Intermediate Product E is a commodity chemical made from the reaction
of Monomer C with Monomer D in the presence of a catalyst. Fluctuation
s in the yield and color of Product E were observed at a production pl
ant. A screening, fractional factorial, experimental design was used t
o initially identify which process parameters were affecting the yield
and color of Intermediate Product E. Of the six process parameters in
vestigated, namely, temperature, reaction rime, catalyst concentration
, reactant concentrations, water concentration and stoichiometry, only
three of the factors were found to have a statistically significant e
ffect on the color and yield. These three factors were further evaluat
ed with a face-centered, factorial design which revealed the significa
nt nonlinearity effects and allowed for the modelling of a response su
rface. The complexity of this process might never have been discovered
without the aid of the structured and powerful approach of statistica
lly designed experimentation. The design steps, the analysis, and the
conclusions and recommendations leading to subsequent experimentation
are discussed in this paper. (C) 1996 Elsevier Science Ltd.