22 December 2004 -- It's not often that the simplest fix turns out to be the best. But engineers hunting for an efficient way to transmit terahertz waves along a narrow path have discovered that a simple metal wire does the trick. "It's really a very beautiful and unexpected result," says Michael Pepper, professor of physics at the University of Cambridge, in England, and cofounder of TeraView Ltd., a Cambridge, England start-up that is commercializing terahertz technologies. The discovery ends a search spanning many years for a waveguide that can carry radiation from the part of the spectrum lying between the deep infrared and microwaves, a region scientists have only recently learned to exploit.
Terahertz waves have generated huge interest in recent years because they are associated with the vibrations of whole molecules. Every large molecule has a unique terahertz signature. So in theory, reflected terahertz waves can be used not only to image an object but also to determine its chemical makeup.
What's more, terahertz frequencies pass easily through many common clothing and packaging materials. Many experts, therefore, hope that terahertz waves can be used to detect hidden weapons, illegal drugs, and explosives. The potential of these kinds of applications is driving intense research efforts to overcome the many hurdles that prevent the widespread use of terahertz waves.
One of these hurdles has been finding a way to channel the waves accurately. "The historical approach has been to try waveguides that work at other frequencies," says Daniel Mittleman an associate professor at Rice University in Houston. At optical frequencies, for example, a simple optical fiber does the trick, while at microwave frequencies coaxial cables work well. But neither approach is efficient at terahertz frequencies.
Nobody had even thought of using a simple wire. But then in early 2004, Mittleman and his colleague Kanglin Wang were attempting to create terahertz images using the tip of a wire to build up an image pixel by pixel. The pair expected only the tip of the wire to play any role in the experiment but soon discovered that the wire somehow carried the waves along its entire length.
This came as something of a surprise, not least because terahertz waves cannot pass through metal. "If you had asked me whether a wire would work, I would have said no," says Pepper.
The key to why a wire works turns out to be related to the way waves influence electrons on the surface of a metal. Terahertz radiation causes electrons to move back and forth on the metal surface, setting up a kind of wave known as a surface plasmon polariton. This becomes coupled to the terahertz wave's electromagnetic field and forces the wave to hug the surface as the polariton travels up the wire. "The wave becomes trapped," says Mittleman.
The wire, which works even when bent, turned out to be a better waveguide than anything he could have imagined. The signal attenuation as it travels along the wire is small and the frequency dispersion, which results from different frequencies propagating at different speeds, is "essentially too small to measure," says Mittleman. Minimizing frequency dispersion is important because one of the most common ways of producing terahertz waves is in square wave pulses that are made up of many frequencies. "If these were to disperse along our wire, it wouldn't be much use," admits Mittleman.
Still, there are a few teething problems. The trickiest is in finding a better way to couple the wave to the wire in the first place, say Wang. At the moment, less than 1 percent of the terahertz energy reaching the wire actually becomes trapped in the waveguide. This is because terahertz sources generate linearly polarized waves, like those seen when light is filtered through polarizing sunglasses, while the wire can carry only radially polarized waves, where the waves vibrate in many directions all radiating from the wire. Mittleman says that a new terahertz source he and Wang have designed that produces radially polarized waves will increase the conversion efficiency by two orders of magnitude.
The team has already designed a terahertz endoscope, which could become an important tool for medical imaging. There is evidence that terahertz images can be used to identify tumors. But because the waves can pass through no more than a few millimeters of skin, the technique works only for skin cancers. A terahertz endoscope would allow imaging to be done inside the body as well as outside. "It could also work as a chemical dipstick that identifies trace gases," says Mittleman, perhaps for examining hard-to-reach areas where explosives or drugs might be hidden.
Practical difficulties remain. Pepper, whose company TeraView is one of the leaders in terahertz technology, says that a wire would almost certainly need some kind of covering before it could be used as an endoscope, and that could prevent it working as a waveguide.
Mittleman isn't phased however. He's confident that waveguides will become an important component of the emerging field of terahertz science. "It's too early to tell how well some of these applications will work with waveguides. But it's certain they won't work without them.