Thermal oxidative decomposition estimation combining TGA and DSC as optimization targets for PMMA

Abstract

Thermal analysis techniques play a key role to determine and characterize solid phase thermal decomposition. In this sense, Simultaneous Thermal Analysis (STA, i.e. TGA and DSC tests carried out simultaneously) are widely employed, since it provides information about how mass is lost and energy released while the temperature of the sample increases. Fire computer models combined with methods numerical methods are widely used to represent the results from tests and to achieve the values of the kinetic and thermal parameters. Previous works looked forward achieving those parameters using, as unique optimization target, the mass loss curve (TGA) or its derivative (DTGA). As the study of heat release rate is a decisive element to characterize the material properly, most recent works were adding additional measures. These extra measurements concern the heat transfer and the energy required or released during temperature programmed heating, such as heat rate release, heat of gasification, or the surface temperatures of the samples. The information about the energy is provided by the Differential Scanning Calorimetry curve (DSC). Despite of the employment of the information provided by the DSC, this information usually is not used as a target to approach the DSC simulated curve to the experimental one as TGA does. Based on the lack of use of the DSC curve as numerical approaching process to set the kinetic properties, we decide to explore the possibility of adding this as a new target in the process. Therefore, kinetic and thermal properties might be achieved fitting experimental and simulated curves simultaneously, which should allow us to take into account the decomposition process and their energy released. Results obtained in the present work reveal the major challenge of getting a set of parameters, which can fit DSC curve. The level of accuracy reached when only TGA is utilized as target to approach is higher than the level of accuracy of DSC curve. This fact makes increase the value of the errors when both curves are used as targets to approach. In other words, an approach to both curves simultaneously cannot be directly made. With this consideration in mind, this paper proposes an alternative methodology in order to fit TGA curve considering the optimization of the DSC curve. The methodology proposed in the present work is applied to the analysis of poly(methyl methacrylate) (PMMA).