Bioparticle Separation in Microfluidic Devices for in-Line Application

Bioparticle Separation in Microfluidic Devices for in-Line Application

Zhang, L.

French, P.J. (promotor)
Boscche, A. (promotor)

Electrical Engineering, Mathematics and Computer Science

Microelectronics & Computer Engineering

2009-11-20

There was an explosive growth in the bioprocess industry market during the last decade. The tight control of these processes is often very critical in order to optimize the process efficiency or even achieve the right product. Capillary electrophoresis (CE) system is a good option for process monitoring and controlling, since it has already been proved to be a powerful analytical tool used in laboratory situations. However the extremely high separation voltage required by CE operation is preventing the system miniaturization and integration, and thus hindering its way to in-line application. Therefore, in this work, two low-voltage separation methods are proposed and investigated with the aim of substituting the conventional microfluidic CE. In the first DEP method, the curved parts of a serpentine channel are effectively used for dielectrophoretic separation. The second method moving field CE is basically an evolution version of the traditional CE, in which the required separation voltage is significantly reduced by decreasing the span of the electrodes in the channel. For the DEP method, the curves are the essential parts where the dielectrophoretic separation takes place. A DC-biased AC electric field allows small particle separation in a curve as the dielectrophoresis is decoupled from the electroosmosis. The separation sensitivity was found to have a second order dependence on the ratio of AC to DC amplitudes and a linear dependence on the particle size. The separation resolution is limited by the Brownian motion and is voltage dependent. Since the curves are bended in the opposite directions, the separation will be wiped out in the next curve. Therefore, focusing must be conducted to bring all the particles to the same entering point for the next curve. In terms of focusing, the top-and-bottom electrode structure performs superior to the planar array for negative DEP because stronger electric field gradients are generated and particles are forced away from walls where nonuniform flow and nonspecific particle adhesion can occur. Because the hydrodynamic flow profile in a curve also decreases monotonously from the channel center to the outer sidewall, the same device can be used for both hydrodynamic and electrokinetic DEP separation. The hydrodynamic DEP,bearing a higher flow gradient, is expected to produce a larger separation than the electrokinetic DEP. The DEP device was fabricated by bonding of two processed substrates with an SU-8 layer sandwiched in between. The retention time of 2 µm and 4 µm PS particles in a channel segment containing 4 curves was measured in the case of hydrodynamic DEP. The 4 µm particles turn out to take an average 10% longer retention time than the 2 µm particles, which proves the concept. It was revealed that applying a DC voltage over a fraction of the channel will evoke flow distortion in moving field CE operation, and will in turn cause plug dispersion and peak broadening. For a given voltage, the smaller fraction it is applied on, the larger flow distortion will be caused in the EOF region. By extracting and analyzing the equivalent circuits of the fluid system, the flow distortion was found to be the consequence of a backward flow. The method we proposed is to generate, by coupling an electroosmotic pump (EOP), a forward flow and compensate the backward flow in the EOF segment. Specific pump structures were designed to relieve the driving voltage to the same level as the separation voltage. They are an array of slender channels and a broad thin channel. In order to minimize the flow loss, the channel arrays are proposed to be used in the injection channels as well to increase the flow resistance. The moving field CE prototype device was fabricated by laminating two layers of dry film resist on a processed glass substrate. The first layer is used to define the channel structure and the second acts as the lid of the channel. The flow distortion compensation was observed when the EOP driving voltage was increased to a certain level. The required EOP driving voltages with respect to different separation voltages were measured, which exhibit a linear relationship as expected. To detect cells at extremely low concentrations, we additionally proposed a biomass concentration detection method in which an electrokinetic technology is used for rapidly increasing the concentration of the cells while measuring the concentration. Due to the dielectric contrast between the viable cell and the surrounding medium, the cells could be trapped and accumulated by positive dielectrophoresis. Based on the simplified homogeneous sphere model, the frequency dependence of the induced dipole strength of the yeast cell was studied by computer methods. The threshold DEP forces required to dominate over the Brownian motion for cell trapping were investigated. To achieve accurate and sensitive detection, a contactless 4-electrode conductivity detector was suggested to be used. Readout circuits and measurement setups were verified by using ideal resistors to mimic the environment of a liquid suspension. This method is recommended to be further investigated by using viable yeast cells in future research.

microfluidics
separation
capillary electrophoresis
dielectrophoresis

http://resolver.tudelft.nl/uuid:167d7879-7f61-4c5e-9390-7d1435e9eda1

9789081331678

Institutional Repository

doctoral thesis

(c) 2009 Zhang, L.