Characterisation of waste fuels with TGA

In this study an attempt is made to develop a method to determine the composition of an unknown waste mixture. The basic idea is that the waste fuels are composed of some major components, e.g. paper, plastic, textile and biomass. These single components and waste mixtures are characterised by thermogravimetric analyser (TGA). The measured weight loss curve (TG-curve) or its differential (DTG-curve) is used as a fingerprint of each material. To obtain the composition of an unknown waste mixture the fingerprints of single components are correlated with that of the mixture. It is assumed that the mixture curve is obtained as a weighed sum of the curves of its single components - in other words that no interaction takes place between the single components when they are mixed. Synthetic four-component mixtures with a known composition were prepared. The weighed-sum method was first applied to the TG curves and then to the DTG curves. The results show that modelling with the DTG curves distinguishes better between materials decomposing in a narrow temperature range. However, the decomposition temperature must differ with tens of degrees before the model can set two materials apart. Therefore, another approach is required. Instead of trying to distinguish between all single waste components, one could divide the single components into classes based on their chemical structure and DTG curves. This will be studied closer in the future.