Use of Factorial Design of Experiment and Response Surface Methodology for the Study of the Effects of Different Types of Oil on the Yield of Soap.

Abstract

The work was carried out to study the yield of soap using different oils and oil blends. The experimental design, randomization, optimization and statistical analysis were carried out using Box Benken Design Expert 10.0. The design of experiment was done using four independent factors (variables), at three categorical levels. The optimum oil yield for palm kernel oil (P.K.O), coconut oil, P.K.O and olive oil, P.K.O and coconut oil blends and olive oil and coconut oil blend are 179.90 g, 150.22 g, 118.90 g, 130.99 g and, 105.00 g, respectively. The yield of soap at these optimized values for P.K.O, coconut oil, P.K.O and olive oil, P.K.O and coconut oil blends and olive oil and coconut oil blend are respectively, 123 g, 12.09 g, 79 g, 88.6 g and59.99 g. This showed that P.K.O is the best oil for soap production for it guarantees very good yield of soap, and other oils could be blended with PKO when special soaps are required. The coefficient of determination (R2 ) for P.K.O, coconut oil, PKO and olive oil, PKO and coconut oil blend, olive and coconut oil blend are 0.8007, 0.6649, 0.9739, 0.9266 and 0.9131, respectively. Mathematical models developed were found significant (p <0.05) for predicting the response variable. The various contour and 3-D surface graphs plotted can be used to determine the response variable in lieu of the mathematical models.