Friday, June 30, 2006

"CHIPS: Silicon laser harnessed"

Continuous-wave silicon lasers were demonstrated for the first time last year by Intel Corp. But the new lasers generated too much heat to be practical for CMOS devices. Researchers at the University of California at Los Angeles now claim to have the solution: Harness the excess energy with a photovoltaic effect that converts heat back into electricity to power the chip. The researchers also beat Intel in the race for the world's first silicon laser in 2004, when they reported a pulsed silicon laser. The laser could not operate continuously because it produced excess hot electrons. In 2005, Intel followed up with a technique for sinking hot electrons, clearing the way for the first continuous-wave silicon laser. The first version did not use excess energy. The UCLA group has again topped Intel by showing how a Raman scattering architecture can be used in CMOS devices to harness excess energy from hot electrons. The mechanism, stimulated Raman scattering, sidesteps silicon's indirect bandgap that prevents normal lasing. However, Raman scattering works through double-photon absorption, which creates hot electrons as a byproduct and generates excess heat. The researchers claimed that a photovoltaic sink not only enables continuous operation of silicon laser chips, optical amplifiers and similar photonic devices, but it can now be economically incorporated into CMOS. Moreover, the excess energy can now be harvested to drive other CMOS circuitry.

Monday, June 26, 2006

"CHIPS: Defects spur light emission in nanotube FETs"

IBM Corp. researchers say they have characterized four types of carbon nanotube field-effect transistsor defects that can stimulate nanotube FETs to emit light. The first light-inducing defect was discovered on the end contacts, where a natural Schottky barrier exists at the semi- conducting nanotube interface with its metal electrode. The second was found anywhere that charge had been inadvertently trapped in the oxide-covered silicon wafer; trapped charge locally inverted the carriers in the nanotubes atop it by forming a light-emitting intratube npn or pnp junction. The third defect characterized was on a nanotube loop where the tube bent around and crossed over itself. At the crossover point, hot carriers could tunnel from one leg to the other, where they affected other carriers and caused light emission. The first three defects were studied by randomly laying down nanotubes on a substrate and looking for those anomalies. But the fourth type of defect was intentionally made by partially covering a nanotube with a polymer. Where the polymer ended, a voltage drop caused carriers to collide, inducing electroluminescence.

Wednesday, June 21, 2006

"CHIPS: Molecular switch promises to solve sticky situations"

Microfluidic devices sense and sort through molecules by channeling them down nanoscale pipes etched from polymer substrates. Unfortunately, the tiny channels often clog when biological materials stick to them, degrading performance. Researchers at the Rensselaer Polytechnic Institute (RPI, Troy, N.Y.) think they have found a remedy: a material that optically switches from slippery to sticky. When exposed to UV light, the polymer dislodges molecules stuck to its surfaces by becoming more slippery. The property enables even the most clogged microfluidic channels to be flushed clean. The researchers predicted that the polymer will be useful for filtering specific proteins from biological fluids, which often clog the pores of conventional filters.

Monday, June 19, 2006

"CHIPS: Record-low gate voltage for nanotube field emitters"

Scientists at the U.S. Naval Research Laboratory in Washington D.C., said they have fabricated arrays of high-current carbon nanotube field emitters with a record-low gate voltage of just 60 volts for emissions of up to 1.2 amps per square centimeter. Its low voltage-high current operation also yields the high transconductance--the rate of increase in current with increase in voltage--that is necessary for many electronic device applications. The arrays of carbon nanotube field emitters were grown by researchers David Hsu in the Chemistry Division and Jonathan Shaw of the Electronics Science and Technology Division, using chemical vapor deposition (CVD) on silicon substrates that had been prepatterned with 1-micron-diameter posts centered on 2.5-micron-diameter metal aperatures. Nickel or iron catalysts were deposited over the gated structures and then removed from the surfaces. The carbon nanotubes grew from the catalyst particles during CVD using ammonia and acetylene gas. By applying a relatively low voltage to the gate apertures, a relatively high electric field was produced at the nanotube tips, thereby causing electrons to spew out by field emission, the researchers reported. The proximity of the gates yielded the high local electric fields from the relatively low voltage input. The high-density field emitter arrays packed more than 75,000 emitters per square millimeter, according to the researchers. The scientists said their low-voltage emitters do not degenerate during extended usage as rapidly as high-voltage field emitters do. The low gate voltages avoid the damage typical in high-voltage devices, such as dielectric breakdown and residual ion aputtering, they said. The field emitters were said to perform well at temperatures up to 700°C and in the presence of water vapor and other common gases. The Navy researchers said that the 1.2 amps per square centimeter current density is more than adequate for applications such as spacecraft propulsion systems (ion thrusters and tethers), miniature x-ray sources, cathodoluminescent devices (flat-panel displays) and mass spectrometers. Another factor of ten to twenty higher current density is needed for high frequency electronics.

Thursday, June 15, 2006

"ALGORITHMS: Technique takes flight to quickly erase hard drives"

In 2001, an American spy plane collided in the air with a Chinese fighter and was forced to land on Chinese island. Since then, researchers have been looking for a way to quickly erase computer hard drives to deny access to sensitive intelligence data. Scientists at the Georgia Institute of Technology (Atlanta), working with L-3 Communications Corp. (New York), said they have developed a technique for quickly erasing hard-disk drives. The team reports development of a prototype fast-erasure system to prevent sensitive information from reaching enemy eyes. At the time of the U.S.-China incident, there was no way the U.S. crew could quickly erase hard drives on the surveillance aircraft before landing on Chinese soil. The Chinese eventually gained access to U.S. military secrets. Erasing a hard drive usually takes hours using special procedures that repeatedly scramble information on a disk drive. Still, given unlimited resources and time, special magnetic snooping techniques can often recover at least some of the original information. The researchers sought a method that not only securely erased information but also performed the erasure during emergency situations where minutes, not hours, were available. The researchers concluded that permanent magnets are the best solution. Other methods, including burning disks with heat-generating thermite, crushing drives in presses, chemically destroying the media or frying them with microwaves all proved susceptible to sensitive, patient, recovery efforts. Permanent magnets for erasing magnetic media have been available since the dawn of disk drives, but the team found that commercial systems were either magnetically too weak, too large and heavy or could not meet air-safety standards. Instead, the team crafted a new generation of super-powerful magnets to penetrate hard disk enclosures to quickly erase magnetic media. Special high-strength magnets as powerful as those in medical imaging equipment proved sufficient for permanently erasing all information on a disk drive in a single pass.

Tuesday, June 13, 2006

"CHIPS: Indium oxide may best GaAs for spintronics"

Harnessing electron spin for optomagneto-electronic devices will depend on materials that, like silicon, can separately adjust their densities and dopant levels to designer specifications. Indium oxide doped with chromium may fill the bill, according to a research team from the Massachusetts Institute of Technology and Boise State University. Lacking the limitations of gallium arsenide-based ferromagnetic materials, chromium-doped indium oxide could enable durable, transparent thin-film spintronic devices, said researchers at MIT's Francis Bitter Magnet Lab. According to the researchers, ferromagnetic memories will need to harness spintronic approaches within a decade as they scale down into molecular-sized magnetic domains to store information. Chromium-doped indium oxide and similar formulations could enable the magnetic spin of even individual molecules to be flipped from "up" to "down," potentially packing a bit of data into every atom. By the time FRAMs scale down to molecular-sized domains, MIT researchers hope to have chromium-doped indium oxides fully characterized and ready to build spintronic devices. For now, they are merely reporting that they have overcome the basic limitations of other ferromagnetic formulations, especially those using GaAs, by adjusting the location of molecules in the crystalline lattice of indium oxide and by setting the dopant levels separately.

"CHIPS: Decoupling capacitors return to simpler times"

AVX Corp. has made what it claims is a revolutionary step backward with its invention of a patented new land-grid array (LGA) architecture for decoupling capacitors. Decoupling capacitors provide on-demand power reserves that can keep the edges sharp on the quickly changing signals from high-speed microprocessors, graphics coprocessors, digital signal pro- cessors and field-programmable gate arrays. Unfortunately, as clock frequencies have climbed, capacitor designers have quelled the parasitic inductance that can limit switching speed by resorting to hard-to-use multilayer architectures that require terminals that exit from the side. AVX has lowered parasitic inductance by returning to the simpler design of old-school capacitors, which feed directly into the printed-circuit board from the bottom. Most capacitors are just parallel metal plates separated by an insulator, and hence have essentially no inductance. In working capacitors, however, parasitic inductance is a fact of life. When current flows, the flux creates a magnetic field that loops around the internal electrodes of the capacitor and its external termination, the power planes, vias, mounting pads and solder fillets of the printed-circuit board. AVX's LGA capacitors solve the parasitic inductance problem through a reorientation of the internal electrodes. Instead of a horizontal orientation, in which the electrodes are parallel to the substrate, the electrodes are given a vertical orientation. The LGA architecture enables multilayer capacitors to gang together their multiple electrodes internally, with only two terminals exiting the capacitor from the bottom of the package. This allows printed-circuit-board makers to use a single two-terminal component with bottom exits, rather than using multiple small capacitors, each of which might have up to eight terminals exiting from the sides.

Monday, June 05, 2006

"MATERIALS: Team posits 'cloaking' nanomaterial"

The researcher who introduced the concept of negative-index-of-refraction metamaterials in 2000 is now positing that materials with a variable refractive index could enable such fantastic applications as a Harry Potteresque "invisibility cloak." Sir John Pendry, a physicist at Imperial College in London, predicted six years ago that metals could be engineered to make electrical fields behave oppositely to normal, yielding negative-index-of-refraction metamaterial composites. Since his prediction, such metamaterials have been created and demonstrated from gigahertz to optical frequencies. Now Pendry has teamed with Duke University EEs David Schurig and David Smith to predict that both the electrical and the magnetic properties of an inhomogeneous composite with embedded nanoparticles could be altered to create a variable-index-of-refraction material. They postulate that such a material could adapt at the nanoscale to conceal what's under it by preventing electromagnetic energy from entering an area. Light hitting the material would "flow" around it and continue, undistorted, on the other side. The material thus would neither reflect light nor cast a shadow.