ABSTRACT
Chemists usually jump directly into optimization when experimenting on formulations, relying on their subject-matter knowledge to narrow down the field of components. The authors demonstrate screening for main-component effects can uncover vital ingredients that might otherwise be overlooked. Screening designs only fit the linear model. The authors begin by detailing of screening designs with ones that cover a simplex region, for example, when the amounts of each component vary from zero to one-hundred percent. The authors explain Appendix 6A, this plot displays the predicted response as any given component deviates from any chosen reference point, while holding all other components in constant proportion. The trace plot shows the effects of changing each component along an imaginary line from the reference blend to the vertex. When component constraints violate the simplex geometry, the authors recommend extreme-vertices screening design. The interpretation of linear effects must be modified to account for complex constraints.
