5] Response surface methods for optimizing and improving reproducibility of crystal growth

Abstract

Publisher Summary This chapter discusses response-surface methods for optimizing and improving the reproducibility of crystal growth. Response-surface experiments produce a functional representation, or surface, to represent the behavior of a system—that is, its response, as input conditions change; they provide a coherent way to simultaneously study relationships between a measurable result—such as crystal shape or volume—and a local region of several input variables that influence that result—such as protein concentration, crystallizing-agent concentration, and temperature. Response-surface experiments are generally used to determine the approximate conditions for which the response assumes a maximum or other desirable value. As with screening, experiments are performed, according to a multidimensional factorial design, and scored. However, this approach should be distinguished from screening because the experimental matrix is designed with a view to fit the results to a quadratic polynomial function in all significant variables. The advantages of using response surfaces to identify optimal conditions for crystal growth include improvements in crystal volume and shape and improvement in reproducibility. These benefits outweigh the cost of such experiments, and as response-surface methods become more familiar, their use may increase.