Infrared spectroscopy can be used to identify and quantify individual (bio)chemical species within complex mixtures provided the spectroscopic fingerprints are strong enough. It is often the case, however, that the characteristic fingerprints are concealed by the much stronger absorptions of the most concentrated species. For example, in the case of blood serum, the very strong absorptions of protein severely hinder detection of diagnostically relevant metabolites. A microfluidic device [1,2] is modeled here with the objective of separating serum metabolites from proteins, or – more precisely – to enhance the metabolite/protein concentration ratio, and in so doing to provide a new metabolite-rich sample for spectroscopic characterization [3-5]. Focusing upon a representative metabolite (creatinine) and protein (albumin) we have modeled the diffusion of serum components into a water stream when the two streams flow in contact through a microchannel. With the laminar fluid diffusion interface (LFDI) thus created, creatinine diffuses rapidly into the water, while the albumin hardly diffuses at all. Thus, at the channel exit one may recover a product containing significant creatinine but very little albumin. The goal of the simulations is to predict the composition of the product stream. To that end, we have tested various two-dimensional models to see if the essential features are contained in them. Various fluid mechanical insights are found: diverging flow at the inlet, converging and separating flow at the exit, and the difference between 2D and 3D simulation results.
分享
引用
