By Saren Johnston
Photonic crystals don't occur naturally. If you want one, you have to make it -- not an altogether easy job. But scientists are aggressively pursuing this task, manipulating photons to control electromagnetic radiation for potential applications that include optical switches, micro-fabricated lasers, waveguides, light-emitting diodes and telecommunications.
Building a photonic crystal requires creating a periodic structure from dielectric material -- material that is an electrical insulator or in which an electric field can be perpetuated with a minimum loss in power. Scientists must arrange the photonic material in a lattice structure that repeats itself identically and at regular intervals. If the assemblage is precisely made, the resulting crystal can have a photonic bandgap, a range of forbidden frequencies within which a specific wavelength is blocked, and light is reflected.
With the exception of the bandgap, a properly constructed photonic crystal will transmit wavelengths of light up and down the electromagnetic spectrum. Now scientists can predetermine the bandgap by engineering the lattice spacing of the photonic material to match the wavelengths they wish to block. This fortunate circumstance has given them the ability to control and manipulate light, sending it down assigned routes and around loops and bends.
A good deal of the credit for light control belongs to an inspired group of Ames Laboratory researchers led by senior physicist Kai-Ming Ho. Ho knows his way around photonic crystals. In 1990, he and Costas Soukoulis, senior physicist, and Che-Ting Chan, a former Ames Lab physicist now at the University of Science and Technology in Hong Kong, theoretically demonstrated the existence of the first photonic bandgap crystal.
"While exploring a number of structures, we discovered that the diamond structure had a bonafide three-dimensional photonic band gap," says Ho. "We were just doing the calculations, so the structure existed only on the computer at that time."
However, it wouldn't be long before the diamond lattice structure that Ho and his fellow theorists developed through computer simulations would become a tangible object. Their design led to the construction of the first successful photonic crystal, which was built later in 1990 by Eli Yablonovitch, a scientist at Bellcore Labs in Red Bank, N.J.
The crystal Yablonovitch fabricated from the Ames Lab design had a bandgap in the microwave region of the electromagnetic spectrum and was produced by drilling a series of carefully positioned holes in a dielectric material. However, drilling holes small enough and precisely enough to position the diamond lattice bandgap in the infrared or visible range of the spectrum where more potential applications exist turned out to be extremely difficult.
After seeing the problem that plagued Yablonovitch, Ho, Soukoulis and Chan started looking for a different way to construct the diamond lattice structure. Ho came up with a variation of the structure that could be built in a layer-by-layer fashion. "The idea was that instead of trying to drill holes, we'd stack alumina rods in alternating layers that ran perpendicular to one another, and we'd position the rods in every layer halfway between the rods two layers away," says Ho. The researchers received a patent for their novel design, which was fabricated in 1992 by former Ames Lab associate physicist Ekmel Ozbay and Gary Tuttle, an Iowa State University associate professor of electrical and computer engineering.
Between 1992 and 1994, experimentalists Ozbay and Tuttle built a number of three-dimensional diamond lattice bandgap structures to work in the microwave and millimeter-wave regions.
"Then Ekmel came up with the idea of making the structures smaller," says Ho. "Building the crystal with alumina rods takes quite a bit of time because you have to lay out all the rods and glue them together, stick by stick. So the idea was to make use of wafer technology. Ekmel and Gary would fabricate a whole layer at a time on a wafer and then stack it. That approach was quite successful."
Rana Biswas, a physicist and a member of the photonic crystal theoretical team with Ho, Soukoulis and associate scientist Mihail Sigalas, explains the push to make smaller structures. "We wanted to make these structures work at optical wavelengths, or at least work at infrared wavelengths so there would be applications for lasers and telecommunications.
"Since 1994, we have been working with Gary Tuttle at Iowa State University's Microelectronics Research Center and Ekmel Ozbay, now of Bilkent University in Turkey, to make the structures smaller and smaller," Biswas continues. "We had some success with structures made from metal strips deposited on plastic, but that still didn't solve the problem of getting the bandgap into the optical region."
When Biswas was giving a talk at a 1997 American Physical Society meeting, it was not by chance that Shawn Lin from Sandia National Laboratories was in the audience. Lin had been corresponding with Ho and knew that Biswas would be at the APS meeting. "After the session, Shawn came up to me and said he'd like to try to make the Ames structure in his lab at Sandia," says Biswas. "They have a state-of-the-art processing lab at Sandia, probably one of the leading facilities anywhere in the world."
In 1998, Lin's team built the Ames group's structure using bars made of silicon. But instead of leaving an inner air space, they filled the region inside the bars with silicon dioxide to give the structure more strength. Where other groups hadfailed with this "back-filling" procedure, Sandia succeeded because of their capability to planarize, or flatten, each successive layer of the stacked crystal so well that, as Ho describes the process, "adding another layer was just like starting anew with the initial flat substrate on which the first layer of the structure was readily built."
In the final step, the Sandia researchers immersed the whole structure in acid to etch away the silicon dioxide. The resulting three-dimensional diamond lattice structure had a bandgap of about 12 microns. "That's an important region because the carbon dioxide laser operates around 10 microns," Ho explains. "There are lots of interesting uses you can apply these structures to if you can hit the CO2 laser wavelength."
Later in 1998, Lin's group succeeded in fabricating an even smaller version of Ames Lab's layered lattice design, bringing the bandgap into the region of 1.35-1.95 microns in the near-infrared, a significant achievement because the wavelength used for transmission of telecommunications by optical fibers is 1.5 microns. Photonic crystals with bandgaps at that wavelength could be used to improve the efficiency of optical switches for fiber optic communications.
The combined efforts of Ames Lab and Sandia researchers in designing and fabricating three-dimensional photonic crystals with bandgaps in the near-infrared range of the electromagnetic spectrum was a major step made even more noteworthy because the fabrication method is economical and lends itself to mass production. However, the silicon photonic crystal presents a problem in that it has an indirect bandgap, which lowers spontaneous emission. But Lin thinks he can overcome this problem with certain doping measures.
Further testimony to the success of the Ames structure is the fact that Susumu Noda and Noritsugu Yamamoto, researchers at Kyoto University in Japan, have also used Ho's design to fabricate a layered structure of gallium arsenide rods with a bandgap layer in the infrared. "They use a different technique involving wafer bonding and ion etching, and now they have a sample with eight layers," says Ho.
While the Sandia researchers investigate ways to improve the emission rate for silicon photonic crystals, researchers on Ho's Ames Lab team are turning their attention from rods to spheres.
Kristen Constant, an Ames Lab associate and an ISU associate professor of materials science and engineering, came up with a novel ceramic technique for making photonic crystals from a nontoxic mixture of polystyrene microspheres and titania. She explains that polystyrene spheres can be purchased in dimensions less than 1 micron in size and can self-assemble into periodic, close-packed structures at optical wavelengths. "This makes them very attractive candidates for fabricating optical photonic crystals," she says.
Graduate student Ganesh Subramania knows just how good a mold the microspheres provide for fabricating photonic crystals. He's the person who makes the crystals, and he does it with surprisingly low-tech and inexpensive equipment.
Subramania begins the crystal-construction process by spreading a few drops of a slurry of titania suspension and polystyrene spheres on a glass substrate. The sample spends 24 hours drying in a humidity chamber and is then taken to be "squished," as Constant refers to its fate, in a cold isostatic press. Five minutes of compression helps thin the sample and reduces stress cracks in the heat-treatment process that follows.
Slow baking is the final step. Subramania heats the sample to 520 C (968 F) over five hours, during which time the mold of polystyrene spheres is burned off, leaving behind air spheres in a titania matrix. Shiny regions are visibly apparent in the remaining photonic crystals. These regions have characteristic colors that depend on the size of the polystyrene spheres used. The crystals made with smaller spheres (395 nanometers) have bright green regions, and those made with larger spheres (479 nanometers) are salmon-red in color.
"The color indicates that the structure that's present is what we're looking for -- basically, a periodic structure at a certain length scale," says Constant. "If we don't see color, either the scale is wrong or the structure is wrong. It's nice because we can visually inspect some of these to see if they have the right structure. Then Ganesh does scanning electron microscopy on the good samples to see the detailed, periodic structure."
Constant hopes they will eventually be able to make large-area optical photonic crystals using the economical and reproducible ceramic technique. Such an achievement would greatly enhance the efforts of scientists to find out more about the control of light emission and propagation in these materials. But there is still work to be done.
Biswas and Sigalas note that although they have now achieved optical photonic crystals with colloidal systems, the structures don't have a fully three-dimensional bandgap like the ones the Sandia team has produced using Ho's original stacked crystal design. On the other hand, the Sandia team is still trying to reach the optical region.
There's plenty of work to be done, but there's also plenty of work that's come out of the photonic crystal research to date. And Biswas says both the theorists and the experimentalists are involved in efforts to expand that work. "One of the next steps we have to focus on is finding new uses for all the research that is coming out," says Biswas. "We also want to partner with industry, and we've had some good feelers."
For more information:
Kai-Ming Ho, (515) 294-1960, kmh@ameslab.gov
Rana Biswas, (515) 294-6987, biswasr@ameslab.gov
Mihail Sigalas, (515) 294-7245, sigalas@ameslab.gov
Kristen Constant, (515) 294-3337, constant@iastate.edu
Current research funded by:
DOE Office of Basic Energy Sciences
Last revision: 12/17/99 sd
Home | Comments
| Search | Disclaimer