Experimental Designs: Response Surface Design


Central Composites are designed to estimate the coefficients of a quadratic model i.e. one can model curvature. They are ideal for predictive models.

Figure 1: Diagram depicting a Central Composite Design and example of a Response Surface Model showing the dependence of the product formation against Reagent 1 and Reagent 2.
Figure 1: Diagram depicting a Central Composite Design and example of a Response Surface Model showing the dependence of the product formation against Reagent 1 and Reagent 2.

Once key parameters have been identified using screening designs, optimisation designs such as response surface design are used to more precisely model the curvature around our reaction optimum. It allows one to get deeper understanding of a process and identify the region in which the process conditions satisfy one or more goal.