Microelectromechanical systems (MEMS) integrate mechanical and electrical components and have feature sizes ranging from micrometers to millimeters. They may be fabricated using methods similar to those used to construct integrated circuits and they have the potential of providing significant cost advantages when batch fabricated. Their size also makes it possible to integrate them into a wide range of systems. Feature sizes may be made with size on the order of the wavelength of light, thus making them attractive for many optical applications. Microsensors (e.g., accelerometers for automobile crash detection and pressure sensors for biomedical applications) and microactuators (e.g., for moving arrays of micromirrors in projection systems) are examples of commercial applications of MEMS.
Microelectromechanical systems (MEMS) are micrometer-scale devices that integrate electrical and mechanical elements. They have been used in diverse applications, from display technologies to sensor systems to optical networks. MEMS are attractive for many applications because of their small size and weight, which allow systems to be miniaturized.
To further explore the challenges and opportunities of MEMS, this article describes some fundamental technologies associated with MEMS design and operation. First, the advantages and challenges of MEMS are discussed. Then manufacturing processes are presented, followed by the methods used to generate micro-scale forces and motions. Finally, several example MEMS applications are highlighted.
The small size of MEMS is attractive for many applications because feature sizes are typically as small as 1 micrometer or less. Hence, for optical applications, features may be made with size on the order of the wavelength of light. Their small size also allows applications to be developed which would otherwise be impossible. For example, micromechanical switches fabricated as part of a communications circuit allow phase shifting and signal switching at speeds that would be impossible to achieve using macro-scale switches. To illustrate the scale of a typical microsystem, Figure 1 shows a micromachined mirror assembly next to a spider mite. The mirror assembly is about 100 micrometer wide, and it is dwarfed by the spider mite. The mechanical devices surrounding the mirror allow it to be positioned accurately as part of an optical network.
Figure 1: A spider mite dwarfs a micromachined mirror assembly. Courtesy Sandia National Laboratories, SUMMiTTM Technologies, www.mems.sandia.gov.
Other advantages include the on-chip integration of electromechanical systems and the circuitry used to control them, allowing further miniaturization. Furthermore, many MEMS fabrication technologies allow parallel fabrication of thousands of systems by leveraging the parallel fabrication techniques of the integrated circuit industry. This may lead to a reduction in the manufacturing cost and improvement in reliability.
Like any technology, microsystems present some challenges. Because micromechanisms operate at a size scale far below that of typical mechanical devices, surface forces such as adhesion and friction may dominate over other forces in the system, leading to failure of the device. Careful device design, fabrication, and testing are required to reduce these effects. Their small size also makes it difficult to interact mechanically with MEMS components. In many microelectromechanical systems, electrical or optical signals are used to interface with, provide power to, and control the device, rather than manual, hydraulic, or pneumatic control typically seen in macro-scale mechanical devices. In addition, packaging of MEMS components has often presented a challenge because each device must be packaged in a way that keeps the components clean and free from contamination, while also allowing mechanical motion and, in many cases, interaction with the environment. For example, a MEMS pressure sensor requires a package that exposes the sensor to the ambient pressure while protecting the electronic circuitry from dust or other particles. Finally, while parallel fabrication techniques can reduce the manufacturing cost of many units, MEMS development may be more costly because few units are produced at a time using complex and expensive fabrication equipment.
Several books are available that can provide general information about MEMS [1, 2, 3, 4].
A large variety of fabrication methods have been employed for MEMS. However, many of these methods may be broadly described under three headings: surface micromachining, bulk micromachining, and molding.
Bulk micromachining, the oldest of the micromachining technologies, is accomplished by removing material from a substrate to create holes, cavities, channels, or other desired shapes. Early bulk micromachining was accomplished using isotropic or anisotropic wet etching of silicon or glass substrates. In particular, several chemicals, such as KOH (potassium hydroxide) or TMAH (tetramethylammonium hydroxide) etch a silicon substrate preferentially depending on the crystalline planes in the direction of etching. The etch rate for these chemicals is tens to hundreds of time faster in the  crystalline plane compared to the  plane. This effect has been used to create a wide variety of features using simple wet etching.
Another common technique of bulk micromachining uses a deep reactive ion etch (DRIE) plasma etcher. Using this technique, silicon, as well as some other materials, can be etched very quickly and very anisotropically, making possible very thick structures with small widths. For example, the mirrors in Figure 2 were deep etched from a titanium substrate. These mirrors are flatter and more rigid than mirrors made using surface micromachining.
Figure 2: Titanium mirrors bulk micromachined using deep etching of a titanium substrate.
Another important technology frequently used with bulk micromachining is wafer bonding. After etching the desired parts in several different substrates, the substrates may be bonded together to create systems incorporating several parts. In some cases, the wafers are bonded using an adhesive, but this method often has poor accuracy in spacing and alignment between bonded parts. Silicon and glass wafers can be bonded more accurately by applying high temperatures (about 1,000 degrees C) to a clean interface between the surfaces. In many cases, the bond generated in this way is as strong as the intrinsic strength of glass or silicon. If high temperatures are not desired, anodic bonding can be using. In this case, the wafers are pushed together at a lower temperature (about 400 degrees C), and then a large voltage (about 1,000 V) is placed between the wafers. This generates a high electric field at the interface, accelerating bonding even at the lower temperature. Wafer bonding has also been explored as a way to create a vacuum package enclosing MEMS parts.
Surface micromachining is one of the most common technologies used to manufacture MEMS devices. In surface micromachining, films are deposited on a substrate and patterned, using photolithography, to create micromechanical devices. The films normally alternate between structural and sacrificial layers, with the MEMS parts being made from the structural layers. The sacrificial layers serve to support the structural components during fabrication. After the structural layers are patterned, the sacrificial material is removed, often using wet chemical etching. The result is freestanding MEMS parts that can move relative to the fixed substrate.
Most early surface micromachining used polycrystalline silicon (polysilicon) as the structural layers and an oxide of silicon as the sacrificial material. However, as surface micromachining has further developed, numerous other materials have been used. Depending on the desired application, MEMS developers have used metals, oxides and nitrides of silicon, and even polymers for both structural and sacrificial films. Several foundry processes have also been developed to allow users to design their own MEMS and have them fabricated using the foundry's fabrication facilities. For example, Figure 3 shows a surface micromachined electrostatic drive manufactured using the SUMMiT foundry process developed at Sandia National Laboratories. The device is made from two layers of released polysilicon, with a layer of polysilicon. The fingers of the electrostatic drive are about 35 micrometer long, and each layer is about 2 micrometer thick. The gaps (between the substrate and the first layer, and between the two layers) are determined by the thickness of the sacrificial oxide film - 2 micrometer.
Figure 3: An electrostatic comb drive fabricated using SUMMiT, a surface-micromachining foundry process.
Finally, MEMS parts are often made by creating a mold, which may then be filled to create the desired part. Molds have been made from a variety of polymers, including some types of photoresist, as well as metal and deep-etched silicon wafers. Photolithography is normally used to define the mold pattern. If metal parts are desired, the mold may be filled by electroplating. Polymer parts may be created by pouring or pressing the precursor into the mold. After the part has been molded, it may be removed from the mold by either etching the mold away, or, if the mold is to be used again, by peeling away the mold.
Micro-molding was first performed in Germany, where it was called LIGA, an acronym for the German words lithography, electroplating, and molding. The original LIGA process required an X-ray source to fully expose thick layers of photosensitive material, but many molding techniques have since been developed that use visible or ultraviolet light sources. However, because of this history, many molding processes are still referred to as LIGA or LIGA-like processes. Figure 4 shows a copper waveguide fabricated using true LIGA technology. Because of its high thickness and exact specifications, the waveguide has very low losses, especially compared to other micromachined waveguides.
Figure 4: An extremely low-loss coplanar waveguide fabricated using LIGA.
Each of the processes discussed above has advantages and disadvantages, depending on the application. Several are summarized in the table below. The table describes the relative complexity of parts which are normally created using each process, the typical relative size of the MEMS resulting from each process, the complexity of the fabrication steps required, and the typical aspect ratio of the parts, meaning the ratio of their height to their width. The table is very general, but provides a rule-of-thumb comparison of the different processes.
Part Complexity Typical Size Fabrication Complexity Aspect Ratio Surface Micromachining High Small High Low Bulk Micromachining Low Large Low High Molding Medium-High Small-Large Medium High
On-chip actuation of microsystems has been a particularly challenging aspect of MEMS development. Common macro-level actuation approaches, such as hydraulics, pneumatics, electric motors, internal combustion engines and turbines, are either too difficult to fabricate at the micro level or do not work well at that scale. Electrostatic attraction is one approach that has been widely used for actuation of microsystems. While electrostatic actuation is suitable for many applications, some systems require either lower voltages or higher output forces. Electrostatic and thermal actuation approaches are described in more detail.
According to Coulomb's law, the electrostatic force acting between two charges is inversely proportional to the distance between the charges. For macro-scale objects, this force is normally negligible. However, micro-scale devices may have very small gaps, making electrostatic attraction an important source of mechanical motion. This actuation technology is especially attractive because it uses very little power. On the other hand, large voltages (typically tens to hundreds of volts) are required.
The simplest type of electrostatic actuator consists of a movable plate or beam which is pulled toward a parallel electrode under the application of a voltage difference. This type of actuator is illustrated schematically in Figure 5. The movable electrode is suspended by a mechanical spring, which is often simply a micromachined beam. When voltage is placed across the electrodes, opposite charges on each one attract each other. However, unless they touch, the electrodes only draw sufficient current to charge the actuator's effective capacitance, resulting in low power requirements. The attractive force is larger when the movable electrode is closer to the fixed electrode, with the force proportional to the reciprocal of the square of the gap.
Figure 5: A parallel-plate actuator consists of two parallel electrodes. Attractive electrostatic force pulls the movable electrode toward the fixed electrode.
Because of this inverse relationship, these parallel-plate actuators suffer from instability for voltages beyond a threshold known as the "pull-in" voltage. For voltages beyond the pull-in voltage, the electrostatic attraction grows more quickly than the mechanical restoring force, causing the electrodes to crash into each other. Unless the electrodes are protected with a dielectric coating or mechanical stops to prevent contact, this normally results in catastrophic melting or vaporization of the actuator due to sudden current flow between the electrodes. For systems with a linear spring, the pull-in voltage is the voltage which causes the movable electrode to deflect one third of the gap. Hence, these actuators cannot be stably operated for deflections larger than this. However, many types of mechanical springs, including fixed-fixed beams, exhibit nonlinear deflection characteristics, leading to a larger usable deflection range.
Comb drive electrostatic actuators avoid the pull-in instability and remove the dependence of the force on the deflection. As with parallel-plate actuators, comb drives consist of one fixed and one movable electrode. The electrodes are shaped like interdigitated combs, however, as shown in Figure 6. Using straight combs, like those shown in the figure, results in an attractive force that is nearly constant over a wide range of deflection of the movable comb. The attractive force falls rapidly when the combs disengage and rises rapidly when the combs approach full engagement, but most comb drives are designed to operate entirely in the constant-force region.
Figure 6: A comb-drive actuator uses interdigitated comb electrodes to produce an attractive force which does not depend on actuator displacement.
Comb drives also suffer from instability when the applied voltage is too high, but this instability is not directly related to the deflection of the movable comb. As long as each comb finger is perfectly centered between two opposing comb fingers, the net transverse force acting on that finger will be zero. However, it is common for the fingers to become slightly off-center during motion. If the resulting transverse force is large enough, it will either cause bending of the fingers or it will pull the entire comb in the transverse direction. Hence, for sufficiently large applied voltage, the fingers on opposing combs can touch, which is frequently catastrophic due to melting or vaporization of the fingers.
A change in temperature causes an object to undergo a change in length, where the change is proportional to the material's coefficient of thermal expansion. This length change is usually too small to be useful in most actuation purposes. Therefore, a method of amplifying the displacement is an essential part of thermal actuators. Figures 7a to 7d illustrate four examples for achieving amplification of thermal expansion in microactuators.
Bimetallic devices use two materials with different coefficients of thermal expansion that are fused together. As the temperature increases, one material expands more than the other and the actuator bends to accommodate the different deflections. An example is demonstrated in Figure 7a [5, 6, 7]. Challenges associated with bimetallic actuators include somewhat complicated fabrication and the potential for delamination of the layers.
Figure 7: Example thermal amplification approaches, including (a) bimetallic, (b) pseudo-bimorph, (c) geometry-based amplification, and (d) a thermomechanical in-plane microactuator (TIM).
While bimetallic actuators heat both materials to the same temperature but exploit their differences in coefficients of thermal expansion, pseudo bimorphs use a single material with a uniform coefficient of thermal expansion, but with different parts experiencing different temperature changes. This approach makes it possible to construct an actuator from a single layer of the same material. An example is the device shown in Figure 7b [8, 9] which has one leg thinner than the other. An electric current runs through the legs, but the thin leg will have a higher electrical resistance and will heat up more than the wide leg. As the hot leg expands it will cause the actuator to rotate in the direction shown.
Another approach for amplifying the thermal expansion is to use geometric constraints that force the actuator to move in the desired direction. An example of this type of amplification is the "bent beam" actuator, illustrated in Figure 7c [10, 11]. As the thin legs heat up, the expansion causes an amplified deflection in the direction shown.
The Thermomechanical In-plane Microactuator (TIM) also exploits geometric constraints [12, 13], as illustrated in Figure 7d. It consists of thin legs connecting both sides of a center shuttle. The leg ends not connected to the shuttle are anchored to bond pads on the substrate and are fabricated at a slight angle to bias motion in the desired direction. As voltage is applied across the bond pads, electric current flows through the thin legs. The legs have a small cross sectional area and thus have a high electrical resistance, which causes the legs to heat up as the current passes through them. The shuttle moves forward to accommodate the resulting thermal expansion. Advantages of this device include its ability to obtain high deflections and large forces, as well as its ability to provide a wide range of output forces by changing the number of legs in the design. A scanning electron micrograph of a TIM is shown in Figure 8.
Figure 8: A scanning electron micrograph of a TIM.
One of the most visible commercially available microelectromechanical systems is Texas Instruments' Digital Micromirror Device (DMDTM) which is used in applications such as portable projectors, rear-projection televisions, and cinema projectors. The DMD is a rectangular array of moving micromirrors that is combined with a light source, optics, and electronics to project high quality color images .
Figure 9 shows the architecture of a single DMD pixel. A 16 micrometer square aluminum mirror is rigidly attached to a platform (the "yoke"). Flexible torsion hinges are used to connect the yoke to rigid posts. An applied voltage creates an electrostatic force that causes the mirror to rotate about the torsion hinges. The electronics and structure are designed to allow the mirror to be rotated by 10 degrees in either of two directions (the "on" and "off" positions). When tilted in the on position, the mirror directs light from the light source to the projection optics and the pixel appears bright. When the mirror is tilted in the off position, the light is directed away from the projection optics and the pixel appears dark.
Figure 9: Architecture of the Texas Instruments Digital Micromirror Device (DMD). (Illustration courtesy of Texas Instruments.)
The micromirrors can be combined in an array on a chip, and each micromirror is associated with the pixel of a projected image. Figure 10 is an illustration of two pixels in the on and off states. The micromirrors can be switched thousands of times per second. A gray scale image is obtained by varying the amount of time each pixel is on or off, with lighter gray pixels associated with mirrors being on more than off. Color images are created by combining this switching method with a color filter, such as a color wheel. The on and off states of a mirror are coordinated with flashes of red, green, and blue light to create the combination needed to create the desired color. An illustration of the combined system is shown in Figure 11. In high-end applications, such as cinema projectors, three DMD chips are used, with each chip dedicated to either red, green, or blue, and each pixel is created by combining light from a mirror from each of the three chips.
Figure 10: Two mirrors in the on and off position. A mirror in the on position directs light toward the projection optics. The mirrors are shown transparent to show device components. (Illustration courtesy of Texas Instruments.)
Figure 11: The Texas Instrument DMD shown as part of their Digital Light Processing (DLP) technology. (Illustration courtesy of Texas Instruments.)
The DMD architecture nicely illustrates several MEMS concepts - a few of these are:
A wide variety of optical components may also be implemented in MEMS. For example, the Digital Mirror Device discussed above uses micromachined mirrors to redirect light. MEMS mirrors have also been used for optical switching applications, allowing optical communication to be routed without requiring conversion to electrical signals. Adaptive optics systems using MEMS mirrors have been built to correct distortions due to air refraction or lens anomalies.
MEMS optical waveguides are often used to route optical signals within a MEMS optical chip. These waveguides consist of a core with low loss at optical wavelengths. Using micromachining, the waveguides can be patterned on the same chip as other optical components. Both mechanically suspended and fixed waveguides have been demonstrated. Suspended waveguides can also be designed to deflect mechanically. Hence, they can also switch or attenuate a signal when the waveguide is moved in and out of alignment with other components.
Micromachined lens arrays have also been demonstrated for optical applications. Figure 12 shows an example of three types of micromachined lens arrays. Similarly, diffractive gratings can be made using the fine dimensional control available from micromachining. These lens arrays and gratings have been used in optical filters and switches.
Figure 12: Three types of MEMS lens arrays: cylindrical (top), circular lenses, square packed (middle), and hexagonal lense, hex packed (bottom).
A sensor is a device that responds to a physical input (such as motion, radiation, heat, pressure, magnetic field), and transmits a resulting signal that is usually used for detection, measurement, or control. A transducer (often used as a synonym for sensor) is a device that is actuated by power from one system and converts it to a different form to another system. Advantages of MEMS sensors are their size and their ability to be more closely integrated with their associated electronics.
Piezoresistive and capacitive sensing methods are among the most commonly employed sensing methods in MEMS. Piezoresistance is the change in resistivity caused by mechanical stresses applied to a material. Materials with high piezoresistivity (such as some semiconductors which have more than an order of magnitude higher piezoresistivity than metals) are useful for transducing mechanical deformation to electrical signals. This is particularly useful in applications such as pressure sensors and accelerometers.
Capacitive sensors rely on the physical input being sensed to cause a change in capacitance. This capacitance change can be caused by changing the distance between the capacitor plates (e.g. pressure pushing two plates closer together) or by changing the dielectric (such as relative humidity sensor using a dielecric with a permittivity that changes with moisture content). The resulting change in capacitance can be very small and specialized electronics are required to detect the changes and convert them into a usable output signal.
An example of MEMS sensors include bulk micromachined pressure sensors, which have been commercially available since the 1970's. A typical design is illustrated in Figure 13. A cavity is etched to create a thin diaphragm which deflects under pressure. A backside port is etched in another substrate and bonded to the first. Piezoresistive pressure sensors have piezoresistive elements on the diaphragm that change resistance as the pressure increases. Another approach is to use the diaphragm as a plate in a capacitor and to detect the capacitance change as the diaphragm deflects under pressure.
Figure 13: An example of a MEMS pressure sensor.
Accelerometers are another example of commercially successful MEMS sensors. Applications include automotive airbag safety systems, mobile electronics, hard drive protection, and others. These have been successful enough that Analog Devices, a leader in MEMS sensors, had shipped over 200 million MEMS inertial sensors by April 2005. Figure 14 illustrates an example of a surface micromachined capacitive accelerometer. An acceleration causes a displacement of the inertial mass and the capacitance change between the comb fingers is detected.
Figure 14: A sketch of a capacitive MEMS accelerometer.
Other MEMS sensors include rate sensors, gyroscopes, radiation sensors, gas sensors, microphones, and mass flow sensors, to name a few.
Several types of MEMS components have been designed to operate in radio-frequency communications circuits. Low-power MEMS filters, variable capacitors, and switches have all been identified as promising MEMS components of RF communications systems. MEMS filters use mechanical vibrations to filter RF signals. They have demonstrated extremely low-power operation. MEMS variable capacitors are used in tuning circuits and oscillators.
MEMS switches are especially attractive because they exhibit "nearly ideal" switch behavior . When on, their insertion loss is typically about 0.2 dB or less, and off-state isolation is normally 20 dB or better even at high frequency (20-40 GHz). In addition, MEMS switches are normally electrostatically actuated, so that they consume very little cycling power.
Two types of MEMS switches have been used. The simplest type opens and closes a contact between micromachined metal electrodes. For example, the switch shown in Figure 15 consists of a suspended gold beam that can be pulled down using electrostatic force to make contact with an underlying gold contact electrode. Metal contact switches have very wide bandwidth and excellent performance. However, if they carry too much power, heating and degradation of the contact can occur.
Figure 15. A metal contact RF MEMS switch. When contact is made between the suspended gold beam and the underlying gold electrode, current can flow.
Capacitive switches do not suffer from this limitation. Capacitive RF switches also consist of a suspended electrode which is pulled toward a second electrode; however, a dielectric layer separates the two electrodes to prevent the flow of electrons between them. Instead, a capacitive switch works by changing the capacitance between the two electrodes. The ratio of on-state to off-state capacitance can be as high as 200. Because DC current does not flow between the electrodes, capacitive switches can often carry more power than metal contact switches. However, because capacitive impedance is a function of signal frequency, they operate in a more narrow frequency band than metal contact switches.
In 1984, Hewlett-Packard introduced the Thinkjet printer, the one of the first desktop printers to use inkjet technology. The technology was based on micromachined inkjet print heads used to expel drops of ink onto paper in well-defined patterns. Since that time, the technology has developed and spread, until many manufacturers offer inkjet printers based on micromachined print heads 
Inkjet printing depends on ink drops being ejected through micromachined orifices to create the desired pattern. Many different techniques have been used to eject the ink drops. Figure 16 illustrates a micromachined printhead that uses one method, called thermal inkjet printing by Hewlett Packard, where it was developed. In this method, a thin-film heater inside an ink-filled cavity heats a thin layer of ink. A bubble forms as the ink layer is superheated. The bubble rises out of the cavity through the micromachined orifice, carrying with it a drop of ink, which is expelled toward the paper. The size of the ink drop can be controlled by designing and fabricating an appropriately-sized cavity and orifice. Other methods for ejecting ink drops rely on mechanical pumping motions, often using piezoelectric materials. This application of micromachining has become extremely wide-spread in the printing industry; in addition, it is starting to be expanded to other industries, including automotive fuel injection, drug delivery, and other areas where precise control of fluid volume is required.
Figure 16. A micromachined inkjet printhead. The thin-film heater heats a thin layer of ink, causing a bubble to form. Expulsion of the bubble through the orifice ejects a drop of ink onto the paper.
Achieving motion at the micro level presents some interesting challenges. Because bearings are not feasible and lubrication is problematic, friction and wear present major difficulties. Assembly of parts at this scale is difficult. The constraints introduced by the planar nature of MEMS fabrication also introduce a number of unique challenges in constructing mechanical devices.
Nature provides an example of how to affectively address problems with motion at small scales. Smaller living organisms are more likely to use the deflection of flexible members to achieve motion. In a similar fashion, compliant mechanisms gain their motion from deflection of flexible members and they present solutions to many of the problems discussed above. The advantages of compliant mechanisms at the micro level include the following : Can be fabricated in a plane
Figure 17 illustrates a bistable switch that employs compliant members with long flexible legs that combine the function of hinges and springs . Other examples of compliant MEMS include the thermal actuators of Figure 7, the mirror torsional hinges in Figures 2 and 9, the pressure sensor of Figure 13, the accelerometer of Figure 14, and the RF switch of Figure 15.
Figure 17: An example of a compliant MEMS device. It is a bistable switch that uses flexible members to achieve the function of hinges and springs.
The performance of compliant mechanisms is highly dependent on the material properties, which are not always well known at this scale. In spite of this, tests have demonstrated that micro compliant components can be very robust at the micro level.
MEMS offer opportunities to miniaturize devices, integrate them with electronics, and realize cost savings through batch fabrication. MEMS technology has enhanced many important applications and it holds great promise for continued contributions in the future.
 Maluf, N.; An Introduction for Microelectromechanical Systems Engineering; Artech House: Norwood, MA, 2000.
 Gad-el-Hak, M., Ed.; The MEMS Handbook; CRC Press: New York, NY, 2002.
 Madou, M.J.; Fundamentals of Microfabrication: the Science of Miniaturization; CRC Press: New York, NY, 2002.
 Senturia, S.D.; Microsystem Design; Kluwer Academic Publishers, 2001.
 Barth, P.W.; Beatty, C.C.; Field, L.A.; Baker, J.W.; Gordon, G.B. A robust normally-closed silicon microvalve. Solid-State Sensor and Actuator Workshop 1994, 248-250.
 Kabei, N.; Kosuda, M.; Kagamibuchi, H.; Tashiro, R.; Mizuno, H.; Ueda, Y.; Tsuchiya, K. Thermal-expansion-type microatuator with paraffin as the expansive material. JSME International Journal 1997, 40, 736-742.
 Reithmuller, W.; Benecke, W. Thermally excited silicon microactuators. IEEE Transactions on Electrical Devices 1998, 35 (6), 758-762.
 Comtois, J.H.; Michalicek, M.A.; Barron, C.C. Electrothermal actuators fabricated in four-level planarized surface micromachined polycrystalline silicon. Sensors and Actuators A 1998, 70, 23-31.
 Huang, Q.A.; Lee, N.K.S. Analysis and design of polysilicon thermal flexure actuator. Journal of Micromechanics and Microengineering 1999, 9, 64-70.
 Que, L.; Park, J.S.; Gianchandani, Y.B. Bent beam electro-thermal actuators for high force applications. Proceedings of the 12th IEEE International Conference on Micro Electro Mechanical Systems 1999, 31-36.
 Park, J.S.; Chu, L.L.; Siwapornsathain, E.; Oliver, A.D. Long throw and rotary output electro-thermal actuators based on bent-beam suspensions. Proceedings of the 13th IEEE International Conference on Micro Electro Mechanical Systems 2000, 680-685.
 Cragun, R.; Howell, L.L. Linear thermomechanical microactuators. Microelectromechanical Systems (MEMS) at the ASME International Mechanical Engineering Congress and Exposition 1999, 181-188.
 Lott, C.D.; McLain, T.W.; Harb, J.N.; Howell, L.L. Thermal modeling of a surface-micromachined linear-displacement thermomechanical microactuator. Sensors & Actuators: A. Physical 2002, 101 (1-2), 239-250.
 van Kessel, P.F.; Hornbeck, L.J.; Meier, R.E.; Douglass, M.R. MEMS-based projection display. Proceedings of the IEEE 1998, 86 (8), 1687-1704.
 Prophet, G. MEMS flex their tiny muscles. EDN 2002, 47 (3), 63-72.
 Le, Hue P. Progress and trends in ink-jet printing technology. Journal of Imaging Science and Technology 1998, 42 (1), 49-62.
 Howell, L.L. Compliant Mechanisms; John Wiley & Sons: New York, NY, 2001.
 Wilcox, D.L; and Howell, L.L. Fully compliant tensural bistable micro-mechanisms (FTBM). Journal of Microelectromechanical Systems 2005, 14 (6), 1223-1235.