Microengineering Aerospace Systems
Henry Helvajian, editor
Chapter 1: Introduction to MEMS
M. Mehregany and S. Roy
1.1 Overview
Interest in the development of microelectromechanical systems (MEMS) has mushroomed during the past decade. In the most general sense, MEMS attempts to exploit and extend the fabrication techniques developed for the integrated circuit (IC) industry to add mechanical elements, such as beams, gears, diaphragms, and springs, to the electrical circuits to make integrated microsystems for perception and control of the physical world. MEMS devices are already being used in a number of commercial applications, including projection displays and the measurement of pressure and acceleration. New applications are emerging as the existing technology is applied to the miniaturization and integration of conventional devices.
This chapter starts with an overview of MEMS technology, followed by fabrication technologies, selected applications, commercial aspects, trends in MEMS technology, and journals and conferences (Table 1.1). The review covers both the potential and the limitations of MEMS technology.
| Silicon anisotropic etching | pre-1950 |
|---|---|
| Piezoresistive effect in silicon | 1953 |
| Semiconductor strain gauges | 1957 |
| Silicon pressure sensors | post-1960 |
| Solid state transducers | post-1970 |
| Microactuators | post-1980 |
| Mechanisms and motors | 1987–89 |
| Microelectromechanical systems Microsystems Micromachines |
post-1988 |
1.1.1 Historical Background
The transistor1 was invented at Bell Telephone Laboratories on 23 December 1947. This invention, which led to a Nobel Prize awarded in 1948 to Schockley, Bardeen, and Brattain, initiated a fast-paced microelectronic technology. The transition from the original germanium (Ge) transistors with grown and alloyed junctions to silicon (Si) planar double-diffused devices took about 10 years. The IC concept was conceived by several groups, and included RCA's Monolithic Circuit Technique for hybrid circuits (1955). The first IC, shown in Fig. 1.1, was built by Jack Kilby of Texas Instruments in 1958, using Ge devices (a patent was issued to Kilby in 1959). A few months later, Robert Noyce of Fairchild Semiconductor announced the development of a planar Si IC. The complexity of ICs has doubled every 2 to 3 years since 1970. The minimum dimension of manufactured devices and ICs has decreased from 20 µm to the submicron levels of today. Currently, ultra-large-scale-integration (ULSI) enables the fabrication of more than 10 million transistors and capacitors on a typical chip. ULSI-based microprocessors and microcomputers have revolutionized communication, entertainment, health care, manufacturing, management, and many other aspects of our lives. Low-cost, high-performance electronic systems are now available to the public and have improved the quality of life in many ways. However, control and measurement systems, as well as the actual automation facilities used in IC fabrication, would be "deaf," "dumb," and "blind" without sensors to provide input from the surrounding environment. Similarly, without actuators, control systems would be powerless to carry out the desired functions. While IC technology (and more specifically, microfabrication) has provided high-speed, miniaturized, low-cost signal conditioning and signal-processing capabilities, conventional sensors and actuators (also referred to as transducers) are far behind in performance, size, and cost.
The success of Si as an electronic material in ULSI technology was due partly to its wide availability from silicon dioxide (SiO2) (sand), resulting in potentially lower material costs relative to other semiconductors. Consequently, a significant effort was put into developing Si processing and characterization tools. Today, some of these tools are being utilized extensively to advance transducer technology. In this area, attention was first focused on microsensor (i.e., microfabricated sensor) development. Si microsensors initially addressed the measurement of physical variables, expanded to the measurement of chemical variables, and then progressed to biomedical applications.The first, and to date, the most successful microsensor, is the Si pressure sensor. The history of Si pressure sensors is representative of the evolution of microsensors.
1.1.1.1 Silicon Pressure Sensor Technology
|
Fig. 1.1. First integrated circuit consisting of one transistor, three resistors, and one capacitor. The IC was implemented on a sliver of germanium that was glued on a glass slide. (Texas Instruments, Inc) |
In 1953, Dr. Charles S. Smith of Case Institute of Technology (now part of Case Western Reserve University [CWRU]), during a sabbatical leave at Bell Telephone Laboratories, studied the piezoresistivity of semiconductors and published the first paper on the piezoresistive effect in Ge and Si in 1954.2 The piezoresistive effect is the change in the resistivity of certain materials due to applied mechanical strain. The measured piezoresistive coefficients indicated that the gauge factor of Ge and Si strain gauges could potentially be 10 to 20 times larger, and therefore much more sensitive than those based on metal films. As a result, discrete Si strain gauges were developed commercially in 1958 by Kulite Semiconductor Products, Honeywell, and Microsystems. Such Si strain gauges were integrated on a thin Si substrate as diffused resistors in 1961 by Kulite. The thin Si substrate was then mounted on a base to act as a diaphragm. In 1966, Honeywell developed a method to fabricate thin Si diaphragms by mechanically milling a cavity into an Si substrate. Isotropic Si etching was used to produce micromachined Si diaphragms in 1970, and anisotropic etching was introduced for this purpose in 1976. Both techniques were introduced by Kulite. Glass frits were introduced to bond the Si wafer (in which the pressure-sensitive diaphragms were fabricated) to a base wafer in the 1970s, allowing wafer-scale fabrication of pressure sensors. The first high-volume pressure sensor was marketed by National Semiconductor in 1974. This sensor included a temperature controller (in a hybrid package) for constant temperature operation. At this point, piezoresistive pressure-sensor technology had become a low-cost, batch-fabricated manufacturing technology. Further improvements of this technology have included the utilization of ion implantation for improved control of the piezoresistor fabrication, etch stops for improved control of the diaphragm thickness after the etch, deep Si-reactive ion etching for increased packing density, anodic bonding (electrostatic bonding), and more recently, Si-to-Si fusion bonding for improved packaging of the pressure sensors. Currently, Si pressure sensors are a billion-dollar industry and growing.3
The first monolithic integrated pressure sensor with digital (i.e., frequency) output was designed and tested in 197l at CWRU,4 as part of a program addressing biomedical applications. Miniature Si diaphragms with a resistance bridge at the center of the diaphragm and sealed to the base wafer with a gold (Au)-tin (Sn) alloy were developed for implant and indwelling applications. During field evaluation, it was found that the packaging of the sensors determined their performance, and that piezoresistive sensors were very sensitive to interference, such as sideways forces, making them inaccurate for many biomedical applications. To achieve better sensitivity and stability, capacitive pressure sensors were first developed and demonstrated at Stanford University in 1977 and shortly afterward at CWRU. The first integrated monolithic capacitive pressure sensor was reported in 1980.5 In general, capacitive pressure sensors exhibit superior performance compared to traditional piezoresistive pressure sensors. However, the relatively complex design and implementation of signal-processing circuitry required for electronic readout initially limited the widespread availability of capacitive pressure sensors. During the last 15 years, various processing and transduction techniques have been used to develop new or improved Si pressure sensor designs. While such developments are ongoing, advanced piezoresistive Si pressure sensors still account for almost all of the Si pressure sensor market. During the same period, Si microsensor technology has matured substantially, and a variety of sensors have been developed for measuring position, velocity, acceleration, pressure, force, torque, flow, magnetic field, temperature, gas composition, humidity, pH, solution/body fluid ionic concentration, and biological gas/liquid/molecular concentrations. Some of these sensors have been commercialized.
1.1.1.2 Micromachining
Development of Si microsensors often required the fabrication of micromechanical parts (e.g., a diaphragm in the case of the pressure sensor and a suspension beam for many accelerometers). These micromechanical parts were fabricated by selectively etching areas of the Si substrate away to leave behind the desired geometries. Hence, the term "micromachining" came into use around 1982 to designate the mechanical purpose of the fabrication processes that were used to form these micromechanical parts. Isotropic etching of Si was developed in the early 1960s for transistor fabrication. Anisotropic etching of Si was reported in 1967 by Finne and Klein6 and in 1973 by Price.7 Various etch-stop techniques were subsequently developed to provide further process flexibility. Together, these techniques have been used for fashioning micromechanical parts from Si materials, and they also form the basis of the "bulk" micromachining processing techniques. Bulk micromachining designates the point that the bulk of the Si substrate is etched away to leave behind the desired micromechanical elements.
While bulk micromachining has been a powerful technique for the fabrication of micromechanical elements, ever-increasing needs for flexibility in device design and performance improvement have motivated the development of new concepts and techniques for micromachining. For example, the application of the sacrificial layer technique (first demonstrated by Nathanson and Wickstrom in19658) to micromachining in 1985 gave birth to the concept of "surface" micromachining.9 Surface micromachining designates the point that the Si substrate is primarily used as a mechanical support upon which the micromechanical elements are fabricated. More recently, the introduction of Si fusion bonding and deep reactive ion etching, as well as high-aspect-ratio lithography and plating processes, have expanded the capabilities of micromachining technology.
Prior to 1987, Si micromachining had been used to fabricate a variety of micromechanical structures, such as thin Si diaphragms, beams, and other suspended structures, in single-crystal Si or in films deposited on an Si substrate. These micromechanical structures were generally limited in motion to small deformations and were physically attached to the substrate. Such elastic components could be used as flexible joints, but their overall usefulness in the design of "mechanisms" was limited. "Mechanism" as used here is a means for transmitting, controlling, or constraining relative movement and refers to a collection of rigid bodies connected by joints. During 1987 to 1988, a turning point in the field was reached when, for the first time, techniques for integrated fabrication of mechanisms on Si were demonstrated.10,11 It was then possible to fabricate mechanical parts that could execute unrestrained motion in at least one degree of freedom (e.g., gears, gear trains, linkages). Shortly thereafter, this technology enabled the development of electrostatic micromotors12,13 and motivated the development of other types of microactuators, such as valves, pumps, switches, tweezers, and lateral resonant devices.
1.1.1.3 MEMS
Recent progress in microactuators is transforming the conventional field of solid-state transducers into what has become known as MEMS. The term "MEMS" was coined around 1987, when a series of three workshops on microdynamics and MEMS was held in July 1987 in Salt Lake City, Utah; in November 1987 in Hyannis, Massachusetts; and in January 1988 in Princeton, New Jersey. These workshops ushered in a new era of microdevices. Equivalent terms for MEMS include "microsystems," which is preferred in Europe, and "micromachines," which is favored in Japan. MEMS is application driven and technology limited, and has emerged as an interdisciplinary field that involves many areas of science and engineering.
Miniaturization of mechanical systems promises unique opportunities for new directions in the progress of science and technology. Micromechanical devices and systems are inherently smaller, lighter, faster, and usually more precise than their macroscopic counterparts. However, the development of micromechanical systems requires appropriate fabrication technologies that enable the following features in general systems:
- Definition of small geometries
- Precise dimensional control
- Design flexibility
- Interfacing with control electronics
- Repeatability, reliability, and high yield
- Low cost per device
