Design of Magnetic Blood Cleansing Microdevices through experimentally validated CFD modeling

J. Gómez-Pastora, C. González-Fernández, I.H. Karampelas, E. Bringas, E.P. Furlani, I. Oriiz
Flow Science Inc.,
United States

Keywords: magnetic particle separation, flow patterns, mass transfer, CFD modeling, two-phase liquid-liquid microfluidic systems, blood detoxification

Summary:

Continuous flow bioseparators allow magnetic particles that are properly functionalized so that they get attached to dangerous pathogens, to be easily captured in a buffer solution. Specifically, the unique properties of magnetic particles enable their separation from blood, after fulfilling their role of pathogen capture, in a continuous process [1,2]. However, the optimization of the process, i.e. achieving higher magnetic particle recovery while minimizing blood loss or dilution and avoiding diffusion between fluid phases, remains a technological challenge. Therefore, in this study, we introduced a CFD-based Eulerian-Lagrangian approach using the commercial CFD software package FLOW-3D (www.flow3d.com) to model particle magnetophoresis in a two-phase continuous-flow microseparator where the particles were continuously separated from the blood stream and collected into a co-flowing buffer solution. A combination of mass transfer, magnetic and fluidic computational models were used to accurately describe the particle motion, along with any effects on the interface between phases, and the diffusion of blood components to the buffer phase. A numerical CFD analysis was used to predict the particle-fluid transport whereas an analytical approach is employed for the prediction of both the magnetic field generated by permanent magnets and the corresponding magnetic force on the particles. The parametric analysis focuses on the impact of different variables, such as flowrates, particle and magnet sizes and the viscosity of the fluid phases on particle recovery. The model was also experimentally validated through fluorescence microscopy. The simulation data fits the experimental results within an absolute error of 15%. Overall, this computational analysis helps promote the fundamental understanding of the underlying phenomena while offering a powerful platform for the design of magnetophoretic components that can be integrated into lab-on-a-chip systems for bioseparation processes.