Researchers at Cornell University have developed a simple silicon device for speeding up optical data. The device incorporates a silicon chip called a “time lens,” lengths of optical fiber, and a laser. It splits up a data stream encoded at 10 gigabits per second, puts it back together, and outputs the same data at 270 gigabits per second. Speeding up optical data transmission usually requires a lot of energy and bulky, expensive optics. The new system is energy efficient and is integrated on a compact silicon chip. It could be used to move vast quantities of data at fast speeds over the Internet or on optical chips inside computers.
Most of today’s telecommunications data is encoded at a rate of 10 gigabits per second. As engineers have tried to expand to greater bandwidths, they’ve come up against a problem. “As you get to very high data rates, there are no easy ways of encoding the data,” says Alexander Gaeta, professor of applied and engineering physics at Cornell University, who developed the silicon device with Michal Lipson, associate professor of electrical and computer engineering. Their work is described online in the journal Nature Photonics.
The new device could also be a critical step in the development of practical optical chips. As electronics speed up, “power consumption is becoming a more constraining issue, especially at the chip level,” says Keren Bergman, professor of electrical engineering at Columbia University, who was not involved with the research. “You can’t have your laptop run faster without it getting hotter” and consuming more energy, says Bergman. Electronics have an upper limit of about 100 gigahertz. Optical chips could make computers run faster without generating waste heat, but because of the nature of light–photons don’t like to interact–it takes a lot of energy to create speedy optical signals.
The new ultrafast modulator gets around this problem because it can compress data encoded with conventional equipment to ultrahigh speeds. The Cornell device is called a “time telescope.” While an ordinary lens changes the spatial form of a light wave, a time lens stretches it out or compresses it over time. Brian Kolner, now a professor of applied science and electrical and computer engineering at the University of California, Davis, laid the theoretical groundwork for the time lens in 1988 while working at Hewlett-Packard. He made one in the early 1990s, but it required an expensive crystal modulator that took a lot of energy. The Cornell work, Kolner says, is “a sensible engineering step forward to reduce the proofs of principle to a useful practice.”
Here’s how the Cornell system works. First, a signal is encoded on laser light using a conventional modulator. The light signal is then coupled into the Cornell chip through an optical-fiber coil, which carries it onto a nanoscale-patterned silicon waveguide. Just as a guitar chord is made up of notes from different strings, the signal is made up of different frequencies of light. While on the chip, the signal interacts with light from a laser, causing it to split into these component frequencies. The light travels through another length of cable onto another nanoscale-patterned silicon waveguide, where it interacts with light from the same laser. In the process, the signal is put back together, but with its phase altered. It then leaves the chip by means of another length of optical fiber, at a rate of 270 gigabits per second.
The physics are complex, but the net effect, says Bergman, is to “take a stream of bits that are kind of slow and make them go much faster.” The time telescope transmits more data in less time, and does so in an energy-efficient manner, because the only power required is that needed to run the laser.
The Cornell device is one of a series of recent breakthroughs in silicon photonics. “Silicon is this amazing electronic material, and for a long time it was viewed as being a so-so optical material,” says Gaeta. Over the past five years, researchers have been overturning this notion. In 2005, researchers at Intel made the first silicon laser; subsequently, other optical components, including modulators–devices for encoding information on light waves–have been made from the material. “People keep saying you have to replace silicon to do very high-speed processing, but silicon may be the way to go,” says Gaeta.
Sticking with silicon has two advantages. First, manufacturers already have the infrastructure for making devices out of silicon. “You can leverage all the technologies that have been developed for electronics to make optical devices,” says Gaeta. And if electronics and optics can be made out of the same material, it could be much easier to integrate them on the same chip and have each do what it does best: processing in the case of electronics, ultrafast data transmission in the case of optics.
DeepMind’s cofounder: Generative AI is just a phase. What’s next is interactive AI.
“This is a profound moment in the history of technology,” says Mustafa Suleyman.
What to know about this autumn’s covid vaccines
New variants will pose a challenge, but early signs suggest the shots will still boost antibody responses.
Human-plus-AI solutions mitigate security threats
With the right human oversight, emerging technologies like artificial intelligence can help keep business and customer data secure
Next slide, please: A brief history of the corporate presentation
From million-dollar slide shows to Steve Jobs’s introduction of the iPhone, a bit of show business never hurt plain old business.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.