Concerns over the characterization of the thermal properties of coal ash using thermomechanical analysis (TMA) in carbon assemblies (crucible and penetrating ram) has led to a study which critically examines the effect of sample assembly materials on thermomechanical analysis measurements. This study quantifies the influence of the sample assembly material on the TMA traces. For temperatures up to 1600 °C and in inert conditions, ash chemistries were identified where the slag reacts with a range of sample assembly materials [molybdenum (Mo), platinum (Pt), carbon (C), zirconia (ZrO2), alumina (Al2O3), and boron nitride (BN)]. Only ZrO2 and Pt sample assemblies were found to be unreactive to ash samples during the tests. Iron-containing ashes were found to react with C assemblies, so that previous TMA data published for such assemblies and inert gases is not characteristic of ash fusibility alone. The implications of the reactions to the interpretation of this TMA data are outlined. For ashes that contain iron, TMA results in graphite sample assemblies may be considered to have been obtained under stronger reducing conditions than those used in the standard AFT test.