Flow-Based Passive Microfluidic Architecture for Homogeneous Mixing

Accurate sample preparation, an essential component in microfluidic design automation, poses a great challenge to lab-on-chip (LoC) designers. Although continuous-flow based microfluidic biochips (CFMBs) offer many advantages, the problem of achieving fast, stable, and controllable sample concentration becomes more pronounced when such chips are used. Most of the flow-based dilutions and/or mixing methods need certain settling time to ensure an output flow with stable concentration profile resulting in initial loss of fluids. Control-valve sequencing along with time management for injecting input fluids further complicates the scenario. In this paper, we address the problem of preparing a homogeneous mixture of several input-fluids with different concentration factors. We present a binary tree-based free-flowing microfluidic architecture, which obviates the need for any control valve. The passive fluidic network supports balanced hydraulic resistance along all input-output fluid paths and ensures low Reynolds number. We show that the homogeneity of mixing can be expedited by properly assigning inlets to fluids with different concentrations based on a graph-matching algorithm. The method reveals how design automation tools can be effectively used for solving a problem of fluid dynamics. Theoretical attributes of the proposed architecture, when compared with COMSOL Multiphysics based simulations, provide excellent agreement with high accuracy.