Project Description

Di-electrophoresis traps are used to separate biological contaminants such as viruses, bacteria or cells, in general, from fluids. The different polarizability of the biological material and the carrier fluid is used to achieve the separation.

In the commonly used single-particle model, electric field gradients polarize the individual cells. It is not considered in this model, that polarized particles produce an electric field themselves, which again influences the motion of other cells.

In this project a multi-particle approach has been pursued. Thereby, the previously untreated collective phenomena like dipole induced dipole charges and collisions, as well as hydrodynamic interactions due to the effect of the particle motion in the fluid are modeled. It is expected that the multi-particle approach shows formations of chains and clusters which significantly change the dynamics of the particles.

We use the Lattice-Boltzmann method which allows for the calculation of the impulse exchange of fluid and particles, and thus gives the drag forces applied onto the particles. The di-electrophoresis forces are computed by considering all existing electric fields, i.e. outer and induced fields.

Simulation example

A particle-loaded fluid is purified by deflecting particles using di-electrophoresis forces induced by an electrode and by exhausting the cleaner fluid layer.

In the single-particle approach, every particle is exposed to the same external force independent of its initial position. Therefore, all particles reach the same final position of about 30 microns.
In the multi-particle model, the final position of the particles depends on their initial position. The final position is in average much greater, i.e. about 55 microns, than in the single-particle case.
This simulation example shows that it is indispensable to use the multi-particle model since it has a strong effect on the predicted filter efficiency.

Further validation

A filter test system, which has been built during the project, allows to investigate filter performance depending on a number of parameters. The observed trends are qualitatively confirmed by the simulation (see picture on the left).