Interpretation and Sensitivity of Thermal Stability Measurements

The correct assessment of thermal instability is of paramount importance in the prevention of incidents in the chemical industry. There are many ways of assessing the thermal instability ranging from a simple bond energy calculation to carrying out the thermal runaway reaction in an Accelerating Rate Calorimeter (ARC™), The various ways of assessing the thermal instability have been assessed by some of these methods on a selection of five compounds having a diverse spread of reactivity. Several inadequacies were observed in the theoretical methods, making the usage of these methods a little limited. Preliminary screening by Differential Scanning Calorimetry (DSC) has been shown to be more reliable providing that the crucibles used for DSC are inert and capable of retaining all materials over the entire temperature scan. Usage of high pressure gold plated crucibles has been shown to be particularly successful in the gathering of accurate data for thermal stability evaluation, although it has also been shown that a slower scan rate is necessary with these crucibles, due to their larger size and weight. Comparisons of the results have shown that there are the expected differences between the DSC and the ARC™ data. Historically, all simple models for processing ARC™ data have been based on the original Townsend and Tou decomposition model. This model has an obvious limitation that has not been previously addressed, i.e. the time to maximum rate corrections for the heat lost to the bomb (thermal inertia factor) do not reflect the elevated temperature that is expected by the correction to the overall temperature rise. This study shows how the thermal inertia factor relates to the observed maximum rate, and shows how it may then be modified to generate a new factor which is used to construct an extrapolated heat-rate plot. By integrating this new rate plot, a new time to maximum rate curve may be produced which will run from the observed onset temperature to the thermal inertia factor corrected maximum temperature. A new simplistic model (Arc_Off) has been developed taking into account not only these new methods, but also includes and improves some of the more useful features of the ArcLink software previously developed by Mores. The new model runs on a Windows™ platform, to provide instant visualisation of all the many graphical possibilities simultaneously. In order to provide a complete system for ARC™ users, a sister program (Arc_On) has also been written which will not only aid the setting up of the ARC™ processor, but will ensure that the resultant saved ARC™ data is in the correct format for processing by Arc_Off. To reduce computing times a compiled program rather than an interpreted one was required, and therefore both these programs were written in Borland®’s Turbo Pascal for Windows™. The use of more sophisticated nodal modelling software (ARCSIM by Drayton) has also been evaluated using data for one of the five compounds. The model needs well-defined, and accurate kinetics in order to run successfully. The use of this model was shown to be successful once the program had been modified for the postulated reaction, and the large numbers of physical data necessary for correct operation had been found, or determined experimentally.