The development of bio-MEMS and lab-on-a-chip devices requires efficient mixing of fluids. With this aim, a variety of passive and active micromixers have been developed. Computational Fluid Dynamics (CFD) performs now an important role in the development of microfluidics components, basically for the analysis of flow fields to adjust the design parameters and for the evaluation of the final design. But intuition and experience of the designer is usually behind the application of CFD for the design improvement. Therefore, the process has basically an empirical approach and it is still a build-and-test design procedure. There have been very few attempts in which automatic optimization techniques are exploited in the design of microfluidics components.

A new design and optimization strategy for microfluidics devices is presented which systematically integrates CFD with an optimization methodology based on the use of Design of Experiments (DOE), Function Approximation (FA) and Multi-Objective Genetic Algorithm (MOGA). The DOE method is a statistical approach to explore the design space for a given set of design parameters; the DOE provides an “experimental table” of design points. These designs are then evaluated with CFD to obtain the corresponding performance parameters values. The design and performance parameters are used by the FA method to create the objective functions (response surfaces) that correlate them, on which a sensitivity analysis can be carried out. Finally, a MOGA is run on the objective functions to determine the Pareto Front (PF) of optimum solutions and the best compromise of the performance parameters can be chosen depending on the required design specifications.

Passive micromixers are attractive to MEMS engineers since no extra forcing mechanism is required for mixing. The Staggered Herringbone Mixer (SHM) has been studied and its geometric parameters optimized by applying the proposed methodology which also allows the evaluation of their impact on the mixing performance; the evaluation results are in agreement with previous published studies. Because high mixing efficiency usually results in large pressure drop, they both are used as the performance criteria and the resulting optimum designs in the PF combine the highest mixing efficiency with the lowest pressure drop at different Reynolds numbers for any combination of the design parameters.

The design and optimization strategy presented here is an automated approach to the design of passive micromixers based on their own topology manipulation and for different flow conditions.