Partial factorial designs for multivariate calibration: extension to seven levels and comparison of strategy

Multilevel experiments are necessary in chemistry for calibration and mixture experiments. This paper presents methods for generating designs for seven concentration levels (l=7) using l2 experiments paying special attention to the problem of orthogonality between successive factors. In addition, a method to extend all l level designs to orthogonal arrays with double the number of mutually orthogonal factors is shown. Some comparisons are made between different designs using leverage. Finally, recommendations are made in order to select the most appropriate design for each specific situation.