Characterization And Evaluation Of Geometric Features Of MEMS Microstructure

In recent years, Micro-Electro-Mechanical Systems (MEMS) with their own, such as small size, light weight, low power consumption, high-speed, high sensitivity, has been very rapidly applied to the aero space, military, biomedical, information, automation and other fields. A metrology infrastructure has underpinned all industrial revolutions, and this infrastructure is weak or nonexistent for many of the proposed MEMS devices. Such systemic theoretical research is more urgent and needs to be carried thoroughly on. Metrology and testing technology of MEMS is one of the key requirements of MEMS designing, fabricating, quality control and evaluation. The main research contents of this paper fall into several parts as below:the theories of analysis, extracting, characterization and evaluation of MEMS microstructure are proposed, and the mathematical modelling for datum plane of characteristic microstructure as well as the identification of characteristic functional regions with the microstructure division technology are studied; the research above mentioned will systematically study the metrology and testing technology of MEMS, and establish the initial standard evaluation system of MEMS characteristic structure, and meet the further needs of MEMS technology. For MEMS geometrical parameters, there is no standard definition acceptable for worldwide, which brings much difficulty in MEMS communication, that is why it need to be further described.One of MEMS microstructure surface is most composed of a number of steps, grooves and slots; the other is a type of pattern structures included hemisphere, cylinder, and pillar. Pattern analysis and recognition of these microstructures with particular feature is one of the key problems in metrology and testing technology of MEMS.In this paper, the mathematical model to evaluate MEMS microstructure is shown by extracting datum plane of characteristic microstructure. The reference plane for evaluation is defined by the least square method and B-spline theory, a group of statistical parameters base on the reference plane are presented.This paper explains that how the SU-8 photoresist lithography process and the character of material caused the microstructure surface roughness and the error of geometrical size and shape-position error, and how these parameter affect the delivery efficiency of MEMS devices.Based on the wolf pruning morphological segment and a novel method of pattern detection are proposed, which can detect functional areas in most situations. It is the most important that this method is consistent topologically. Experiment results prove the efficiency of the method proposed even in measurement data with very low SNRThe geometrical parameter of MEMS microstructure and roughness of sidewall are distinguished. In line with the development trend of microstructure surface characterization, as well as the practical demand in the semiconductor and lithography industry, the evaluate procedures, bring forward a group of parameters to describe character of MEMS microstructure, is given out.