Skip to Content

Silicon Photonics Comes to Market

For the first time, a silicon-based optical cable will be commercially available.
August 16, 2007

Silicon photonics–using silicon chips to send and receive data-carrying light signals–promises to revolutionize telecommunications, but so far, it’s been largely confined to the lab. Now Luxtera, a startup based in Carlsbad, CA, that spun out of the California Institute of Technology, has announced the first optical cable based on the same silicon technology used to make microprocessors. The company says that the cable, called Blazar, can send 40 gigabits of data per second through its fiber but will cost as little as today’s 20-gigabit-per-second optical cables. Built using standard complementary-metal-oxide-semiconductor, or CMOS, processing, the cable is likely to find its first applications in data centers and computer clusters.

Silicon photonics: (1) The thin lines on the surface of this chip are waveguides, channels carved into silicon that direct light. (2) Five indium-phosphide-based lasers mounted to a die. (3) Luxtera researchers modified a standard piece of semiconductor test equipment so that it can test optical as well as electrical functions. (4) This device couples fiber to Luxtera’s silicon-based chip.

“This is the world’s first CMOS photonics product,” says Cary Gunn, Luxtera CTO. It’s the “culmination of eight years of development: six at Luxtera and, prior to that, two years at Caltech.”

Today, most photonics devices, such as lasers, modulators–the gadgets that encode data onto beams of light–and detectors are made of expensive nonsilicon materials such as indium phosphide and lithium niobate. It’s been the conventional wisdom that silicon, while great for electronic applications such as microprocessors and memory, is not well suited for generating, modifying, or detecting light. But silicon photonics is a rapidly advancing field, thanks to research at companies and universities such as Intel, IBM, the University of California, Santa Barbara, and Caltech. (See “Intel Speeds Up Silicon Photonics.”)

The customers that Luxtera is initially targeting are high-performance computer centers and data centers with racks of servers connected into systems that sprawl across giant rooms and warehouses. These centers make up the backbone of the modern information infrastructure and are used by Internet service providers, the government, the finance industry, researchers, and utility companies. And in the vast majority of them, equipment is connected with copper cabling, which has limited bandwidth that decreases with distance. For instance, thick copper cabling can provide only about 10 gigabits of data per second over one and a half meters.

Because copper, while relatively inexpensive, is slow and constrains the physical arrangement of complex computing systems, engineers have been turning instead to fiber-optic cables, which provide much higher bandwidth and can stretch distances of more than 300 feet without losing data-carrying capacity. Optical cables are also thinner and much more flexible than copper, so they help keep data centers cool: thick copper cables tend to trap heat and block the cooling air from fans. However, since a standard optical cable’s transceivers–the chips at its ends that send and receive data–are made of expensive semiconductors, the cable itself can get pricey.

Luxtera hopes to make optical cables more affordable by using silicon in parts of the transceiver. Gunn explains that each transmitter (one half of the transceiver) in the new cable contains a relatively inexpensive, off-the-shelf indium-phosphide laser. The light from this laser is split into four beams, and each beam passes through a modulator made out of silicon. An electric signal, Gunn says, is sent to each modulator to “imprint the signal directly on the light wave” at the rate of 10 gigabits of data per second. Then another silicon photonics device, called a holographic lens, launches the light into low-cost fiber. The holographic lens is lithographically etched into the surface of the chip and replaces the array of lenses found in a standard optical cable. After it’s transmitted down the fiber, the data-encoded light passes through another holographic lens and is directed onto an array of indium-phosphide-based photodetectors, which convert it back into an electrical signal. Electronics integrated into the same CMOS chip that houses the photonics devices amplify and clean up the signal and send it to an electrical receiver. Each end of a Luxtera Blazar cable has a chip that contains both a transmitter and a receiver, says Gunn.

The new product “represents the next logical step in utilizing silicon photonics for low-cost assembly,” says Alan Willner, professor of electrical-engineering systems at the University of Southern California. “The ability to house together various optical components makes for an even more compelling argument why 40-gigabit-per-second optical cables can beat out copper,” he says.

And the market for such devices is definitely good, says Fred Zieber, president of Pathfinder Research, a technology-analysis firm. “It’s a great product for things like server farms and computer complexes,” he says. “And that market’s growing pretty doggone fast.”

Luxtera may have to overcome the obstacle of being a startup and therefore not well known and trusted in the industry, Zieber says. But he adds that the reliability of its product could give the company an advantage. “They’re not using arcane material and not putting a lot of chips together to do this. All of that should give you a more rugged package.”

Gunn is confident that the cable will find customers, and as the transceiver chips are manufactured in larger numbers, their cost could drop to the point that silicon-based optical cables can even compete with copper–something that nonsilicon optical cables will never do. As costs fall, Gunn notes, silicon photonics could move beyond simply providing connections between systems. “Looking further out, silicon photonics will address bottlenecks within systems,” he says. Home computers whose circuitry used silicon photonics instead of copper wiring, he adds, would be faster and keep power consumption to a minimum.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

How scientists traced a mysterious covid case back to six toilets

When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

It’s time to retire the term “user”

The proliferation of AI means we need a new word.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.