Traditionally, electrical engineers working on radar and microwave communications and optical engineers captivated by the promise of fiber optics and high-bandwidth applications have remained separate. Separate conferences. Separate spectrums. Separate devices.
Figure 1. Coplanar electrodes and nonlinear polymers combined with ion-exchange waveguides could lead to highly integrated devices for communications and millimeter wave sensing systems. Courtesy of David Mathine, University of Arizona.
That's changing, thanks to new millimeter-wave devices and development in gigahertz and terahertz photonics. For the past several years, phased-array radars have melded photonic technology with radar systems to produce highly sensitive defense devices in the military. Millimeter waves are long enough to penetrate fog and adverse weather, but short enough to provide good-resolution images of objects (or, in the military vernacular, targets).
Military development continues to be a big part of the picture. However, civilian communications using these technologies is grabbing some of the military's thunder. In fact, MCI already has telecommunication systems that use gigahertz technology.
As a result, more and more researchers are working on devices that bridge the gap between the optical and RF spectrums through novel sources, switches, modulators, and integrated circuits.
Meetings tell the tale
Jennifer Hwu, director of the University of Utah's Center for Excellence for Electronic Systems Technology, looks forward to the millimeter, gigahertz, and terahertz conferences planned for the 1999 SPIE Annual Meeting.
"Some people see it as RF controlling optics and others see it the other way around," Hwu said. "That's why this conference is so important. I see both and that's why I'm so excited about these conferences. It'll bring two camps of people who can help each other to the same area."
Although applications such as optically controlled phased-array radar have already found commercial success in the military and in medical imaging, this broad area of research remains a nascent area of development with a bright commercial future. Consequently, scientists continue to look at every aspect of design, from material physics of superconductors, polymers, and organic materials, to new optical source designs and, of course, versatile semiconductor configurations.
"At the systems level, we're talking about phased array and signal processing functions like optically controlled microwave devices," Hwu said of millimeter, gigahertz, and terahertz development. "At the next level, we're talking about optical distribution and control signals. At the components level, we're seeing development of mixers, detectors, modulators, filters, couplers, multiplexers, and antennae -- an incredibly large number of elements. From there, we come down to a physics understanding of the components and materials."
According to Hwu, optical and RF signals are a great match, regardless of which way one uses the technology. "When you use optical signals to control RF functions you have the great bandwidth of the optical signal to play with," she said. "In the other direction, you have the speed and mature understanding of RF technology."
Phased-array radars are the standard example of an RF system that uses optical bandwidth to control an RF signal. These systems, comprised of many individual antennas, quickly shift the phase or frequency of RF signals. The greater the bandwidth of the incoming signal and the capacities of the modulation -- and assuming adequate antenna design -- the more flexible and secure the radar system.
New sources seek longer wavelengths
As with any optical system, optically controlled RF signal systems have to start with a light source. For applications where optical signals react with semiconductor or superconductor materials, existing wavelengths are adequate. However, a large part of the electromagnetic spectrum between the far-infrared and microwave ranges remains untouched. Lead-salt semiconductor lasers have demonstrated operation out to 30 µm with limited performance. Lead-salt lasers, however, are unreliable and short-lived devices inadequate for commercial use, although they have found niches in R&D.
Federico Capasso at Lucent Technologies/Bell Laboratories (Murry Hill, NJ) hopes to blaze a trail to the unutilized part of the electromagnetic spectrum through extended wavelength performance of his quantum cascade lasers. These lasers owe their performance to intraband transitions, as opposed to bandgap transitions, between semiconductor layers, and to the fact that each electron injected above laser threshold creates as many photons as the number of cascaded stages (typically 25 to 75). This is responsible for the high power of these devices.
Alessandro Tredicucci from Capasso's group recently demonstrated a quantum cascade laser with ~12 mW emission at 17 µm and operating up to 160 K in pulsed mode. Capasso said he expects that one should be able to extend the performance of quantum cascade devices far beyond these wavelengths into the terahertz region. (Broadly speaking, 1 to 10 terahertz translates to submillimeter optical emission between 30 and 300 µm).
According to Capasso, the highest frequency solid state oscillators are resonant tunneling diodes that have operated near 1 terahertz. However, these devices suffer from low power unless they are combined in an array. Higher power commercial solid state electron devices include Impatt diodes that can emit frequencies up to several hundred gigahertz.
"The area between 30 and 300 µm is particularly exciting on several accounts," Capasso said. "One is that this area lies between electronic and photonic devices. There's a whole area of the spectrum where there are very few electronic and photonic devices."
This year, Capasso's group expects to produce 20- to 25-µm emissions from quantum cascade devices with improved waveguide designs. After 25 µm, the quantum cascade laser will have to jump to 60 to 80 µm because of physics limitations, Capasso said. His collaborators, which include Mike Wanke, Alessandro Tredicucci, Claire Gmachl, Al Cho, and Debbie Sivco, are currently working on several strategies, including internal gratings and other waveguide designs to penetrate farther into the far-IR spectrum.
Although optical losses increase approximately by the square of the wavelength, Capasso does not see any inherent long-wavelength limitation to the quantum cascade laser, at least up to 100 µm.
"I am confident it can be done," Capasso said, "thanks to the superb aluminum gallium indium arsenide and gallium indium arsenide quantum cascade laser material quality grown by Molecular Beam Epitaxy. The atomically abrupt interfaces and the exquisite control of layer thickness made possible by this technique are central to our strategy to achieve far-infrared semiconductor lasers."
Polymers lead to fast switches
Quick switching and modulation of an optical signal is crucial to many applications, including millimeter wave devices and next-generation telecommunication systems. Several researchers at the Univ. of Arizona's Center of Optical Sciences (Tucson, AZ) including Nasser Peyghambarian, Seth Marder, Alex Jen, Seppo Honkanen, Bernard Kippelen, and David Mathine are developing hybrid glass-polymeric switches that operate above 100 GHz.
The group of Harald Fetterman (UCLA), Larry Dalton, and William Steir (USC) has already demonstrated switching speeds in excess of 113 GHz in an all-polymeric switch. The Arizona group is developing a hybrid polymer-glass horizontal coplanar device that could offer significant advantages toward integration with fiber optics, splitters, amplifiers, and other components.
In design, a bottom glass substrate contains an ion-exchanged waveguide that is topped with a layer of chromophore-doped polymer. A gold electrode tops the device, providing a bridge between the electric and photonic signals. Although ion-exchange waveguides are already in commercial use with lower modulation characteristics, Mathine said special chromophores being developed at the Univ. of Arizona are responsible for the higher frequency modulation while the coplanar design should make the commercial transition more attractive. Today, each of the three major components have been tested in pairs with excellent performance. The final step will be to integrate all three layers, said Mathine (Figure 1).
Semiconductor beam formers
In optically controlled phased-array radar, conventional approaches require the RF signal to be modulated into the optical carrier, and then demodulated from an optical signal back to an electrical one before RF transmission. Waveband Corporation (Torrance, CA) is taking a different approach that promises to save considerable expense by directly forming the millimeter wave beam using either optically or electrically induced plasma holograms in semiconductor material.
Figure 2. Plasma generation in semiconductor materials for controlling millimeter wave steering and switching can be completed either through optical excitation or through electrical injection. Although Waveband Corporation (Torrance, CA) has demonstrated both, company officials say electrode injection can provide more compact packaging and is entirely compatible with IC technology.
According to Waveband Director of Research Vladimir Manasson, their device excites a plasma hologram in a silicon planar waveguide, either through carrier injection from up to 1000 separate electrodes or through excitation by photons delivered by optical fibers. The localized excitation creates plasma spots, or abundances of free electrons. These spots form a holographic grating in the waveguide. When the millimeter wave travels through the planar waveguide, it interacts with the plasma-created hologram and is modulated, shaped, or directed based on the holographic pattern.
Because of short plasma lifetime, Manasson said the hologram can be quickly rewritten and, therefore, it can quickly change the direction of millimeter wave beams.
Waveband Corp. Vice President and CTO Lev Sadovnik expects these kinds of low-cost devices could greatly open up the application base for millimeter wave radars. Potential applications include automobile sensing devices for guidance and collision avoidance, he said.
Modeling the future
These examples of innovative millimeter wave development in gigahertz and terahertz photonics are only a scant few in an area undergoing intense development. "We see a lot of applications today that can employ this kind of technology," Hwu said. "In communications, there are so many aspects; so many different things that need to be developed."
Perhaps the most important tool needed does not lie with any one device, but rather with the computing tools needed to model these extremely high-performance devices and systems.
"One area that really needs to be improved is simulation modeling, because most of the tools for gigahertz device simulation are very time and resource consuming." Hwu said. "When you move up to the terahertz or even higher, the computer tools cannot keep up with the frequency simulations. I think we understand the basic physics, but teraflop computers may be necessary so that we can slice our frequencies and model and develop these devices.
"We almost need to build an optical computer to meet these needs," Hwu said with a smile.
375 Van Ness Ave.
Torrance, CA 90501
(1) 310/212 7808
Bell Labs Innovations
600 Mountain Ave.
Murry Hill, NJ 07974
(1) 800/458 -2368
R. Winn Hardin
R. Winn Hardin is a science and technology writer based in Jacksonville, FL.