A penalty method to model particle interactions in DNA-laden flows.

We present a hybrid fluid-particle algorithm to simulate flow and transport of DNA-laden fluids in microdevices. Relevant length scales in microfluidic systems range from characteristic channel sizes of millimeters to micron scale geometric variation (e.g., post arrays) to 10 nanometers for the length of a single rod in a bead-rod polymer representation of a biological material such as DNA. The method is based on a previous fluid-particle algorithm in which long molecules are represented as a chain of connected rods, but in which the physically unrealistic behavior of rod crossing occurred. We have extended this algorithm to include screened Coulombic forces between particles by implementing a Debye-Hückel potential acting between rods. In the method an unsteady incompressible Newtonian fluid is discretized with a second-order finite difference method in the interior of the Cartesian grid domain; an embedded boundary volume-of-fluid formulation is used near boundaries. The bead-rod polymer model is fully coupled to the solvent through body forces representing hydrodynamic drag and stochastic thermal fluctuations. While intra-polymer interactions are modeled by a soft potential, polymer-structure interactions are treated as perfectly elastic collisions. We demonstrate this method on flow and transport of a polymer through a post array microchannel in 2D where the polymer incorporates more realistic physical parameters of DNA, and compare to previous simulations where rods are allowed to cross. We also show that the method is capable of simulating 3D flow in a packed bed micro-column.