Altekar, Maneesha (1991, ASQC) Hercules Incorporated, Wilmington, DE
While studying mixture variables, studying the effects of process variables may be also be helpful. This paper discusses techniques for applying mixture variables in developing resins for lithographic applications. The objectives were (1) to develop correlations between reaction conditions and raw material formulations with various physical properties of the resins, and (2) to develop models that would enable one to select reaction conditions and monomer formulations based on the properties desired in the product.
A combination of a mixture design and a fractional factorial design (Cornell, 1981) were used to collect the data for the experiment. First, a mixture design was generated for the formulation variables; then, at each set of conditions, a fractional factorial design was developed. The main effects of process variables can be estimated by analyzing the interaction effects of a combined mixture-factorial design in addition to the main effects of the formulation variables.
The study was successful in obtaining comprehensive and useful models for all properties of interest. In most cases, a first order polynomial was an adequate fit to the data. The results of this study have provided chemists and engineers with a knowledge base that will enable them to save time and be more cost effective as they develop future products and applications.
Experiments,Manufacturing,Process improvement,Statistics,Data analysis