Breakdown of deterministic lateral displacement efficiency for non-dilute suspensions: A numerical study

We investigate the effect of particle volume fraction on the efficiency of deterministic lateral displacement (DLD) devices. DLD is a popular passive sorting technique for microfluidic applications. Yet, it has been designed for treating dilute suspensions, and its efficiency for denser samples is not well known. We perform 3D simulations based on the immersed-boundary, lattice-Boltzmann and finite-element methods to model the flow of red blood cells (RBCs) in different DLD devices. We quantify the DLD efficiency in terms of appropriate "failure" probabilities and RBC counts in designated device outlets. Our main result is that the displacement mode breaks down upon an increase of RBC volume fraction, while the zigzag mode remains relatively robust. This suggests that the separation of larger particles (such as white blood cells) from a dense RBC background is simpler than separating smaller particles (such as platelets) from the same background. The observed breakdown stems from non-deterministic particle collisions interfering with the designed deterministic nature of DLD devices. Therefore, we postulate that dense suspension effects generally hamper efficient particle separation in devices based on deterministic principles.