| The increasing penetration of MEMS technology into new application domains suggests the need for sophisticated engineering tools that can automate routine MEMS engineering design functions. This thesis discusses the development of algorithms and automated software tools that are intended to automate the mask-layout process for bulk etch micro-machining.; At present, a designer conceives of a MEMS function, then (informally) creates a mask-layout that the designer believes will process into a shape that will exhibit the desired function. Because of the highly anisotropic nature of the bulk etching process, the mask design process relies heavily upon the designer's intuitive understanding of the etching process. A prototype device is created from the candidate mask, and its actual function is tested. This process can result in many iterations, and many prototypes.; This dissertation presents a method to automatically synthesize the mask layout for a bulk etching process. That is, given a desired part geometry and process parameters, the algorithm determines a candidate mask geometry that will etch to the final desired shape even in the case of highly anisotropic etchants. It will also compute compensation structures for difficult to etch features. Conceptually, the algorithm is based on the use of a forward etch simulation in reverse time. Since the forward etch process is a many-to-one map, the reverse time simulation is augmented to include a choice of valid preimages. Timing models are introduced to develop mask layouts that have appearing crystal planes during the etch and shown to allow more complex compenstation structures.; Finally, calibration masks are developed that allow the experimental determination of etch rate parameters without painstaking measurements. |
