Lasers with indirect-band-gap technology could ease integration with computer chips and therefore help optical computing, researchers say.
Stephen Shanklandprincipal writer
Stephen Shankland has been a reporter at CNET since 1998 and writes about processors, digital photography, AI, quantum computing, computer science, materials science, supercomputers, drones, browsers, 3D printing, USB, and new computing technology in general. He has a soft spot in his heart for standards groups and I/O interfaces. His first big scoop was about radioactive cat poop.
Expertiseprocessors, semiconductors, web browsers, quantum computing, supercomputers, AI, 3D printing, drones, computer science, physics, programming, materials science, USB, UWB, Android, digital photography, scienceCredentials
I've been covering the technology industry for 24 years and was a science writer for five years before that. I've got deep expertise in microprocessors, digital photography, computer hardware and software, internet standards, web technology, and other dee
MIT has demonstrated a laser that's built from germanium and that works at room temperature, a move the university said could be useful for high-speed optical data pathways within computers.
Lasers today are widely used to transmit large amounts of data over long distances, but the technology isn't economical for short-haul trips. However, many researchers are investigating ways to integrate lasers directly with conventional computer chips in an effort to reduce those costs and make high-speed communications more widespread.
Today's lasers typically are made from gallium arsenide and other expensive materials that have to be attached to computing chips after each component has been separately manufactured. In a paper to be published in Optics Letters, the Massachusetts Institute of Technology researchers said Thursday they made the germanium lasers work using a technology called indirect-band-gap semiconductors that other researchers thought wouldn't work. And the germanium technology is more easily integrated during manufacturing with today's chips.
In the current state of affairs, lasers use direct-band-gap technology, in which an energetic electron can move in a semiconductor crystal area called the conducting band. It releases the energy in the form of a photon as it falls back out of the band.
In the indirect-band-gap method, such an electron would squander its energy by releasing it merely as heat rather than light; to produce light, the electron would have to be excited to a higher energy level. The researchers got around this problem by "doping" the crystalline structure with phosphorus atoms, which have an extra fifth electron available compared to germanium's four.
Specifically, the lower energy state in the band is occupied by phosphorus atoms' extra electrons, so the excited germanium electrons can go to the higher level needed to produce a photon.
However, the researchers said they need to improve the technique's efficiency by increasing the amount of phosphorus doping so that less energy is wasted in powering the laser. They also gave no indication about how expensive the technology is to manufacture at scale--no great surprise given the early research stage of the work.
Lasers have been in use for decades in networking, but they're only now coming into reach for computing applications as well. Although lasers have progressed, so have conventional electronics methods--pumping electrons through metal wires.
But moving electrons pose problems. One big one is waste heat, which is expensive to manage within a computer chassis and in data centers packed with computers. Another problem is electromagnetic emissions that can cause problems by interfering with other electronics in a computer or with radio devices such as wireless networks. The faster the data transfer rates in wires become, the more these problems rear their heads.
That's why companies such as Intel are working hard on silicon photonics, in which data is exchanged by light rather than electrons. And the technology is nearing use: Intel's Light Peak optical cabling technology will be ready for use this year. Light Peak is designed to replace today's profusion of high-speed computer ports for video, networking, and countless USB devices with a single optical link.
The next step is shorter-haul data links within computers--from one chip to another. Ultimately, it's possible that photonics could power the actual computations on a chip.
MIT's primary germanium laser investigator was Jurgen Michel. Jifeng Liu was lead author of the paper, and Lionel Kimerling, Xiaochen Sun, and Rodolfo Camacho-Aguilera are coauthors.