: In this report, the procedures for analysis and prediction of thermal hazard of chemical substances resulting from thermal decomposition or runaway reaction are discussed using techniques of differential scanning calorimetry (DSC) and adiabatic calorimetry (ARC).
    Recent years have seen rapid development in the application of thermal analysis such as DSC or DTA (Differential Thermal Analysis) as a tool for the thermal hazard evaluation of chemical substances. This is because thermal analyses require a small amount of sample and no complicated operation. Thermal analyses, however, have inherent issues in principle and operation, so an adequate knowledge is necessary in interpreting the thermal analysis data.
    The data available from thermal analyses are more or less influenced by shape of sample cell, sample weight and environmental pressure as well as heating rate. The measurements such as decomposition temperature or decomposition heat are particularly greatly influenced by these factors, and hence discussion on thermal hazard by DSC data should be carried out on the data which were obtained at the same analyzer and under the same operation conditions. However, even the DSC data obtained by the same procedure vary a little in each measurement such as decomposition heat, with an error of 5 to 10%.
    If these inherent characteristics of thermal analysis are taken into consideration, the analysis and prediction of thermal decomposition or reaction hazard by DSC become feasible to some extent, and DSC is regarded as an effective method as a screening test for the primary hazard evaluation of chemical substances. In particular, a special attention should be paid from the standpoint of reaction kinetics to see whether the reaction proceeds with a single reaction mechanism or not in DSC experiments. While the single reaction mechanism is able to estimate from a regular deviation of DSC curves with heating rate, identification of decomposition products during the heating may be essential.
    An accelerating rate calorimeter (ARC) was used for adiabatic measurements for the evaluation of thermal hazard of chemical substances. ARC has an unique function of keeping a few grams sample in an adiabatic condition and measuring self-heat rate and pressure precisely.
 However, even ARC has a difficulty in catching up the sample temperature when temperature rise rate exceeds about 10°C/min, giving non-adiabatic condition in the ARC system.
    Moreover, the ARC system is not capable of agitating in the experiment, hence temperature gradient in the sample will be estimated, especially in the case of reaction of solid or mixed liquids. Thermal correction due to heat capacity of sample bomb is also required as sample quantity is relatively small as compared with weight of sample bomb. Therefore, the ARC data should be dealt with care when these data are applied to large scale chemical processes.
    ARC, however, is clearly distinct from thermal analysis such as DSC in the points of measuring self-heat rate., adiabatic temperature rise and pressure rise, and analysing reaction kinetics such as time to maximum reaction rate. Hence, the ARC data will be of importance in predicting such danger of chemical processes as runaway reaction or thermal decomposition.
    From the above results, thermal analytical methods are effective for a screening test in evaluating thermal hazard, and an adiabatic method such as ARC may be recommended as a secondary test if extensive, thermal hazard of the chemical substances is presumed from the screening test.