An extremely sensitive and accurate infrared detector, which can ferret out a misfiring transistor from the billions on today's Pentium and PowerPC chips, has been developed by researchers at the University of Rochester. The technology behind the device, outlined in the August 6 issue of Applied Physics Letters, began life in Russia as a way to see the heat radiated from cool stars in deep space, and it is now under non-exclusive license to a California company that tests processors.
The new device can detect single photons, making it sensitive enough to meet the demands of new chip makers who need to test billions of transistors on each chip as quickly and efficiently as possible. The key is to exploit a quirk of physics; when transistors switch, they sometimes emit a very brief flash of infrared light. This flash can reveal much about how the transistor is behaving-but only if a detector catches it. Conventional semiconductor detectors either can't see in the infrared or can't see such brief flashes, or they report flashes when there aren't any.
Surprisingly, the answer to the chip problem came from an astronomy team in Russia.
"We were working on the photoresponse of superconductors and we contacted this group at Moscow State Pedagogical University that was using superconductors for radio astronomy," says Roman Sobolewski, professor of electrical and computer engineering at Rochester and co-creator of the device. "The upper radio bands are essentially far infrared bands, so we got together with the Moscow team and worked on putting their materials into our detector."
Though the Russian astronomy team, led by physics professor Grigory Gol'tsman,, and the American engineering team ran in the same superconducting electronics circles, it wasn't until they met at a conference five years ago that they came up with the idea of merging Russian technology with U.S. optical instrumentation and aiming it toward developing a completely new class of optical single-photon detectors. Together they won a small, international collaboration grant from Naval International Cooperative Opportunities in Science and Technology Program, administered by the U.S. Office of Naval Research, and a grant from the North Atlantic Treaty Organization Scientific and Environmental Affairs Division in Bruxelles.
The experiments showed that ultrathin strips of a metallic compound called niobium nitrite a millionth of a meter wide and only several atoms thick could detect single visible and infrared photons. Inside a kind of thermos of liquid helium at temperature near absolute zero, the strips, fabricated in Moscow, become a superconductor-able to conduct electricity without any of the resistance found in normal conductors like copper wires. This lack of resistance is essential because it makes the superconductor like a calm pond; toss in the smallest stone and you'll notice the ripples. A single photon of infrared light plunking into the material could be detected, unlike the case with conventional types of detectors that are full of noise, like a storm on the pond obscuring all but the largest perturbations.
The detector also boasts practically negligible "dark counts"-the reporting of a photon when there is none-so it can count accurately even when there are very few photons to count, such as 10 per second or less.
Such precision is crucial because when a clock pulse passes through a chip and the transistors emit just a photon or two of infrared light, most conventional semiconductor single-photon detectors will miss the flash or misinterpret their own stormy perturbations as an incoming photon. The flash can divulge quite a bit about the way a chip is working; for instance, whether or not the transistor is switching at the correct time, a vital consideration for today's incredibly high-speed chips. The Sobolewski and Gol'tsman group is currently the only one in the world with a detector able to efficiently measure such ultra-fast flashes. "We're measuring bursts on the order of picoseconds," explains Sobolewski. Picoseconds are one trillionth of a second.
Sobolewski and his Russian colleagues have a patent pending for the device together with engineers from Schlumberger Semiconductor Solutions, a California company that builds testers for integrated circuits and sponsors the research through the University of Rochester Center for Electronic Imaging Systems (CEIS). Because of the superconducting detector's great speed and accuracy, Schlumberger can now test every chip made, even the fastest ones, which will power future workstations and desktop computers.
Sobolewski sees the superconducting single-photon optical detector making a difference in other fields as well. "NASA is interested in this detector for communications between Mars and Earth," he says. "When you're dealing with such incredible distances, you may only be able to catch a few photons from a transmitter on Mars." Down-to-earth communications may also benefit. The superconducting detector is regarded as the best candidate for future practical quantum cryptographic systems. Quantum cryptography, where bits of information are coded and transmitted as single photons, offers unconditionally secret, undecipherable communication.
The Russian team has successfully brought its original radio-astronomy detector to fruition as well. It has just been installed in the Sub-Millimeter Telescope Facility on Mount Graham in Arizona, to study, among other things, how stars are born.
The Moscow-Rochester collaboration is now partially funded by the U.S. Civilian Research and Development Foundation for the Independent States of the Former Soviet Union.