JOURNAL of the Tunisian Chemical Society

The study of the robustness of a process related to the preparation of sulfated amide of pomace oil fatty acids has been carried out in two steps. First, we performed a Plackett - Burman screening design to select from eleven potentially influential factors those which are really influent. Second, we reduced the variation domain of the four significant factors resulted from the Plackett - Burman design and considered them among the 31 parameters which may disturb the process efficiency if little variations of their levels occur when performing the reactions. The study of the process robustness has been then carried out by applying a two-level supersaturated design related to a half fraction of a 36-experiment Hadamard matrix. The total variance of the response in this design, compared to the repeatability variance showed that the process is affected by the deliberate small variations in its parameters. Multiple regression methods namely ridge regression, stepwise selection procedure, and all-subset regressions were used to analyze the supersaturated design data. Results indicated that five factors, namely, molar ratio SO₃/Ester, amidation time, amide addition rate, alkali concentration, and amidation temperature, remain active despite the decrease of their variation domain and thus they must be well controlled when performing the process.

F. Rais, A. Kamoun, M. Chaabouni, M. Claeys-Bruno, R. Phan Tan Luu, M. Sergent

Robustness, Plackett-Burman, supersaturated design, sulphate amide of fatty acids, olive pomace oil

Pages: 191-202