Reversal of flow in Stokes regime was demonstrated in a filmed experiment by G.I.Taylor. Aref and Jones numerically showed that separation of particles based on diffusion can be enhanced by combining reversibility of chaotic advection in Stokes flow with differential diffusive irreversibility of the particles. Separation of particles by diffusive irreversibility (SDI) could be an extremely simple method of chemical separation that would not require electromagnetic fields or membranes. Hence SDI is desirable in several micro-fluidic applications and bio-chemical processes. I have investigated laminar flows in a micro-channel patterned with staggered herring-bone shaped grooves; these flows exhibit Lagrangian chaos. The study of the reversibility of the trajectories of tracers (passive and diffusive) in a chaotic flow in stokes regime offers an opportunity to understand the interplay of diffusion, geometry and inertia in controlling the mixing process in the flow field.

I will present experiments and simulations performed to understand the flow and to characterize the reversibility as a function of Peclet number. I will discuss the Stretch model proposed by Ranz and its applications to the prediction of the extent of reversibility within the given flow. I will use the characteristics of chaotic flow evaluated using the simulations in the model to predict a scaling for mixing, and the percentage of return of particles on reversal. I will compare the results with numerical simulation and experiments performed in a chaotic flow with the results from the model. I will investigate the performance of a chaotic and non-chaotic flow to show that chaotic flows achieve efficient separation by decoupling mechanical stirring from diffusive mixing. I will discuss the design and experimental implementation of microfluidic device that works on the principle of SDI. The experiments and the model together could improve our understanding of the fundamental aspects of reversibility in chaotic Stokes flows.