Research in the field of detergent formulation is deeply involved in the optimization of numerous detergent systems. The use of experimental designs in looking for optimal formulations can save a large amount of laboratory time. This paper describes a coherent strategy of use for a special class of simplex designs adapted from Scheffé’s work. The technique applies to multicomponent systems, as they are generally met in detergent formulation. Proper designs provide polynomial equations and, in addition, convenient graphical representations that enable the chemist to predict responses for a wide range of mixtures. Calculations are easy. A computer may help for routine application of these designs, but is not strictly required. The method is illustrated by a problem in the solubilization of light duty liquid detergents and at the same time provides arguments for better use of statistics in formulation. Finally, this simplex method is positioned among some other statistical techniques dealing with mixture designs.
[1]
Raymond H. Myers,et al.
Response Surface Designs for Experiments with Mixtures
,
1968
.
[2]
John W. Gorman,et al.
Simplex Lattice Designs for Multicomponent Systems
,
1962
.
[3]
Virgil L. Anderson,et al.
Extreme Vertices Design of Mixture Experiments
,
1966
.
[4]
G. R. Hext,et al.
Sequential Application of Simplex Designs in Optimisation and Evolutionary Operation
,
1962
.
[5]
M. N. Das,et al.
Design and Analysis of Experiments with Mixtures
,
1968
.
[6]
G. Box,et al.
A Basis for the Selection of a Response Surface Design
,
1959
.
[7]
J. Kiefer,et al.
Optimum Designs in Regression Problems
,
1959
.