Plasmonic Imager Could Slim Down Spy Satellites

10 April 2008--The U.S. Department of Defense hopes that a new type of imaging system will improve its spy satellites by allowing them to see in both the visible and infrared spectra using one simplified camera system. The DOD’s National Reconnaissance Office, responsible for the nation's reconnaissance satellites, has given HRL Laboratories, of Malibu, Calif., a research award for an undisclosed sum to explore a concept known as plasmonic imaging. The idea is based on how light interacts with structures made of metal and dielectrics.

”This is just a concept, and it has not been proven yet,” says Keyvan Sayyah, a senior research scientist in HRL's Applied Electromagnetics Laboratory, who came up with the idea and is leading the study. HRL is a corporate research lab owned by Boeing and General Motors.

Photons, which are electromagnetic waves, have an oscillating electrical field. When photons pass through a dielectric material, including air or a vacuum, and strike a metal surface, that field interacts with free electrons in the metal, causing their density to resonate at the same frequency as the incoming photon's wave. Those oscillations are known as surface plasmons.

Sayyah says that he may be able to use this plasmonic resonance to create images. He won't go into great detail because HRL is in the process of submitting a patent application for the technology. He says that each of the device's pixels is a detector made up of metals and dielectrics (arranged in a structure he won't describe) and is designed to have a plasmonic resonance absorption band, where the frequencies of the plasmons match a band of photon frequencies. When the structure absorbs a photon, it amplifies the photon's electrical field by many orders of magnitude, he says.

The pixels are divided into multiple subpixels, each covering a different band of wavelengths so that the device can perform multispectral imaging from the visible range--between 400 and 700 nanometers--to the long-wave infrared, about 8 to 12 micrometers. The absorption frequency of each pixel can be tuned within a certain range. ”What that range exactly is, we don't know ourselves at the moment,” Sayyah says. The pixels are at most one-tenth of the size of the wavelength, so the imager's resolution should be high.

If semiconductor-based imagers are not constantly cooled, their heat can create so much background noise that the incoming infrared light gets lost. The metal-based plasmonic imager shouldn't have that problem, eliminating the need for electrical power to produce cooling. And unlike compound semiconductor-based imagers, which are generally rigid, the metals and dielectrics could be deposited on, say, a curved sheet of plastic, creating a curved focal plane that would require much simpler optics and therefore lighten the satellite's load.

Though it is known that light can excite surface plasmons, several experts say they don't know of any method for using the effect to capture an image. Still, there has been some related work. In near-field imaging for microscopy, researchers use laser light to create plasmons with far shorter wavelengths than the laser, thus boosting the microscope's resolution. And Thomas Ebbesen of Louis Pasteur University, in Strasbourg, France, recently described a system of nanostructures that channels the plasmonic resonance and uses it to efficiently sort polarization features and spectral features from an image. The surface plasmon effect has also been proposed as a method of creating waveguides for on-chip optics and as the key to a single photon transistor.

William Barnes, professor of photonics at the University of Exeter, in England, studies surface plasmons in different metallic structures. He says he is not aware of anyone who has yet shown a way to do imaging based on the effects but that it's possible. ”It's really hard to know whether we'll have some sort of plasmon-based imaging system that gives us something we haven't had before, but I wouldn't rule it out,” Barnes says.

The HRL project, which should take 10 to 12 months, will consist of detailed modeling and simulations, and perhaps some initial experimentation, to demonstrate whether the concept is even feasible. ”We have confidence in it, but I can't promise it will work, because that's the nature of this high risk, high pay-off program,” Sayyah says. ”However, if it works, a lot of people will be knocking on our door.”