(741b) A Hybrid Control Technique to Achieve Spatial and Temporal Control of Droplets in a Microfluidic Device

Droplets moving in micro-channels act as micro reactors providing a unique environment to carry out reactions precisely. Spatial and temporal control of these droplets give advanced flexibility to perform high throughput screening and combinatorial chemistry. Such systems have huge potential in developing continous manufacturing technology for pharmaceutical industries. Control of these micro reactors can be achieved through active and passive control techniques. Although passive methods provide good control without any moving components, they are sensitive to experimental disturbances. In contrast, active control techniques can handle experimental uncertainties with ease; further, designing passive elements require several experimental trails. Therefore based on the broad applicability of the droplet systems, implementing advanced control techniques in the field of droplet microfluidics is essential.

Droplets affect the decisions, at bifurcation points, and velocities of the other droplets in the channel due to hydrodynamic interactions. The dynamics of these phenomena even in a simple bifurcating channel is astonishingly complex. The present dropletsâ?? decisions affect past droplet decisions and droplet velocities. Droplet motion in microfluidic systems is highly nonlinear and interactive leading to chaos and presents a unique challenge for conventional control techniques. In this paper, we demonstrate the application of a hybrid model predictive controller on a microfluidic loop device to produce droplets of desired sequence given an input sequence. A custom MINLP technique is formulated for droplet systems that efficiently predicts the actuations to control both decisions-discrete states- and velocities-continous states- of droplets. This technique is demonstrated in microfluidic loop device by generating a droplet sequence `AABB' from `ABABâ??.