Thin and Thick Samples Tests Using Heat Flow Meter Instruments

Abstract

Heat Flow Meter instruments (ASTM C518, ISO 8301, EN 12667, JIS A 1412-2 [1-4], etc.) usually are calibrated using Standard or Certified Reference Materials like NIST 1450d SRM (26.0 mm thick), NIST 1453 (13.4 mm thick), and IRMM-440 CRM (~34.5 mm thick). In ideal world, all the calibrations would look the same, because calibration factor is physical property of the heat flow meter. In reality, calibrations are not exactly the same, and also thin samples’ calibrations usually are slightly higher because all the Heat Flow Meter Method Standards, mentioned above, assume that thermal surface (or contact) resistance is negligibly small in comparison to the sample’s thermal resistance (=thickness divided by thermal conductivity). In case of samples of small thermal resistance this can cause noticeable errors. Two-thickness procedure to exclude the surface thermal resistance to improve accuracy has been used successfully in the Heat Flow Meter instruments for number of years already [5]. Surface thermal resistance of regular FOX heat flow meter now can be measured (and then excluded) using tests of one, and then two layers of thin rubber sheets. Very thin samples like e.g. paper or films also can be tested as stacks of two different thicknesses, but possible thermal resistance between the thin samples cannot be separated and excluded—this was shown earlier [6]. Traditional method for testing samples/materials of smaller thermal resistance is ASTM E1530 [7] where the thermal contact resistance correction is used. Tests of very thick thermal insulation samples, Vacuum Insulation Panels and other superinsulation materials are much more difficult. Duration of such samples’ tests is enormous because time to reach their full thermal equilibrium is proportional to the square of their thickness. New method of calculations, based on the thermal problem analytical solution, can “predict†the resulting values - both thermal conductivity and thermal diffusivity can be calculated using new formulas. Recommendations how to accelerate such thick samples tests and to minimize the edge heat losses/gains and distortions are presented.