Seeking new innovations in optical MEMS technology
By ROBERT PEASE
While few can predict the overall effect of a sluggish economy on the optical telecommunications industry, it can be stated with some certainty that technology rarely adopts a "wait and see" attitude. In fact, telecom research laboratories may actually be finding more time to go "back to the drawing board" in developing new innovations and testing radical ideas that may improve the way we communicate.
Micro-electromechanical systems (MEMS) technology is one of today's hotbeds for research and discovery, particularly in optical networking. Already, these devices are being lauded for offering such benefits as excellent optical performance, low insertion loss, wavelength flatness, minimal crosstalk, and the promise of improved integration, scalability, and reliability in optical networks.
With the success of MEMS technology in these and other areas, it's no wonder new applications are continually surfacing. According to the report, "The Little Chips That Could: A MEMS Industry Overview and Forecast," published by Cahners In-Stat Group, "MEMS technology is about to shrug off the flash-in-the-pan mantle some have labeled it and emerge as a bona fide industry." As indicated by research reported at such events as the Conference on Lasers and Electro-optics (CLEO), MEMS is gaining momentum as scientists find new optical uses for these tiny machines-above and beyond their well-known popularity in all-optical-switching components.
Researchers at the University of Minnesota, for example, have tested a new method for monitoring absorption characteristics of MEMS devices. Etch-released microstructures generally fare poorly when it comes to thermal conductance and the ability to handle high powers without thermal damage or warping. The problem is compounded in tunable structures where optical characteristics are dynamic because maximum acceptable intensity is dependent on actuation conditions.
"We have developed a method to use micromirrors as a probe of the average optical power they are reflecting," says Joseph Telghader, assistant professor for the Department of Electrical and Computer Engineering at the University of Michigan. "This technique could replace separate monitoring components, such as photodiodes, that add complexity to a telecommunications system."
Average power monitoring is used in many places in fiber-optic systems, says Telghader, and would be particularly useful in large optical crossconnects where it is necessary to know exactly where power is distributed inside the module.
This newest method uses the MEMS optical device as a thermal probe of its own absorption. The device design does not need to be changed. The absorption of a diffractive microbeam is dynamically measured during the actuation and light-source spectral changes.
In Figure 1, for example, a microbeam cavity filters light from an external source. The beam temperature changes with absorbed intensity in the same manner as a microbolometer in a focal plane array. If the temperature coefficient of resistance for the beam is known, the resistance change allows the extraction of the effective temperature. According to Telghader, excessive absorption can be prevented by feeding the temperature information to the light source-known as quenching the microbeam-until the excess signal has passed.
"I think that optical MEMS has its most important communication applications in dynamically reconfigurable networks," says Telghader. "We're already seeing MEMS devices used in signal routing, but I think there will also be significant MEMS penetration into 40-Gbit/sec long-haul markets, where there will be a need for tunable gain equalizers and dispersion compensators. Tunable lasers and filters could be impacted as well."
At the University of Central Florida's School of Optics, researchers reported a new design and first realization of an electrostatically actuated MEMS 1x2 switch based on planar waveguide SiON technology. A cantilever beam carries two vertically displaced output waveguides. A voltage applied between the substrate and the chromium electrode drives the beam tip down. As a result, the bottom (input) waveguide is aligned with the top (output) waveguide to reach the switched state.
The entire device is just 500 microns long, 400 microns wide, and 6 microns thick, with 4 microns of separation between the waveguides. The performance of this switch was tested using a continuous-wave semiconductor laser operating at 1.55 microns. On 15 tested samples, the voltage required to move the tip of the beam to the "switched on" state averaged 25 V, with 15 V needed for release voltage.
The school's researchers operated the switch for more than 300 million cycles with no functional degradation of the device. The device switch time never exceeded 100 microsec with driving voltages of 25 V. In fiber-optic networks, the device should allow for NxN switching scalability.
Meanwhile, at Stanford University's Solid State and Photonics Laboratory, scientists have found at least one way to more easily calculate the mechanical deformation of the moving top mirror for surface micromachined optoelectronics. They've also combined a model for optical diffraction loss with a mechanical model to provide a useful tool for future design of the high-finesse vertical-cavity devices.
"The industry can benefit from this modeling because it is easy to implement with a reasonable accuracy at the beginning stage of designing," says Chien-chung Lin, a research assistant at Stanford. "Later, if more detailed information is necessary, finite element analysis can be introduced."
According to the Stanford researchers, surface micromachining of deformable mirrors has been widely applied in wavelength-tunable optoelectronic de vices. The reconfigurable structure makes it very robust in WDM systems. However, changes in the reflective surface structure have not been characterized mathematically.
Additionally, it is essential to combine the mechanical model to determine the deformed shape of the mirror. At Stanford, scientists have constructed the first comprehensive optomechanical model to estimate losses and linewidth broadening of the cavity.
The work on designing more efficient MEMS devices is ongoing in universities and laboratories around the globe, and each new success is helping pave the way to lower-cost, more reliable optical communications of the future.
"I think MEMS technology is an enabling role in the future optical communication system because they are robust, agile, and, most important of all, reconfigurable for the system," says Stanford professor James Harris, Jr. "The feature of our optomechanical model can help us understand more in the performance of the passive devices, such as tunable filters and photodetectors."
MEMS technology continues to move forward with new innovations for optical communications applications. As re searchers continue to spin out new ideas for squeezing the most value from fiber-optic networks, MEMS technology has the potential to play a major role.