Welcome to Prof. Toshiyoshi's research group on MEMS at the Institute of Industrial Science (Komaba Campus), the University of Tokyo. In this page, we provide MEMS-beginners with an introduction to MEMS, including its history, concept, techniques and applications.
MEMS is an abbreviation of the technical term "Micro Electro Mechanical Systems" and it is pronounced with an accent at "E." It is usually spoken in the plural form*1. The concept of MEMS is a very small electromechanical systems, where those micron-scale "machines" are involved that are physically operated by means of electronics. In that sense, we sometimes call it "micromechatronics" when we refer to its academic discipline. We also sometimes call "micromachines" to mention the small machines made by the MEMS technologies. Micromachining is a term used to mention the fabrication technology.
The small texture seen behind the ant*2 (photo)*3 are such "MEMS" devices made by the semiconductor fabrication processes. Now, what is the difference between precision machines and MEMS? You may find it in the history of MEMS.
MEMS has been already diffused in our daily life. The photo on the left is the MEMS accelerometer produced by a US company Analog Devices*4. It carries a micromechanical proof-mass made of silicon, and it is suspended with tiny flexure hinges made of silicon, too. When an external shock is given to this chip, the mass stays where it is by its inertia mass, while the chip frame displaced a bit with respect to the mass. The relative displacement change is detected by means of the electrostatic capacitance change, and the acceleration is electronically estimated. There are plenty of different MEMS accelerometers, and they are mainly used in the airbag ignition system in the car. Recently, a MEMS accelerometer is also used in TV game controllers*5 and smartphones. Images on smartphone are always displayed in an appropriate orientation because the accelerometers inside sense the direction of the gravity.
The photo on the left is the DMD*6 from a US company Texas Instruments. A micromechanical mirror of 16 microns by 16 microns is arranged in an array of more than million copies on the LSI chip. Each mirror is electromechanically controlled by the signals generated by the LSI transistors, and the light reflection is individually controlled to create reflection image patterns. The mirrors are operated in a real time manner to make motion images. You may sometimes find image projectors in conference or meeting. They may look like an LCD (Liquid Crystal Display) projector but you can tell a DMD type by carefully looking at the screen from a close distance; a shape of mirror pixel is seen on it.
MEMS is such a high-tech and it looks so new. However, there were many trials in the filed of micro mechanical devices even before the word MEMS was born. In 1975, a US company Westinghouse created an array of microelectromechanical mirrors (50 microns) on an optically transparent substrate*7, and it was used as a projection display engine. No LSI *8 technology was available at that time, and thus the mirrors were individually controlled by using an electron beam traveling in the vacuum. The entire mirror chip was encapsulated in the vacuum, and the electrically charged-up mirror was tilted by the electrostatic force. It sounds like a hybrid of micromechatronics with vacuum electronics. They demonstrated a flight-information-like display.
The same group had also reported a resonant-gate transistor (RGT) in 1967*9. CMOS transistors today have a fixed gate electrode. On the other hand, they intentionally made the gate mechanical movable such that the threshold voltage could be tuned by the mechanical vibration of the resonating gate. The RGT was intended to be a physical sensor by means of semiconductor electronics.
#ref(): File not found: "feynman.png" at page "What is MEMS?/Introduction"
Dr. Richard Feynman, a Nobel Prize winner and a Professor of California Institute of Technology, is known as the first person who pointed out the possibility of what we call today the nanotechnology. In his lecture "There's plenty of room at the bottom" in 1959*10, he forecast several stories. For instance, he suggested that one would be able to have the entire volume of encyclopedia onto the head of a pin by the photo reduction technology. He also mentioned a use of electron beam and atom handling to read and write such data. The most famous suggestion he made was an electrical motor of 1/64 inch in diameter. He offered a prize of US$1000 for the first inventor.
In the following year of Professor Feyman's lecture, an electrical micromotor of 1/64 inch in diameter was presented by Mr. W. McLellan*11.
#ref(): File not found: "MEMSmotor.png" at page "ja"
A real "micro" motor had not been demonstrated until 1989 when Dr. Y. Tai et al. with University of California Berkeley made one by using the semiconductor micro fabrication processes*12. The cross-shaped structure in the middle of the photo is a micro rotor, and it was pinned down with a rotation shaft. Rotation torque was generated in between the twelve fixed stators around the motor, where voltages were applied in turn.
The micro motor of Dr. Tai was an epoch making device in a sense that the semiconductor process was applied to produce micro "mechanical" devices instead of micro electronics. His idea drew a lot of inspiration from the researchers. In Japan, Prof. H. Fujita with the Institute of Industrial Science, the University of Tokyo and his colleague with IBM Tokyo Research Labs proposed an electrical micro motor by using electroplated nickel in 199
02. The diameter of the motor was as small as 100 microns, which was equivalent to the size of a human hair (typically 80 microns). Thanks to the very small mass, the motor could reach maximum rotation speed of 10,000 rpm within a fraction of second.
The above micro motor is now featured in a science textbook for junior high school. "It was the very early stage of MEMS R&D at that time, and people came up with various ideas of micro actuators. Well, it was also difficult to answer to such questions like ... what is this for?" said Dr. Toshiki Hirano (currently with HGST, San Jose, CA) who designed and developed the micro motor. Almost a quarter-century has passed since then, and the MEMS technologies have diffused into various fields including silicon microphones and gravity sensors in smartphones, tire-pressure sensor (TPS) in cars, and fiber telecom applications. It is impossible to think of a life without MEMS devices today.
In no time, we will see new students in the university who learned science with this textbook. The lab faculty (Prof. H. Toshiyoshi) is teaching a 101 course on electromagnetism for the freshmen in Komaba campus of the University of Tokyo. He uses the micro motor as an example to explain the Coulomb's force and the Gauss' law.
Can you imagine how the micro motor was energized? It was driven electrostatically. Electrostatic force that we feel in our daily life is so small, and we do not pay attention to it until we feel a sharp electric shock when we touch the door knob after long walk on a carpet. You may have experience using a plastic sheet to swab on the head to make your hair spread out. Electrostatic force is so small in the macro scale but it becomes to be significant in the small world, because the electrostatic force is a surface force that is proportional to the square of dimension, while gravity is proportional to the power of dimension.
In a case of the micro motor, the rotor is gradually pulled toward one of the stator in turn where a voltage is applied. When the voltage hops to the next stator, the rotor follows it. The motion can be understood by an image of a compact disk (CD) wobbling around a finger. The rotor and the center pivot are in contact to each other but the rotor moves without friction, i.e. without energy loss. Surface friction is also a surface force, and it has an significant effect in the MEMS scale. So, friction is usually avoided to have smooth motion of MEMS actuators.
Let us now take a look at the characteristic fabrication process of MEMS. A watch movement is an good example to compare with MEMS*13. It is made up of more than 100 pieces of mechanical components. Each part is made of an appropriate material and is produced by an appropriate fabrication process. Most precision machines are manually assembled by engineers. As a natural consequence, it takes long time to finish one device, and only one product is made at a time. Of course, a lot of industrial efforts are paid to put it in the mass production phase.
On the other hand, most MEMS devices are free from the final assembly step. It is made by using the semiconductor thin film processes by repeating film deposition and patterning. Device materials are chosen out of the semiconductor-process compatible ones such as silicon (single crystalline, poly crystalline, and amorphous), metals, and silicon oxide. MEMS structures are supposed to have micromechancally movable parts that could be made by the "selective release" step. For instance, let us imagine a micromechanical structure of silicon sitting on a silicon dioxide layer. Silicon dioxide can be selectively etched in hydrofluoric acid without damaging the silicon structure, and thus such mechanical structures can stay on the substrate. For this process restriction (of sacrificial release), MEMS structures always have mechanical connection to the substrate called an "anchor."
Most MEMS devices look thin and floppy because they are made of thin films of the semiconductor processes. Nevertheless, three dimensional volumetric structures can be made by lifting such flat structure out of the substrate plane. Careful observation can spot a hinge structure at the bottom of the three dimensional microstructure*14). The inventor of the polysilicon micro hinge is Professor K. S. J. Pister with University of California Berkeley. The photo on the left is a MEMS optical bench created by Professor Ming C. Wu and his colleagues at University of California Los Angeles*15. A semiconductor laser chip was mounted on a silicon chip of 1 cm with some fixing plates, and Fresnel lenses and beam splitters were prepared on the same height. Those optical components were made to be pre-aligned on the mask level, so, one could greatly reduce the effort of optical alignment.
We picked up a rotating micro motor as an example of MEMS. Mechanically active element of MEMS is usually called an actuator. There are various types and mechanisms of micro actuators. The photos on the left include the rotating motor mechanism by means of electrostatic torque, a saloon-door like structure that are retractable inwards the substrate hole (also electrostatic), out-of-plane motion by means of piezoelectric strain, and lateral motion in parallel with the substrate surface. Other actuation principles are, for instance, thermal expansion through electric Joule heat and electromagnetic force.
Mechanical motion seems to be the most visible characteristic of MEMS. However, the significance of MEMS is NOT in the mechanical motion but it is explained by three M's, namely Miniaturization, Mass production, and Multifunctional integration, as pointed out by Professor Fujita with IIS, the University of Tokyo. Use of the semiconductor processes enable us to have a lot of copies of micro machines at a time unlike the conventional precision machining. One may also integrate multiple functions such as electronics, mechanics, optics and else onto a single chip of silicon by the MEMS processes. Due to the MEMS capabilities represented by these three M's, MEMS is expected to be the platform technology for the next-generation electronics industry.
The table on the left is from the MEMS market research in Japan that is periodically reported by the Micro Machine Center, Japan*16. MEMS technology has diffused into various field of engineering such as information technology, medicine, automobile, and biotechnology. Sensors for vehicle are particularly large in market.
Now that we have seen various applications of MEMS, we realize that MEMS application fields includes two major domains: micro actuators that convert electrical signal into mechanical output, such as micro mirror array for image display and ink-jet printer nozzles. The other application is micro sensors that convert mechanical (or chemical) input into electrical signals, such as accelerometers, gyroscopes, and silicon microphones.
Due to the small voltage range of sensor output (micro volts to millivolts), we usually use amplifiers to magnify the signal range. What if the signal is already large enough to electrically drive something else? It could be used as a new energy source or so-called micro energy harvesters, which has become the third application domain of MEMS technology today.
MEMS is a toolbox to create small devices. It has versatility to apply in various fields such as sensors, micro fluidics, micro optics, radio wave devices, micro power generator. MEMS also lies in various academic fields beyond the boundaries, where we enjoy the research in university. There are many MEMS-related conferences organized in electrical engineering, electronics, applied physics, chemistry and else. MEMS also has a comprehensive lateral conference and workshop, where most MEMS researcher on this planet get together. MEMS is a place that we need to establish our own academic discipline, and we really enjoy it.
In our lab we pursue the fundamental fabrication processes and design methodology of MEMS and pursue the application to micro optics (including fiber optic telecommunication, image display, and medical instrument) and radio frequency devices (RF-MEMS switches) in close collaboration with MEMS industry. You may find our recent activity in our webpage Research.
The following link summarizes journals for MEMS-related papers.