Handbook of Silicon Based MEMS Materials and Technologies (eBook)

Veikko Lindroos, Sami Franssila, Markku Tilli, Mervi Paulasto-Krockel, Ari Lehto, Teruaki Motooka, V (Herausgeber)

eBook Download: PDF | EPUB

2000 | 1. Auflage
669 Seiten
Elsevier Reference Monographs (Verlag)
978-0-08-094772-3 (ISBN)

Infos

A comprehensive guide to MEMS materials, technologies and manufacturing, examining the state of the art with a particular emphasis on current and future applications.

Key topics covered include:

• Silicon as MEMS material
• Material properties and measurement techniques
• Analytical methods used in materials characterization
• Modeling in MEMS
• Measuring MEMS
• Micromachining technologies in MEMS
• Encapsulation of MEMS components
• Emerging process technologies, including ALD and porous silicon

Written by 73 world class MEMS contributors from around the globe, this volume covers materials selection as well as the most important process steps in bulk micromachining, fulfilling the needs of device design engineers and process or development engineers working in manufacturing processes. It also provides a comprehensive reference for the industrial R&D and academic communities.

• Veikko Lindroos is Professor of Physical Metallurgy and Materials Science at Helsinki University of Technology, Finland.
• Markku Tilli is Senior Vice President of Research at Okmetic, Vantaa, Finland.
• Ari Lehto is Professor of Silicon Technology at Helsinki University of Technology, Finland.
• Teruaki Motooka is Professor at the Department of Materials Science and Engineering, Kyushu University, Japan.

• Provides vital packaging technologies and process knowledge for silicon direct bonding, anodic bonding, glass frit bonding, and related techniques
• Shows how to protect devices from the environment and decrease package size for dramatic reduction of packaging costs
• Discusses properties, preparation, and growth of silicon crystals and wafers
• Explains the many properties (mechanical, electrostatic, optical, etc), manufacturing, processing, measuring (incl. focused beam techniques), and multiscale modeling methods of MEMS structures

Textauszug

Towards Mass Volumes of MEMS Devices

Early Visions

When Kurt E. Petersen wrote his famous review [1] of silicon micromechanics in 1982, there existed already eight companies working in the field. He covered nearly two hundred essential references in his paper. Above all the vision of the applications and opportunities was already well understood. Petersen discussed a very broad set of potential applications with commercial significance, such as, accelerometers, pressure transducers, torsional mirrors, resonant gate transistors, light modulators, resonating beam arrays, inkjet printer heads, and microelectromechanical switches. The research community was already formed: for example, the IEEE Transducers conference was organized already in 1969, and the Sensors and Actuators journal was established in 1980. The first large volume application of MEMS was only some years away.

Most of the concepts of micromechanical devices and fabrication technologies go back in time to the 1960s and the early 1970s. The mechanical characteristics of single crystalline silicon, the anisotropic etching of silicon by potassium hydroxide (KOH) and several concepts of micromechanical devices were studied [2–6]. An interesting example of early device concepts is shown in Fig. 1. W.E. Newell and his coworkers [7, 8] published a resonant gate transistor based on a micromechanical resonator made of electroplated gold in 1967.

Fig 1. Resonant gate transistor published in 1968 by William E. Newell and his coworkers [8].

Integration of micromechanical structures, such as resonators, was a natural continuum from the pre-silicon electronics to the early concepts of silicon integrated devices. The first commercial application by Hewlett Packard used a MEMS cantilever based frequency detector in frequency synthesizers in 1980 [9]. In spite of these early concepts, the development of practical commercial applications based on MEMS resonators has taken nearly 30 years, even today very few commercial devices exist besides the scientific instruments, such as, the resonating cantilevers of the atomic force microscopes [10, 11] and the biosensors [12].

The capability to create large arrays of similar components with precise dimensions enabled new device concepts. Furthermore, the possibility to benefit from the fabrication capabilities that were scaling up in the IC industry created the starting point towards the commercial applications of MEMS.

Ink Jet Printer Nozzles Create the Industry

This capability to manufacture precise components and arrays of micromachined structures practically enabled the ink jet printers during the 1980s. IBM demonstrated [13, 14] the value of silicon micromachining to achieve the necessary printing precision with the integrated methods to control and manage the inks in the same micromachined device. The first mass volume application of silicon MEMS was created, and the ink jet printers became the main stream of printing in the growing information technology market.

The electrostatic control of ink jets [15] for printing purposes was studied in the beginning of the 1970s. The capability to etch the ink jet nozzles into the silicon wafer using anisotropic KOH wet etching and to integrate the control electrodes into the same device using thin film and semiconductor processes enabled the sufficient miniaturization to improve the quality of ink jet printing to the level where the commercial solutions were possible.

Still today the ink jet printer application forms one third of the total MEMS market. Today the printing is expanding as a paradigm to electronics manufacturing [16]. Reel-to-reel manufacturing solutions based on either inorganic or organic inks will be one of the future solutions for manufacturing low cost electronics.

Automotive Applications Drive the Reliability and the Quality

The automotive applications of pressure and motion sensors practically created the MEMS industry. The manifold air pressure (MAP) sensor introduced by Ford in the mid seventies was the first micromechanical sensor in mass volumes. The accelerometers [17–19] were introduced to replace mechanical switches in airbag launchers, and later they enabled sophisticated chassis control systems.

The automotive industry already in the 1980s was characterized by advanced project management and quality control that included all the module manufacturers and subcontractors. The duration of product development projects was long and required a commitment for several years. The MEMS developers soon learned to apply these strict rules of project and quality management. The high requirements for reliability created long development projects with careful testing and verification phases.

The automotive applications also created the requirements for the sensor electronics, such as voltage levels and system interfaces, and for the sensor module packaging. Robust system-on-package solutions were created. Operating temperatures and shock tolerance were extremely demanding. In addition, the frequency dependence of the sensors in chassis control and airbag launchers required very careful control of intrinsic gas damping and structural parameters.

The motion sensors in automotive applications were soon divided into two categories: the high performance 3–5 g sensors for measurement of the motion of the chassis of the vehicle and the low cost 50–200 g sensors for the airbag launchers. This created the distinct paths for the development of the manufacturing solutions.

Leaps towards a Generic Manufacturing Platform

The two paradigms of micromachining develop almost in parallel. In the beginning the anisotropic etching of bulk silicon to form microstructures into the silicon wafer was a more efficient strategy [20]. The bulk micromachining benefits from the optimal mechanical characteristics of the single crystalline silicon. The doping of the silicon wafers and the optimization of their characteristics for chemical etching required specific development by the wafer manufacturers.

Even the simplest MEMS devices require insulating layers. The successful practical method [18, 20] was to use a sandwich of silicon and borosilicate glass wafers. The wafers were bonded together by a so called anodic bonding process. However, the difference in the thermal expansion of the glass and silicon wafers was a problem causing strong temperature dependence and even warping or buckling of the micromechanical structures. The solution was to use only very thin insulating glass layers between silicon wafers [19] . Later the glass manufacturers, such as, Corning and Hoya, introduced glass wafers with thermal expansion characteristics that matched the silicon.

Bulk micromachining enabled several different products [20]: pressure sensors, accelerometers, lab-on-chip devices, etc. As the bulk micromachining provided an inherent wafer level packaging of the micromechanical structures, the very accurate control of the gas damping of accelerometers or the reference pressure of the absolute pressure sensors was feasible.

The other paradigm was to grow a polysilicon thin film on top of a sacrificial silicon dioxide layer [21, 22]. The polysilicon film was anchored to the underlying silicon wafer and patterned to form the particular mechanical structure, e.g., the proof mass and the capacitor structures of an accelerometer. When the sacrificial silicon oxide layer was removed, the mechanical structure was released to move.

The promise of the polysilicon surface micromachining was in the integration of mechanical structures with CMOS electronics. The approach was very successful in the development of accelerometers for the airbag application that was not as demanding on the acceleration resolution. The thickness of the polysilicon layer determines the proof mass of the accelerometer, and the intrinsic acceleration resolution is inversely proportional to the square root of the proof mass. Thus the smaller mass of the polysilicon structures became a limiting factor for the application that required high acceleration resolution.

The control of the intrinsic stress of the polysilicon membranes was challenging and made the release of the micromechanical structures difficult. The deposition of a thick polysilicon layer was eventually developed [22]. After solving the challenge of wafer level encapsulation, the polysilicon surface micromachining provided smaller dimensions at lower cost, possibility for monolithic integration with CMOS devices and allowed more complex mechanical structures.

In the mid 1990s two disruptive technologies appeared. The deep reactive ion etching (DRIE) of silicon using an inductively coupled...

Produkt-Details

• Erscheint lt. Verlag: 1.4.2000
• Sprache: englisch
• Themenwelt: Sachbuch/Ratgeber
• Themenwelt: Technik ► Bauwesen
• Themenwelt: Technik ► Maschinenbau
• ISBN-10: 0-08-094772-7 / 0080947727
• ISBN-13: 978-0-08-094772-3 / 9780080947723