Modelling solutions of hydrocarbons in dense CO2 gas

Shaping of advanced ceramic parts generally requires the use of organic additives which, obviously, have to be removed prior to sintering. In order to obtain free defect green pieces in a short time, an alternative technique to the classical thermal debinding, based on the unique properties of dense gases such as supercritical fluids, has proven to be very efficient. Supercritical extraction of organic additives from a green ceramic part involved both solubilisation and diffusion. The debinding treatment has to be conducted to remove a high amount of the organic phase but in such a way that the cohesion of the green part was maintained. Then, the capillary migration of a molten organic additive has to be avoided and a low amount of binder has to remain in the green part after the debinding treatment. This residual organic phase could be easily removed during the sintering treatment because the porosity is, at this stage, entirely open and interconnected. In this context, extraction was performed using carbon dioxide, (i) at low temperature (lower than 80°C) for which the organic additives used, i.e. paraffin waxes composed of n-alkanes, remain in a solid state and, (ii) under experimental conditions of extraction (pressure, temperature, time) chosen to maintain a non soluble part of the binder. It is then necessary to be able to predict the solubility of organic binders in supercritical CO2 to define a suitable organic formulation (appropriate paraffin waxes) and an adapted debinding treatment. A review of existing models of solubility of organic molecules in supercritical fluids and an analysis of the principal concepts of modelisation are exposed. The choice of an equation of state for solubility prediction of n-alkanes in supercritical CO2 was made on the basis of (i) the values of average errors between calculations using models of solubility and experimental values of solubility measured in this study and, (ii) the facility of equation employment. Then, calculated values of solubility, using this equation of state, were compared to experimental values (FT–IR) for an alkane (n-C28).