Implementing a robust methodology: Experimental designs and optimisation

The aim of any atomic spectroscopist, as of any analytical chemist, is to develop robust and efficient measurement procedures. Experimental design, sequential optimisation and response surface methodology are three critical tools to extract maximum information from a limited number of systematised and planned sets of experimental trials. The basics of these methodologies are introduced here and some common strategies presented in a user-friendly, practical way and their advantages and drawbacks explained, from full factorial designs to saturated designs, which allow minimum performance of experiments. Tagugh's designs and their philosophy, optimisation techniques and main response surface designs are introduced as well. Several matrices of experiments are presented so that they can be used directly by the reader. In addition, an updated review of published works, with comments related to applications in atomic spectrometry, is included.