Search

The use of a range of modelling techniques in the design of electrostatically actuated micromirrors has been demonstrated. Both simple 1D modelling and full 3D finite element techniques have their own merits and neither can be neglected at the cost of the other. However, as in the case of any modelling exercise, the data produced is only as good as the input information. Consequently, if the concept of the virtual fabrication laboratory is to be applied to the design of micromachined structures for MEOMS, accurate knowledge of both material properties and process tolerances remain key issues. (3 pages)

We review some of the attractive attributes of microengineering and relate them to features of the highly successful siliconmicroelectronics industry. We highlight the need for cost effective functionality rather than ultimate performance as a driverfor success and review key examples of polysilicon devices from this point of view. The effective exploitation of the datagenerated by the cost effective polysilicon sensors is also considered and we conclude that 'non traditional' data analysiswill need to be exploited if full use is to be made ofpolysilicon devices.Keywords: Polysilicon, sensors, arrays, data analysis 1. COST EFFECTWE FUNCTIONALITY For a technology to be commercially attractive it must provide cost effective functionality. There is no point in providing ahigh performance technology if the cost of the technology is two high. Many MEMS devices suggested in the literaturehave very attractive features, however they nearly always require the development of a new processing technology toimplement them and this may be the reason that despite nearly 30 years of MEMS research comparatively few devices are acommercial successes.Polysilicon micromachining has been described many times before,' in the context of cost effective functionality it hasseveral attractive features. Polysilicon process capacity already exists in the silicon microelectronics industry, the processdesign tools and layout tools exist. More importantly perhaps are the existence of well defmed design rules for polysiicon.While this may limit the freedom of the MEMS engineers design choices, it will also force many more applications to sharea common process base and so improve the fmancial viability ofboth the device and the process line.This is a feature that we see in the microelectronics industry 1-3