Diffusion Modeling in a Microchannel for Separation of Species

Recently there is an increased interest in the design of microfluidic devices for research in biotechnological studies, applied to sample detection and analysis of species. When fluids are confined to small volumes, mixing results almost entirely by diffusion due to low velocities of flow in microchannels. As a result, it is possible to design microfluidic systems in which dissimilar fluids flow along side each other over long distances without significant mixing. The H-filter is a microfluidic device used for the extraction of molecular analytes from liquids containing interfering particles. The principle behind H filter is that small molecules will diffuse quickly from a sample stream to the buffer stream while very large molecules and particles will remain indefinitely in the sample stream because of their much larger size and much decreased diffusion rate. Because the Reynolds number in most microfluidic channels is generally kept well below 1, no turbulent mixing of fluids occurs. The only means by which solvents, solutes and suspended particles move in a direction transverse to the direction of flow is by diffusion. Differences in diffusion coefficients can be used to separate molecules of large particles over time. The time spent in flowing in a channel is proportional to the length of the channel. Before carrying out experiments, it is worthwhile to simulate the diffusion process in a microfluidic device for various properties of species and channel geometry. This paper attempts to model the diffusion process in an H-filter for typical species using CFD-ACE+, a software for solving problems in fluid dynamics with multi-physics capabilities. A module of CFD ACE+, called user-scalar that allows the user to define scalar quantities and boundary conditions for this scalar is used in the simulation. As seen from the studies, the diffusivities of species A and B in the buffer influence their diffusion. Optimization of geometry for a given species can be done with this method and separation can be achieved. The results from such a study will be useful for the design optimization and fabrication of such devices.