Currently most MEMS designers begin the geometric design of a new device by creating the masks that would lead to a geometric model. At the macro level, this is analogous to generating a geometric model from the tool paths, which would be a very difficult task. In contrast to MEMS designers, designers of macro devices have the advantage of starting with a geometric model and being able to directly visualize or manipulate their designs. The geometric model is then queried to generate the process specific data. In the case of MEMS, there is no systematic means to generate the mask data after the geometric model of a MEMS device has been refined through behavioral simulations.
This paper focuses on automatically generating masks, given a geometric model of the MEMS device and the process sequence (referred to here as the inverse problem). This necessitates the systematic solution of the forward problem, which involves automatically generating a geometric model of the MEMS device given the masks. A systematic and implementation-independent framework for the geometric modeling of MEMS is presented in order to solve the forward and inverse problems for general surface-micro-machined devices. However, all implementations and examples are two-dimensional, i.e., they do not deal with complexity in the third dimension.