Details of the development and initial evaluation of the slug calorimeter can be found in reference (10). The underlying principles are similar to those in the apparatus originally described by Fitch (11) that is still utilized for estimating the thermal conductivity of leather (12). As shown in Figure 2, the basic experimental setup for the slug calorimeter consists of a sandwich specimen with a central core composed of a 12.7 mm thick steel slug of known mass and heat capacity (AISI Type 304 stainless steel). The central slug contains three holes for the insertion of Type N thermocouples. Samples of the FRM to be evaluated are placed on both faces of the slug and the entire sandwich is guarded by a low thermal conductivity high temperature insulation (13, 14). The guarded sandwich is then placed between two metal retaining plates manufactured from a nickel-chromium alloy to maintain a slight compression on the specimen during testing (see Figure 2). The assembled specimen is placed in an electrically-heated box furnace and exposed to a series of heating and cooling cycles. Both the internal slug temperature and the temperatures of the outer FRM surfaces are monitored during the experiment. The heating curves are chosen to simulate an ASTM E119 (15) exposure, but with a less rapid initial temperature rise due to furnace power limitations. Multiple heating and cooling runs on a single sandwich specimen are used to investigate the contributions of chemical reactions, phase changes, and the pressure-driven mass transfer of gases produced during these reactions to the effective thermal conductivity of the FRM.
Knowing the temperature gradient across the FRM specimens and the heat flow into the slug calorimeter, a straightforward analysis can be applied to compute an effective thermal conductivity for the FRM as a function of the mean specimen temperature (10). One-dimensional heat transfer through the FRM to the slug is assumed, and the sandwich specimen geometry provides an adiabatic boundary (Q=0) at the middle plane of the steel slug that greatly simplifies the subsequent analysis. The testing has also been applied to "non-reactive" materials, such as the fumed-silica insulation boards typically used as the guard insulation material, to further validate the approach (10). While more information on the mathematical analysis and the expected uncertainties accompanying the measurements can be found in reference (10), the most significant contributor to the expanded uncertainty is the uncertainty in the temperature measurement. For example, assuming an uncertainty of 1 ºC for the thermocouple readings and a 5 min sampling interval, the estimated uncertainty in the effective thermal conductivity would be about 5 % for values computed in the temperature range of 400 ºC to 700 ºC during heating (10).
Figure 2 Schematic and photo of the slug calorimeter test setup: left- schematic of a cross section through the middle of the basic slug calorimeter setup, and right- photo of a completed sandwich specimen of the fumed-silica insulation board mounted and ready for testing in the box furnace.