(Dr. Thomas Kurfess, advisor)
"Two-Dimensional Analysis of Turbine Blades and Nozzles"
The current tools for geometric analysis of MEMS are primarily limited to those of the semi-conductor industry. These tools are suited for measuring entities that are two-dimensional in nature such as lines, circles, and planes. Hardware that is capable of collecting three-dimensional data is typically limited by the slope variations in the surfaces of the part. They cannot accurately capture information from steep sidewalls, especially in parts fabricated using the LIGA microfabrication process.
This research will develop a methodology to accurately qualify MEMS, specifically those fabricated from the LIGA process. The main issue that will be addressed is the need for three-dimensional analysis of LIGA parts. The first objective of this research will be to ascertain the feasibility of acquiring three-dimensional data from LIGA parts. Current hardware tools are lacking, as discussed above, but new techniques or small-scale applications of older technologies may prove to be useful. After this initial study, the need to use three-dimensional data for inspection of high aspect ratio parts will be evaluated. One of the main attributes of the LIGA process is the ability to fabricate relatively tall, parallel sidewalls. It may be possible to analyze two-dimensional data sets and infer sidewall characteristics with some accuracy. The main issue to be addressed here is the alignment of the two data clouds to perform the analysis. Depending on the geometry of the parts, a unique solution to the registration problem may or may not exist. This methodology will describe the possible registration errors based on the type of geometries being analyzed and the type of data acquired. If two-dimensional data sets can be correctly registered, implementing a methodology to inspect these parts from just the planar surfaces (i.e., top and bottom) and being able to infer the characteristics of the part sidewalls will be very significant.
Once a methodology is established, research and development will commence
on developing a software platform to implement the proposed methodology. This
platform will be based on a typical computer-aided inspection (CAI) program.
This work will implement algorithms for all the procedures typical to a CAI
program (e.g., P-S-A, registration, data analysis) in addition to new techniques
suited for inspection of high-aspect ratio MEMS.