Three-Dimensional Surface Profiling of Microstructures, 10-9327Printer Friendly Version
Inclusive Dates: 07/01/02 - 09/30/03
Background - The objective of this project is to design and demonstrate a three-dimensional (3-D) measurement capability suitable for characterizing micro-electromechanical systems (MEMS) devices on a probe station. SwRI is developing a MEMS quality assurance program area that will include materials testing and device characterization capabilities. Testing, characterizing, and validating the design of MEMS devices requires accurate three-dimensional measurement of devices. Available methods for 3-D measurement are expensive and not well suited for use on a MEMS probe station. A unique capability for microscale 3-D measurement will strengthen the Institute's position in this emerging area.
Approach - The approach selected is to extend a 3-D imaging capability developed in a previous IR&D project to microscale measurements. The previous project resulted in a unique 3-D measurement method using dynamic structured light (DSL 3-D measurement). A patent application has been filed for this concept based on mathematical construction of a family of quadric surfaces generated by projection from a rotating grating. Preliminary tests indicated that the method could be extended to microscale measurements by projecting the structured light pattern onto a small area and imaging the pattern with an optical microscope. Preliminary calculations show that it should be possible to achieve a 3-D resolution of 100 nanometers. Other preliminary calculations indicated that a much faster method of calculating the 3-D coordinates could be implemented, especially if optimized, low-level code is written for a fast processor.
Accomplishments - A design spreadsheet, developed in the earlier 3-D project, was revised to be more appropriate for microscope optics. The spreadsheet was used to determine system parameters for a measurement resolution of 100 nanometers. Projection systems were investigated, and a non-collimated approach was selected. An improved optical projection system was designed and fabricated.
The signal processing computation method was modified and optimized to generate run-length-encoded data directly from the camera video signals, reducing the computation time by two orders of magnitude. The system was assembled on a probe station microscope and tested. Difficulties were experienced in illuminating the mirror-like surfaces of MEMS devices, determining the correct focal length for the infinity-corrected microscope objective lenses, identifying a suitable surface (smooth at a nanometer scale) for calibration and obtaining focused images over a sufficient depth range for accurate calibration.
Solutions were found for these problems, the system was calibrated and several MEMS devices were measured in 3-D. Measurement noise was found to be approximately ±500 nanometers, which is five to ten times greater than needed for successful use in MEMS characterization. While the project did not achieve the goal of accurate measurements of MEMS devices, it did result in several developments that are important additions to the DSL 3-D measurement capability. The computational algorithms were revised and coded to obtain a 100X reduction in the measurement time and the grating projector was redesigned to a much more compact and robust device. These results are being used in current projects and are an important part of development efforts for this technology. Some of these future projects may lead to a better understanding of error sources that could be used to improve the microDSL system to obtain usable MEMS measurements.