Half-elliptical cross section microfluidic channels were fabricated in soda lime glass wafers using photolithography and etching techniques. An isotropic wet etch defining the channel geometry produced a half-elliptical cross section with a 55 μm major axis and 35 μm minor axis, as shown in Fig. 1(a). Fluorescent polystyrene nanospheres of diameter 970, 780, 390, 210, 93, and 60 nm presenting surface carboxyl groups (Bangs Laboratories, Fishers, IN) were used as model NCs. The zeta potentials of the 60, 93, 210, 390, 780, and 970 nm polystyrene particles in DI water were measured to be −40.4, −46.4, −45.5, −41.2, −38.1, −44.0 mV, respectively, with ±4.2 mV uncertainty for each value. Particles were flown at either a constant particle concentration of (1.0 ± 0.2) × 109 particles/ml or at constant particle volume fraction of 3.1 ± 0.6 vol. %. The particle solution was flown at a flow rate of 15 μl/min for all studies except for the flow rate dependent study, where particles were flown at 15, 25, or 35 μl/min through the straight glass channel. Because the viscosity of blood flow is higher than that of the DI water used in our study, these flow rates are larger than RBC flow rates reported in the literature, which vary from 0.03 μl/min  to 2 μl/min  for vessels of similar sizes, to obtain a comparable wall shear stress. With the viscosity of whole blood taken to be 3.5 mPa·s , the shear stress in real blood vessels can be calculated from the Hagen-Poiseuille equation to be on the order of 8.4 Pa . The shear stresses for the current study are estimated using the following equation from Bahrami et al.