In the quest to make computer processors smaller and faster, computing with light instead of relatively slow electrons has long been a tantalizing goal. One roadblock has been the inability to make lasers tiny enough so that several thousand of them could fit easily on a chip. In late August, two groups of researchers reported the construction of a new kind of nanometer-scale laser. Surface plasmon resonance nanolasers, or spasers, are the smallest lasers yet made, and their creators say the devices could pave the way toward ultrafast optical computing.
Spasers can "bridge the worlds of electronics and optics at truly molecular-length scales," says mechanical engineering professor Xiang Zhang, of the University of California, Berkeley, who led one of the groups with research associate Rupert Oulton.
Surface plasmons are oscillations of electrons that form at the junction between an insulator and a metal. Metals have lots of free electrons, which oscillate when light shines on them. But when the metal borders a dielectric, the movement of the electrons is curtailed because they cannot enter the dielectric. That forces the electrons to move in waves of density along the junction.
Significantly, light will produce plasmons even in structures much smaller than the light's wavelength. It was this phenomenon that the two research groups exploited in creating their nanoscale lasers.
Mark Stockman of Georgia State University, in Atlanta, and David Bergman of Tel Aviv University first proposed the mechanism of a spaser in 2003. The idea was to build a device that operates similarly to a laser to generate and amplify surface plasmons. In a laser, light reflects back and forth through a special material called a gain medium, stimulating the emission of more light of the same phase with each pass. In a spaser, it is the waves of electrons that are amplified, which are then converted to light.
A team led by Mikhail Noginov of Norfolk State University, in Virginia, including researchers at Purdue University, in Indiana, and Cornell University, in New York, has demonstrated the spaser in gold nanoparticles surrounded by a silica shell that has been impregnated with green dye. When light hits the nanoparticles, plasmons form at the gold-silica interface. The plasmons are amplified by their interaction with the dye-filled shell, which acts as the gain medium, and laser light emerges. The nanoparticle system is just 44 nanometers in diameter, less than one-tenth the size of the 530-nm wavelength light emitted by it.
Noginov's new laser relies on organic dye molecules, which would make it difficult to incorporate into existing technology. Future work may involve creating a spaser-based nanolaser driven by an electrical source instead of a light source, which would make the lasers more practical for computer and electronics applications, says Noginov.
The UC Berkeley researchers led by Zhang and Oulton have created a spaser using a different approach, which they call hybrid plasmonics. "We have found a way to tightly confine light and at the same time sustain it for a long time in a very small space," says Oulton.
Their team's spaser consists of a cadmium sulfide semiconductor nanowire on top of a silver substrate. The nanowire and silver are separated by a tiny magnesium fluoride insulating layer that's 5 nm thick. A laser feeds light into the semiconductor nanowire, which acts like a waveguide and transmits the light's electromagnetic modes—harmonic frequencies of laser light that are not limited by diffraction. The plasmons that form on the silver substrate couple with the modes in the waveguide and generate laser light within the magnesium fluoride gap. Despite the gap's puny 5-nm dimensions—about the size of a single protein molecule—the wavelength of laser light emitted is 489 nm.
The spaser approach "looks wonderful," says Dick Slusher, a prominent laser physicist not involved with either team, who is currently the director of the Georgia Tech Quantum Institute.
Though it will be a long time before spasers are powering optical computer chips, there are many intermediate steps that could be useful. "We would like to use plasmon lasers to do something you cannot do with conventional lasers," such as analyzing individual molecules, says Oulton.