Technique for direct measurement of thermal conductivity of elastomers and a detailed uncertainty analysis

Abstract

High thermal conductivity thermal interface materials (TIMs) are needed to extend the life and performance of electronic circuits. A stepped bar apparatus system has been shown to work well for thermal resistance measurements with rigid materials, but most TIMs are elastic. This work studies the uncertainty of using a stepped bar apparatus to measure the thermal resistance and a tensile/compression testing machine to estimate the compressed thickness of polydimethylsiloxane for a measurement on the thermal conductivity, keff. An a priori, zeroth order analysis is used to estimate the random uncertainty from the instrumentation; a first order analysis is used to estimate the statistical variation in samples; and an a posteriori, Nth order analysis is used to provide an overall uncertainty on keff for this measurement method. Bias uncertainty in the thermocouples is found to be the largest single source of uncertainty. The a posteriori uncertainty of the proposed method is 6.5% relative uncertainty (68% confidence), but could be reduced through calibration and correlated biases in the temperature measurements.