"A Litte Push," Feature Article, Oct. 2002

a little push

When machine dimensions run in millionths of a meter, a tiny movement could go a long way.

by Larry L. Howell and Timothy W. McLain

Microelectromechanical systems have found their way into such commercial applications as air-bag accelerometers, pressure sensors, and computer projectors, but forecasters predict that this is only the beginning for these tiny devices. Technology is advancing rapidly in fabrication, modeling, design, materials, and packaging.

A particularly challenging aspect of microsystem development lies in the actuation of the micromachines. Traditional actuation approaches, such as hydraulics, pneumatics, electric motors, and internal combustion engines, are either too difficult to fabricate at MEMS sizes, or simply do not work well at those scales.

Electrostatic attraction is one phenomenon that has been widely exploited for actuation at the micro level. Although capable of handling the actuation needs of many applications, its high voltage requirements and low output force capabilities restrict its use from applications that need higher output force or require voltages compatible with on-chip electronics. Electrostatic actuators have turned up in micromotors demonstrated by Sandia National Laboratories, for example, and in some experimental materials-testing devices.

A thrust in micrometers: As they carry current between the bond pads, the legs of the thermo-mechanical in-plane microactuator heat up and expand, to deflect a translating shuttle.

Another phenomenon seldom used for actuation at the macro level, but very useful at the micro, is thermal expansion. Although the magnitude of thermal expansion is small, the resulting deflections can be amplified to produce large displacements. Early thermal microactuator designs were bimetallic devices that used materials with different coefficients of thermal expansion—similar to the principle at work in thermostats. As the temperature increases, one material expands more than the other, and the actuator bends to accommodate the different deflections.

The bimorph idea was simplified in "heatuators," developed at the Air Force Institute of Technology, which used only one material but obtained bimetallic behavior by having one part heat up while another part remained cold. More recently, researchers at the University of Pennsylvania have built on the original heatuator idea to embed electrothermal-compliant actuators in MEMS devices.

Potential applications are switching of optical components such as mirrors and shutters, or operating tiny valves in microfluidic systems.

Moving away from the bimorph concept, a new type of microactuator uses geometric constraints to amplify the motion resulting from thermal expansion. An example of such an actuator is the thermomechanical in-plane microactuator, or TIM, developed by the authors and graduate students at Brigham Young University as part of research funded by the National Science Foundation.

A challenge in the development of MEMS products lies in the difficulty of actuation on the micrometer scale. Each of these scanning electron micrographs shows a bistable switch using two thermal actuators. Either design could be used as a microscopic electronic switch or relay.

The TIM is a planar device composed of a translating center shuttle connected by slender legs on both sides to bond pads anchored to the silicon wafer. The legs are fabricated in a slight chevron shape to bias them to move in the desired direction.

The thermomechanical actuator works under electric current. When a voltage is applied across the bond pads, it causes an electric current to flow through the thin legs. Because the legs have a small cross-sectional area, they have a high electrical resistance, which causes them to heat up and expand as current passes through them. The legs are constrained, so as a result their expansion is accommodated by bending, which moves the center shuttle forward.

Actuators can be designed for various output forces by choosing the number of legs needed to produce the desired force. Bent-beam electrothermal actuators developed at the University of Wisconsin use a similar concept of geometric constraints and have fewer legs that achieve larger displacements but a lower output force.

The Brigham Young University team fabricated a polycrystalline silicon TIM with legs 250 micrometers long. The legs have a cross section of 3 by 3.5 micro-meters. A steady-state input of 3 volts and 9 milliamps in vacuum (or 7 volts and 26 milliamps in air) results in an output deflection of 8 micrometers and an output force of 250 micronewtons.

A scanning electron micrograph shows a thermo-mechanical in-plane microactuator developed with funding by the National Science Foundation.

This represents an output force that is orders of magnitude larger than is commonly achieved with larger, electrostatic actuators. Because the device is so small, it heats up and cools down much faster than intuition would predict. For example, the polycrystalline silicon TIM made by the authors has been fully actuated in 400 microseconds.

Meanwhile, Michael Baker, a MEMS researcher at Sandia National Laboratories, recently achieved a TIM displacement of 12 micrometers using an input pulse duration of only 50 microseconds.

More effective actuators are one of the keys to opening a wider range of commercial applications for MEMS. For instance, TIM-actuated switches have the potential to reduce size, weight, power consumption, and cost in applications such as telecommunication switching in base stations and hand-held devices.

The actuation of microswitches and microrelays was the chief reason that the National Science Foundation funded the research leading to their development.

Larry Howell is the author of Compliant Mechanisms, published by John Wiley & Sons, and is the chair of the Department of Mechanical Engineering at Brigham Young University in Provo, Utah. Tim McLain is an associate professor at the university.