal) is their calibration factor. We assume that the thermal contact resistances at the two surfaces and for both samples are equal. If the sample is instrumented and the temperature drop across it is measured directly only one equation is necessary. However it is frequently not practical and/or easy to do. In case of thin samples and/or of moderate thermal conductivity the thermal contact resistance 2R may even exceed the sample's thermal resistance Δx/λ For example 1/4" thick Pyroceram has thermal resistance Δx/λ=1.59*10~(-3). The thermal contact resistance 2R is about 3 * 10~(-3) - or almost 2 times bigger. This paper will present test data on polymers and composites used as reference materials exhibiting a range of thermal conductivities to illustrate the effects of not having accounted for the interface resistance.