A fluid restrained to submillimeter channels is typically dominated by complex surface forces from the channel walls rather than volumetric forces from the bulk fluid. Even in relatively simple microfluidic channel geometries, the movement of a suspension of deformable particles from one channel to another is affected by many processes, including long-range hydrodynamic and particle–boundary interactions.
Microfluidics is especially important in many biological applications because it describes how red blood cells move through similarly sized vessels in the human body. During his PhD with Chaouqi Misbah at Grenoble Alpes University in France, Zaiyi Shen—now a postdoc at the University of Bordeaux—and colleagues numerically simulated the flow of deformable particles, similar to red blood cells, traveling through a honeycomb-shaped network of channels. The results show that the particles’ motion depends on the path they’ve already taken—a so-called memory behavior.
In their model, particles without any memory behavior would roughly follow the trajectories taken by rigid spheres (black circles in the illustration) or soft nonspherical particles (blue polygons). But Shen and his collaborators found that shape and deformability affect the particle path: The rigid nonspherical particles (magenta ovals), which represent red blood cells moving through vessels, took a categorically different route. If one of those particles turned left at the first bifurcation, for example, it was more likely to take a left at each subsequent fork in the road. (In the model, each channel segment is about 100 μm long and 10 μm wide.)
In addition to simulating a front of particles spreading laterally across a channel network, the researchers modeled the particle motion for an initially homogeneously distributed set of particles, which is more similar to what’s expected for red blood cells in a real vascular network. In that simulation, the same range of motions was observed, with the deformability of a particle affecting its diffusion through the network. But at concentrations with a particle volume fraction greater than 15%, the additional particle–particle interactions helped produce a random path distribution rather than a deterministic one. In large vessels, the red-blood-cell concentration typically reaches 40–45% and has a range of 5–20% in microcirculations.
Although the simulations are limited to a honeycomb network, the model may be adaptable to more complicated disordered vascular networks found in living animals. (Z. Shen et al., Phys. Rev. Lett. 130, 014001, 2023.)
