Accurate and efficient modeling of microfluidic devices   This project is aimed at overcoming limitations of design approaches that rely on time-consuming and expensive finite element simulation or trial-and-error prototyping. While the financial and temporal costs of trial-and-error experiments are clear, we note that full numerical simulations (e.g., using finite element methods) are inefficient as system-level design tools. For example, when a microfluidic design is modified even with minor changes, the numerical model in general would need to be rebuilt, which can be extremely labor-intensive and time-consuming. On the other hand, design of microfludics, like design of any other devices, is typically a process in which initial designs are iteratively modified. Full numerical simulations are generally not suitable as a tool for evaluation of such design iterations. We have been addressing this issue by developing microfluidic models that have both accuracy and efficiency. Such models are typically in closed form and parametrized: closed-form formulas allow rapid evaluation of device performance, and parameters embedded in the formulas allow efficient performance comparisons of various design choices. Expensive, time-consuming numerical simulations are hence not needed until design verification in the final design stage.    Projects: System-level models for electrokinetic microfluidic separations and mixing Modeling, design and experimentation of diffusion-based chemical concentration generation