Cross-Flow Filtration of White Blood Cells

Virtual prototyping of a cross-flow filtration chip for the separation of WBCs

Chen et al [1] proposed two kinds of microfluidic chips based on the crossflow filtration principle which can be more effective than conventional filter types in the area of avoiding clogging or jamming. The pillar-type design allowed to achieve a 80% separation efficency at a length of 150mm. We tested the geometry in the virtual environment keeping all dimensions except the length. The virtual chip was shortened to explore the chip's efficiency for larger flow-through to total flux ratios. As can be seen in the animation below, increasing the flux ratio will lead to WBCs squeezing through the gaps between the pillars. At a constant flux ratio, even at lower global flow rates (i.e. lower stresses on the cells), the WBCs will get sucked into the gaps (but will get stuck and clog the filter). This was expected. The proposed design requires a minimum length to be efficient.

With the same flow parameters, alternative geometries were explored in virtual experiments. In the animation below a simulation of blood flow through an inclined array of obstacles is shown. The size of the obstacles and the gaps between them are the same as in the original geometry by Chen et al. In the alternative geometry the obstacles have a square cross-section and are slightly inclined to the orientation of the array creating a staircase. The new geometry has the effect that the stagnation point of the flow (the point at the obstacle where flow through the gaps and main flow are separated) is further downstream at the obstacle boundary and that even if an WBC is initially "sucked" into a gap, the stresses from the main flow and other cells will drag the WBC along the main flow. In the case of Chen's geometry with deeper gaps, WBCs are shielded from the main flow and are less likely to be dragged along. Still, some of the WBCs get stuck at the gaps in the new geometry, in particular at the beginning of the channel where shear rates at the obstacle array is lower. Another iteration on the design will be necessary. Doing this in the virtual environment this is no problem of course.


  • [1] X. Chen, D.F. Cui, Ch.Ch. Liu, H. Li, Microfluidic chip for blood cell separation and collection based on crossflow filtration, Sensors and Actuators B 130, pp 216, 2008.