TCS Daily

Tiny Machines Have A Big Future In High-Tech

By James K. Glassman - March 19, 2001 12:00 AM

Picture a million mirrors, each gingerly balanced on a pivot. Now imagine putting them in a rectangle one-inch square. Welcome to the world of micro-electricalmechanical systems, or MEMS for short, that is taking the frontier of integrated circuits and microelectronics into the mechanical, optical and chemical fields. Texas Instruments some 15 years ago began the development of that array of mirrors. It provides the basis for super high quality video imaging and projection televisions and much more. Combined with fiber optic technology it can vastly improve the interoperability of that medium with cable and copper wire, easing the transition and lowering the cost to broadband communication. MEMS technology also is making wireless services more powerful. And in the area of medicine, it not only gives doctors three-dimensional visual images, but full on-site analysis of blood and DNA screens in one-thirtieth the time of five years ago. To explore this exciting, evolving field, TechCentralStation Host James K. Glassman interviewed a leading researcher in the field, Dr. Levant Degertekin of Georgia Institute of Technology (Georgia Tech), which has joined an elite list of schools and key businesses that are expanding the scope of this new technology.

James Glassman: Rather than bigger is better nowadays, smaller has become the rule. You are involved in the exciting new field of micro-electricalmechanical systems, or MEMS as it is called. Can you tell us briefly in layman's term what this is all about?

Levant Degertekin: Basically, this technology is like the silicon technology that has been used for microelectronics. We use the fabrication technology as in microelectronics, but for other functions than electrical, such as mechanical functions, chemical functions and optical functions. We then put them together in a kind of system, hopefully adding the microelectronics circuits, and integrate them on a single foundation, such as a silicon wafer or even a film.

Glassman: How do these developments compare with those in transistors, which kept getting smaller and smaller until they led to microchips?

Degertekin: It's not quite the same because transistors involve the basic physics of electrons. MEMS technology does not use that kind of physics, and those kinds of minute dimensions. The dimensions in regular MEMS devices are of an order of magnitude higher than the smallest transistors that you can make. And smaller is not always better in MEMS. Instead, MEMS borrows the fabrication processes.

Glassman: But these devices, as mechanical devises, are very small. In some cases, we're talking about very thin solid films containing these micro-electricalmechanical devises, right?

Degertekin: Yes. Micro-machining is done in two different ways. One way is what we call surface micro-machining. The machines, so to speak, are deposited and removed selectively, some on the surface of the foundation or substrate. Most of the time these substrates are silicon or integrated circuits. In the other case, we dig holes of different shapes to make larger devices. For example, we may put diaphragms on one side of the substrate and use that as a pressure sensor. And if you look at the integrated circuit (IC) industry, the devices are confined to say a couple of micros of that surface. But MEMS actually makes use of the whole wafer in a way. MEMS turns IC manufacturing labs into machine shops. That's the main difference.

Glassman: Tell us about some of what these machines can be used for.

Degertekin: The hottest topic right now is the optical switches for fiber optic networks and for wavelength multiplexing. Actually, that's where I think MEMS will make its biggest impact as an enabling technology. If you look at the devices right now that Lucent and that a company Nortel recently acquired to do those switching operations are making, they are simple in terms of processing. But used at the right place, they make their biggest impact. This technology becomes an enabling technology, so you can create many other devices. Basically, these are just movable very fast transmitters that you deflect. And the principle of the operation and the processing for these kinds of devices has been done for years and is very well known.

Glassman: So one thing you can use them for is optically to speed the activity on the microchips and semiconductors?

Degertekin: Yes. It's basically the kinds of switches that will connect different kinds of lines, particularly fiber optical lines to other kinds of lines in a matrix or network very fast.

Glassman: Let me ask you about a few specific things. You're working with micro-machine ultrasonic transducers. Now, what are those?

Degertekin: What we're doing is that we're using MEMS as an enabling technology in our research. So-called interactive electric materials have been used to generate and detect ultrasounds for medical imaging equipment. After 30 years of engineering, there are some manufacturing limitations using those interactive materials -- mostly ceramics used to form imaging arrays that you've basically got to cut using a saw. Now, if you look at all the medical imaging pictures that you see on the ultrasound pictures, you get a kind of sliced image of the human body, or the baby's body. It amounts to a one-dimensional picture or array, as we call it.

Glassman: And what does MEMS do to change that?

Degertekin: The trend is to get geometric images so you can see the face of your baby. If you want to get multi-dimensional arrays, there is no way that you can manufacture those using the cut and ceramic technology. Getting multidimensional adds huge complexities to the manufacture of these transistor arrays, and micro-machining enables that. We can make those multidimensional arrays or pictures by using photolithography, and we use micro-machining to make the transducers - devices that convey and can change power between systems -- that allow that. We are using capacitor transducers, and we're just making simple capacitors that move a membrane by applying an electrostatic field. So actually we're using a concept that's been known for hundreds of year for making receivers and for decades in making microphones. Only those capacitors were not efficient for transmitting energy. They became efficient only when the gaps between the electrons, or electrodes, are very thin - one micron or less. And at that point MEMS technology comes into the picture because we can make 0.1 micron thick gaps, and that enables us to transmit and receive energy very efficiently.

Glassman: What is that 0.1 micron? Is that like a few atoms?

Degertekin: Actually, 0.1 micron would be thousandth angstrom (a ten billionth of a meter), so if there were atoms their might be tens of a hundred of them. But the gap is a vacuum empty of everything.

Glassman: You mention energy. There's a lot of work being done on fuel cells as an energy source. Is MEMS technology involved in that?

Degertekin: Actually there's a big push right now by DARPA (Defense Advance Research Projects Agency) to make small micro-reactors. We're working right now on such a system. The idea is to generate 21 watts of power using the space of one or a couple of centimeters of cubic volume.

Glassman: So, is that promising?

Degertekin: It's very promising and they want to use liquid and not gas fuels for that so that the soldiers can use it. And later on this technology is going to run cell phones because the batteries are one of the biggest problems.

Glassman: Can any of these be used or to power automobiles?

Degertekin: I don't think you'll get that much power out of these micro reactors. But what may happen is in the future every house can have its own reactors.

Glassman: Each home may have its own energy source?

Degertekin: Exactly.

Glassman: And what kind of fuel is it converting?

Degertekin: The biggest part would be generating hydrogen from methanol, for example. You have to make a very small combustion chamber, so to speak, to generate hydrogen. But that would be possible, and it's the biggest advantage that MEMS is bringing to that technology. When you can make the channels of the chamber so much smaller, the surface-to-volume ratio goes up. So, instead of having just one large channel on a wafer or surface to transfer the volume, you can put in thousands of small channels to improve the surface area. And that increases the efficiency of those reactions, enabling higher-powered density.

Glassman: At Georgia Tech how is the micro-electricalmechanical systems department organized?

Degertekin: Actually the faculty members dealing with micro-machining are housed in mechanical engineering and also in electrical engineering.

Glassman: Where does it stand compared with other universities involved in this field?

Degertekin: I think Georgia Tech is trying to catch up. The University of California at Berkley has been in the forefront of MEMS work. MIT, Cornell and Stanford are also universities involved in the MEMS fabrication facility network. And those facilities are open to industry for everyone to come and do their processes. So that makes them more attractive. But Georgia Tech already has a microelectronics fabrication facility where we can do all the MEMS processes. And one important point to notice is that the MEMS processes so far have been kind of copying what IC manufacturers had been doing. But actually if you look at the research in the last five years, it turns out that you need some different processes to make MEMS devices, such as drilling holes in wafers. For regular IC manufacturers they would never do that. But you need those kinds of processes for MEMS. In addition, we're using different kinds of materials, other than silicon, such as glass and different polymers, especially for making biocompatible devices. Silicon, though, is itself a nice compatible material.

Glassman: Interesting. I'm afraid we're out of time. Thank you very much for doing this for us.

Degertekin: Thank you very much. It's a pleasure.


TCS Daily Archives