Situation

Today, the design process for micro electronic circuits and micro chip layouts can be carried out under highly standardized conditions using well-defined working environments. Engineers have access to extensive model libraries, allowing simulation of complex circuit designs before manufacturing. The underlying model descriptions for electrical components take into account functional behavior as well as technological parameters. Normally, the obtained simulation results yield reliable predictions concerning system performance and can be verified by measurement.

However, there is no equivalent in micro system design. Micro systems often combine quantities from different physical domains as they allow electrical, mechanical, fluidical or thermal interaction. Developing simulation models for such hybrid micro systems usually results in custom solutions, which are highly specific and provide little reusability. Optimal designs are often found by heuristic approaches as trial and error methods, including many redesign cycles. This can lead to disadvantages concerning economic parameters like manpower, development expenses and time to market.

Micro fluid systems are the basis for multiple new analytic approaches in bio- and medical technology. Accurate modeling of such systems is urgently needed in the design phase. At present it is not possible for manufactures and developers to fall back on uniform design standards, because there is a lack of general valid behavioral models to describe micro systems. Furthermore, the high numerical effort prevents, in most cases, a consideration of such systems on the finite element level.

A good sample for those applications are Medical Test Strips made from plastics. They are an innovative approach towards cheap and therefore seminal one-way diagnostic systems and can be considered as a step towards the long discussed "lab on a chip". Main field of application is the analysis of body liquids as blood and urine in regard to the diagnose of antibodies for certain diseases or the examination of other particular characteristics. After the liquid under investigation has been filled into a reservoir, capillary forces lead to a spread through the micro channel, which is structured with small columns of defined dimensions to increase the fluidic resistance. To obtain defined volume flow and therefore defined time for chemical reactions, the geometry parameters structure pitch and structure heights have to be adjusted. Several physical parameters have to be taken into consideration. The channels are structured with small columns or lamellar of defined geometric dimensions to increase the surface and therefore the fluidic resistance.

On the one hand, the underlying physically effects of such systems are well-known; however, on the microscopic level, they are not yet completely analytical described. For instance, the ratio of surface effects to volume effects changes significantly by increasing the size of the structure.

At ITEM models are developed, which are partly based on heuristic and experimental rudiments. This previous research has shown promising results and will be improved by further investigation in the framework of this SCiE project.

Engineering Aspects

The models used so far have been developed using some heuristic ideas, thus improvements and extension are necessary. On the one hand we will consider the conversion from characteristic curves of the behavioral model to the physical model and their feedback. On the other hand additional physical aspects (e.g. the contact angle between up to three different materials, density and viscosity of the involved liquids, as well as the etching angle of the fabrication process) can be integrated into the models and special attention shall be payed to the multi-scale structure of micro fluidic systems.

For example the simulation of the blood flow is a major problem for the enhancement of existing models. Blood is a two-phased fluid composed of a carrier fluid, the blood plasma, and a hard substance, the blood cells (hematocrit). Due to their complex consistency and the particular properties of the components (e.g. non Newton flow medium), modeling the human blood is very difficult.

Mathematical Aspects

The existing models have been numerically evaluated by finite volume methods. Other numerical methods will be investigated, e.g. adaptive finite element methods, that are more appropriate for such multi-scale problems. In order to improve the integration of fluid dynamics and its simulation into the over-all model we will use model reduction methods.