Title: Practical comparison of traditional and definitive screening designs in chemical process development

Authors: Rebecca Olsen; John Lawson; Nathaniel Rohbock; Brian Woodfield

Addresses: Department of Chemistry and Statistics, Brigham Young University, Provo, UT 84602, USA ' Department of Chemistry and Statistics, Brigham Young University, Provo, UT 84602, USA ' Department of Chemistry and Statistics, Brigham Young University, Provo, UT 84602, USA ' Department of Chemistry and Statistics, Brigham Young University, Provo, UT 84602, USA

Abstract: Traditional Screening Designs, such as resolution III 2^k-p fractional factorials are used routinely in the initial stages of process development. These designs are used to determine which process variables have the largest effect on process outcomes. Once a screening design is complete and the data are analysed, follow-up experiments are normally required in order to develop useful prediction equations involving the important variables and to identify the optimal process operating conditions. Recently developed definitive screening designs allow researchers to identify important variables and optimum process conditions after one set of experiments, eliminating the need for follow-up experiments. This leads to the question: What is now the role of traditional and definitive screening designs in process optimisation? We share our insights gained from using both of these designs in developing a process to produce catalyst support material to shed light on these questions.

Keywords: catalyst support; definitive screening design; resolution III; traditional screening design; chemical process development; process optimisation.

DOI: 10.1504/IJEDPO.2016.080534

International Journal of Experimental Design and Process Optimisation, 2016 Vol.5 No.1/2, pp.1 - 22

Received: 15 Feb 2016
Accepted: 11 Sep 2016
Published online: 28 Nov 2016 *