Nanophotonics researchers at Arizona State University have created the world’s first electrically powered room-temperature nanolasers. These lasers are the single most important step towards building computer chips that use light instead of electricity for ultra-fast and efficient on- and off-chip communications.
Producing electrically powered nanoscale lasers has proven to be one of the toughest tasks facing electronic and photonic engineers. Over the last few years we have created very small on-chip lasers, but they need to be powered by a large, off-chip laser. We have also produced electrically powered (i.e. self-contained) lasers, but they require very cold temperatures (sub-zero) to operate correctly. Now, after seven years of research, Cun-Zheng Ning and fellow researchers from ASU have finally cracked it.
ASU’s nanolaser consists of a rectangular, semiconductor sandwich of indium phosphide and indium gallium arsenide (InP/InGaAs/InP), and then insulated by silicon nitride and silver (forming the metallic cavity that forms the beam of coherent, laser light). The complete laser is a few micrometers (micron) in diameter. The research team had built electrically powered lasers using the same structure before, but the silver layer would overheat at room temperature. By adjusting the thickness of the SiN insulator, and refining the fabrication process, ASU’s nanolaser now operates surprisingly well at room temperature — comparable to the conventional semiconductor lasers that are used in optical routers and optical disc drives, in fact.
Moving forward, Ning and co need to improve their nanolaser’s longevity, perhaps by using quantum wells (smaller metal cavities that operate with lower, less-stressful threshold currents). After that, the team will have to work out how to integrate these nanolasers onto standard silicon chips; at the moment, the lasers are built on wafers of indium phosphide — a III-V semiconductor that isn’t completely compatible with current silicon CMOS processes. III-V semiconductors are fairly well understood, though, and are slated as being one of materials that might take over when silicon runs out of steam. We certainly won’t see nanolasers appear in computer chips overnight, but there’s a solid roadmap that should ensure that they don’t disappear into the research ether.
Once nanolasers do find their way into computer chips, we won’t really know what hit us. As IBM showed with the first chip that integrated optical and electrical components on the same silicon die, optical interconnects will allow for data transfer rates that are thousands of times faster than copper wires, while using considerably less energy per bit transmitted. Beyond being faster and more energy efficient, light can also travel much greater distances at higher fidelity over optic fibers than electrical signals over copper. Ultimately, the end goal of nanophotonics is to integrate optical components and interconnects into everyday devices — and once that happens, you’ll be able to link directly into your ISP’s fiber optic network, and enjoy terabyte-per-second transfer speeds to other devices around the house.
Research paper: http://dx.doi.org/10.1364/OE.21.004728 – “Record performance of electrical injection sub-wavelength metallic-cavity semiconductor lasers at room temperature”