Numerical modeling of flow, diffusion, and reactions in microfluidic microarray systems for oligonucleotide synthesis.

Abstract

One of the primary applications of microfluidic devices in life science research is the genetic analysis in DNA microarrays. DNA microarray technology allows parallel synthesis of molecular libraries, massively parallel assaying, and multiplex data acquisition, and plays an increasingly important role in biological and biomedical sciences including genetic disease diagnostics and classification and toxin detection. In this dissertation, the main objective is to study fluid flow, diffusion, and reactions in DNA microarray devices by a computational fluid dynamics (CFD) technique to aid the design and optimization microarray geometries and their operational procedures in parallel synthesis of DNA. First, the geometry effects on flow uniformity and pressure drop in such devices are investigated to design a new microarray platform that has the desirable flow uniformity within thousands of microreactors. The geometrical variables considered include microreactor depth, microchannel depth and width, and microreactor tapered inlet width within thousands of microreactors. The geometrical variables considered include microreactor depth, microchannel depth and width, and microreactor tapered inlet width and length. The microreactor depth and microchannel width are found to be the most significant effects to the improvement of flow uniformity. Second, the effect of pulsatile flow on the washing efficiency is studied to propose a washing method using minimum solvent amount. The results from various pulse lengths are compared to continuous flow feed suggesting that the washing solvent could be used more efficiently by applying the lowest-possible flow rate, if a longer washing time is still acceptable. If the pump reaches its minimum flow range limitation, the pulsatile feed style would also provide similar results in reducing the amount of solvent that is needed. The results also show that the fluid replacement in the microarrays is strongly affected by the diffusion mechanism. Lastly, a model to study the diffusion-reaction of a photogenerated acid (PGA) in detritylation process is developed based on experimental data. This model demonstrates the ability to replicate the important characteristics of experimental results which could not be previously predicted without experiments. Parameter estimations and their sensitivity analysis provide an important understanding of the PGA reaction kinetics.