Replacing wires with light in chips.
Chips that communicate with pulses of light instead of electrical signals could lead to computers that are more power-efficient than today’s best machines and up to 1,000 times as fast. IBM researcher Solomon Assefa has brought this prospect a critical step closer.
Assefa has developed a new way to make a photodetector, a very sensitive device that amplifies optical signals and converts them into electrical signals that can be shuttled around in a microprocessor. Ordinarily, photodetectors are made using a process called chemical vapor deposition. But sticking with this process for chip-to-chip connections would make microprocessor manufacturing prohibitively expensive. Instead, Assefa seeds germanium onto a silicon wafer, and then melts it to achieve the regular crystal structure that makes for a good photodetector material. He has also determined when in the chip manufacturing process the photodetector should be added in order to get the best performance possible without degrading the surrounding electronics.
Assefa can demonstrate the performance of his photodetector in the lab. But before a chip incorporating his creation can be commercialized, he will have to figure out how all the rest of its elements can be integrated efficiently. Making today’s integrated circuits requires hundreds of steps and dozens of lithographic masks, the stencils used to pattern features on chips. “We don’t want to change any of these processes or it really increases the costs,” he says.
Tailoring polymers for biodegradable implants.
As a graduate student at MIT, Christopher Bettinger created strong, rubbery polymers that mimic natural tissue and can be tailored to break down after anywhere from two months to two years. For Bettinger, the hardest part was making sure the molecular building blocks of his polymers were interconnected enough to yield a material that held its shape but not so strongly interconnected that the result was brittle. He initially used the new polymers to make scaffolds for laboratory-grown tissue. Now, as an assistant professor at Carnegie Mellon University, Bettinger is using them to produce degradable catheters and drug-delivery systems that he’s testing in animals.
As part of his postdoctoral work at Stanford in 2009, Bettinger also created a biodegradable semiconductor for electronics used in temporary medical implants. Simple electronic circuits constructed from biodegradable materials could lead to drug-delivery devices and nerve-regeneration scaffolds that a doctor would trigger with radio frequencies from outside the body. Once therapy was complete, the devices would disappear without a trace.
Using semiconductors to steer light.
PROBLEM: Metamaterials, a new class of artificial materials that can affect light in ways not possible in nature, open the door to things like real-life invisibility cloaks and computers that use photons instead of electrons. But current metamaterials absorb or scatter too much light to make such devices practical.
SOLUTION: Alexandra Boltasseva, a professor of electrical and computer engineering at Purdue University, is replacing the metals normally used in metamaterials with semiconductors, such as zinc oxide, that have been doped with aluminum or gallium. Doping the semiconductor makes it behave more like the metals used in metamaterials, but without the associated optical losses. Currently, these doped semiconductors are suitable for manipulating infrared light, and Boltasseva is working on developing formulations that will work with visible light. Another advantage of these materials is that their properties can be altered by applying an electric field, which would make them suitable for applications such as communications and computing.
“We are talking about a whole new generation of devices that are based on new principles of manipulating light,” she says.
Solar cells that see more light.
As it passes through a solution in a small vial, the green light from a laser pointer in Jennifer Dionne’s hand turns into a sparkling blue beam. By making materials that perform a similar color conversion on sunlight, Dionne hopes to boost the output of solar cells and improve the economics of solar power.
Thirty percent of the sun’s light is wasted in even the best of today’s solar cells because this near-infrared light has too little energy to interact with materials in the cells. Other solar researchers have tried to do what Dionne is doing—”upconversion”—by combining two dyes that interact with each other to convert two low-energy photons into one high-energy photon. But Dionne is taking a new approach that could improve upconversion efficiencies by as much as 50 percent. She added metal nanoparticles to an existing combination of upconversion dyes; the particles shine more light on the dyes and get more converted light out of them.
It’s an early demonstration, but solar-cell maker Bosch is working with Dionne to develop dyes that perform the upconversion. The technology could be incorporated into solar cells in seven to 10 years.
Stretchable electronics for medical devices.
Surgery for a common type of cardiac arrhythmia could be quicker thanks to Dae-Hyeong Kim, an assistant professor of chemical and biological engineering at Seoul National University. Kim has built a balloon catheter that can expand to one centimeter in diameter and is equipped with 150 nanometer-thin metal wires that connect to 13 electrodes. Pushed through blood vessels, this device allows a surgeon to detect electrical misfires in 13 patches of heart tissue at a time and use radio energy to blast any patch where a misfire is found. Previously, surgeons had to detect misfiring regions one by one with a single wire and then zap any problems with a second wire.
The catheter is just one application of Kim’s bendable, stretchable high-performance silicon electronics, which could be used in everything from prosthetic neural interfaces to brain implants for controlling Parkinson’s. Devices made from the materials are sensitive enough to detect subtle changes in human physiology, while the wiring is tough enough to deliver cell-zapping energy.
Kim’s catheter is now being tested in pigs. Another of his devices, a flexible sheet of electrodes that can be draped over delicate tissues, has been used to map abnormalities in the brains of epileptic cats. Both are under development by the startup MC10.
Replacing silicon with graphene.
Fengnian Xia has found a way to build a graphene transistor that blocks electrons efficiently when it’s off—a step in the direction of graphene-based electronics, which could lead to smaller, faster microprocessors.
Graphene is a promising material for transistors because it conducts electricity better than silicon does. Unfortunately, it’s hard to stop the flow of electrons in graphene, and that’s an essential function for any transistor. Xia’s solution is based on a type of transistor in which an electric field is applied to two layers of graphene. In theory, this arrangement should stop the flow of electrons, but in practice, results were disappointing until Xia put a thin polymer layer on top of the graphene during fabrication: the polymer kept the electrical properties of the layers from being ruined as later layers were added to make other parts of the transistor.
Xia has also demonstrated that graphene can be used in ultrafast photodetectors for optical communications. The ultimate goal, he says, is an integrated system that uses graphene for both communications and computing. That would mean chips could use fiber optics instead of having their inputs and outputs bottlenecked by the relatively sluggish speeds of metal wires.