Screening design of experiment (DOE) applied to supercritical antisolvent process

This paper proposes to approach the supercritical fluid antisolvent process (ASES) by design of experiment (DOE). As screening designs, they allow for the identification of key variables at an early stage of experimentation, with only a few experiments. Seven factors have been studied, to which two levels were assigned. A fractional factorial design of 8 experiments, plus 3 additional runs to improve the precision of the estimates was performed, instead of the 2(7) experiments required by a full design. The process resp onses are the yield, the dryness of the produced powder, and the mean parti cle size. The screening design tends to indicate that recovery and drying r esponses use the same selection of factor levels for best operation, i.e., a low temperature, high flow rates for both solvent and antisolvent, and an extensive drying time. These responses are also quite insensitive to press ure and concentration. The relative importance of process parameters on the mean particle size is different, suggesting that a compromise in operating conditions should be found to obtain simultaneously acceptable levels of y ield, dryness, and particle size.