Hwang wins top IBM award for MEMS-semiconductor integration

	Jim Hwang (left) and Ph.D. candidate Steven Peng in the Compound Semiconductor Technology Lab.

Jim Hwang, director of Lehigh’s Compound Semiconductor Technology Laboratory, has won one of the most prestigious awards granted by a major corporation to a university faculty member.

Hwang, a professor of electrical and computer engineering, recently received the 2007 IBM Faculty Award, which carries a gift of $40,000. The unrestricted cash prize is not part of a contract between Hwang and IBM, and is not tied to any work Hwang has done or will do for the company.

A former director of Lehigh’s Center for Optical Technologies, Hwang was cited by IBM for integrating microelectromechanical systems (MEMS) with semiconductor technology. MEMS are valued for their size, versatility and low cost, and have become a multi-billion-dollar industry with applications in automobiles, cell phones, cameras and many other areas.

In several projects in the past four years, Hwang and his students have earned renown for mitigating a problem called stiction, which occurs when electric charge is trapped in the dielectric, or insulating materials that are contained in semiconductors.

Hwang and three undergraduate students have helped understand a stiction-related problem in microshutters that NASA has developed for the James Webb Space Telescope, which will replace the Hubble Telescope. The tiny shutters will enable the new instrument to resolve images of galaxies at the edge of the universe.

Last year, Hwang took the lead when Lehigh joined with the University of Illinois at Urbana-Champaign (UIUC), Purdue University and the Georgia Institute of Technology to form the MEMS/NEMS Science and Technology Fundamental Research Center. (NEMS stands for nanoelectromechanical systems). The center, led by UIUC, will receive $10 million in funding from the Defense Advanced Research Projects Agency (DARPA) over the next six years.

The goal of the DARPA project is to gain a deeper understanding of dielectric charging and of the material and mechanical properties of NEMS. A longer-term goal is to integrate NEMS devices with CMOS circuits. CMOS (Complementary Metal Oxide Semiconductor) is the technology through which more than 99 percent of semiconductor chips are currently made. In addition to Hwang, Herman Nied, chair of mechanical engineering and mechanics, and Rick Vinci, associate professor of materials science and engineering, are participating in the DARPA project.

In another DARPA-funded project, Hwang and his colleagues at MEMtronics, a spinoff of Texas Instruments, have overcome two critical challenges to the use of MEMS in military applications. The group developed an improved method of packaging MEMS switches to protect them from environmental stresses. Working at Lehigh last year, the group became the first to operate an MEMS capacitive switch for more than 100 billion cycles in dry air. The Air Force Research Lab has since operated a switch delivered by the team in room air for more than 260 billion cycles.

Towards smart vehicles

In the most recent MEMS project, Hwang and Steven Peng, a Ph.D. candidate in electrical engineering, have mitigated a stiction problem that could open the way to the inexpensive application of millimeter-wave technology and the expanded use of radar for “smart” vehicles.

Peng completed much of the work on this project last summer during a research internship at IBM’s Thomas J. Watson Research Center in Yorktown Heights, N.Y. Hwang, who has six Ph.D. students, requires each to do an internship with a company.

IBM is pursuing wafer-scale integration of MEMS and CMOS technologies by combining hundreds or thousands of silicon chips, each measuring several millimeters, on a wafer spanning 8 to 12 inches in diameter. A major obstacle to this endeavor is a dielectric charging problem which Peng has reduced by using a superior dielectric material.

“Steven did not completely eliminate the problem of dielectric charging, but he has improved the situation by orders of magnitude by going to a better dielectric material,” says Hwang.

Millimeter wave, says Hwang, is an expensive technology used mostly by the military in radar, communications and imaging applications. It is also an enabling technology for automobile collision-avoidance radar and other applications related to smart vehicles and smart highways. Collision- avoidance radar, under the more risk-averse name of “adaptive cruise control,” is now available in the U.S. only in some luxury cars. However, Japan is planning to equip more than half its autos with the technology in the next five years.

Devices utilizing microwave (wavelengths longer than a centimeter) radar, says Hwang, now help guide airplanes and ships. The devices need to be compacted to fit on cars without affecting aerodynamics or aesthetics, and the best way to do that is with millimeter-wave (wavelengths shorter than a centimeter) radar. More compact devices mean smaller prices and superior sensing capability as well.

“If you advance from microwave to millimeter-wave radar,” says Hwang, “you can compact a device and achieve a more focused beam. The shorter the wavelength, the better the resolution.”

Because of high operating speed, millimeter-wave devices are typically powered by compound semiconductors. But the recent downsizing of silicon chips, says Hwang, gives them the speed needed to support millimeter-wave operation, and the short millimeter wavelength allows engineers to build an antenna on each chip. Hundreds of chips, however, are required to provide the devices with the necessary power, and this is best done by fabricating the required number of chips on a wafer and combining the individual millimeter-wave beams generated by each chip in free space.

“To realize the power at millimeter wavelengths, we need to combine many silicon chips,” says Hwang. “The only practical way to do this is on a wafer and through free space, not with wiring because losses are typically high for wired devices at millimeter-wave wavelengths.

“In essence, you have to build many miniature radar stations and combine them in free space. It is an expensive undertaking unless it is done on a silicon wafer using standard CMOS fabrication technology.”

Hwang and Peng are proposing to use MEMS technology to core out the silicon underneath the antenna, thus suspending the antenna in air and reducing loss. They also propose using MEMS technology to build other passive components, such as switches, inductors, tuners and waveguides, with low loss.

“Steven’s discovery,” says Hwang, “will help make millimeter wave technology orders of magnitude cheaper, and much more prevalent, than it is today.”

--Kurt Pfitzer