Even the whizziest, most cutting-edge digital systems still use microphones that are based on 50-year-old analog electret technologies. Now, all that will change with the announcement this week of an all-digital MEMS microphone--the first to harness standard chip-processing techniques. The single-chip solution from Akustica Inc. offers more than digital outputs that match the digital processor and memories in PCs, PDAs, Bluetooth headsets and cell phones. It also promises to leverage the scaling power of CMOS chips to finally merge this formerly analog component with the mainstream of digital signal processing. Akustica, which is based in Pittsburgh and has 42 employees, will announce the MEMS mic at the Globalpress Electronics Summit, which starts today in Monterey, Calif. Akustica's one-chip model AKU2000 microphone builds on five years of MEMS development of intellectual property the startup licensed from nearby Carnegie Mellon University. Its inventor is MEMS pioneer Kaigham (Ken) Gabriel, an EE professor at Carnegie Mellon who co-founded Akustica in 2001. Gabriel researched MEMS at AT&T Bell Labs and spent five years managing MEMS development as director of the Electronics Technology Office at the Defense Advanced Research Projects Agency. He now says that Akustica's first product is the first single-chip microphone in the world to use CMOS. Akustica may need that capacity--that is, if analysts are right about the size of the market it plans to serve. From cell phones to voice-over-Internet Protocol (VoIP) to PDAs and PCs, industry analysts variously forecast that the microphone market could grow to nearly 1 billion units during the next decade from at least 82 million units in 2005.
Monday, February 27, 2006
The standard way to add a microphone to a digital appliance such as a notebook PC, Bluetooth headset or PDA is through a combina- tion of four components: a miniature electret condenser microphone, a discrete field-effect transistor (FET), a separate operational preamplifier chip and an analog-to-digital converter chip. Akustica's digital-output CMOS microphone chip concentrates all four functions in a single device. An electret microphone uses a metallic diaphragm with an air gap between it and a fixed metallic backplate, which acts as the plates of a capacitor. The gate of the FET is attached to this capacitor, and as sound waves move the diaphragm, they change the capacitance between the diaphragm and the backplate. This invokes a change in voltage output from the FET that is fed to the external preamplifier, which in turn feeds an analog-to-digital converter. Akustica's design, too, is based on a metallic diaphragm inducing a change in capacitance, but eliminates the high-impedance input to the FET by directly connecting the diaphragm to an impedance-matched preamplifier that is on the same chip as the diaphragm. Likewise, the A/D converter is also integrated on-chip. Because of the short distance between the diaphragm and the preamplifier, the single-chip MEMS microphone has better isolation from the power supply and output stage as well as being virtually immune to stray RF, electromagnetic and power supply fluctuations. Moreover, its output is digital, making it a perfect fit for digital appliances. The short trace lengths eliminate antenna effects and make reliable, repeatable impedance matching possible for the on-chip preamp. Also, the diaphragm can be less than 0.5 mm in diameter, compared with 4 to 6 mm for the diaphragm of an electret mic.
Researchers at the University of Wisconsin in Madison have discovered a surface treatment that enables nanoscale membranes of silicon to conduct better as they get thinner, thereby extending Moore's Law all the way to atomic dimensions. The team discovered that as silicon layers dip below 200 nanometers in thickness--what the team calls "nano- membranes"--the normal determinants of conductivity, such as dopants, become irrelevant. Instead, it is the atomic accuracy of the surface that determines conductivity. In fact, the team found that when atomically accurate silicon nanomembranes were cleaned of oxide contaminations, they became 10 times more conductive. And as nanomembrane thickness shrunk toward 10 nm, a careful surface preparation could increase conductivity as much as 1 million times. The team made the discovery by accident: It was not setting out to increase the conductivity of nanomembranes, but merely attempting to clean their surfaces for more accurate imaging with a scanning-tunneling microscope. Working with a silicon-on-insulator substrate, the team was trying to clean off unintended oxidation atop its silicon nanomembranes that result whenever chips are exposed to air. The usual method of removing top-layer oxidation, high-temperature annealing, could not be used, because temperatures above 1,200°C cause nanomembranes to ball up. Instead, the team put its test chip in an ultrahigh vacuum and slowly deposited a few monolayers of silicon and germanium to displace the unwanted oxidation and replace it with an atomically precise cap. The resultant scanning-tunneling microscopy revealed that such carefully prepared nanomembranes had greatly enhanced conductivity. To explain the surprising results, the team characterized nanomembranes ranging from 200 down to 15 nm and found that careful surface preparations increased conductivity as the membranes got thinner--the opposite of what happens to unprepared silicon films.
Monday, February 20, 2006
It won't power the starship Enterprise, but an experimental "dilithium crystal" pyroelectric technology is said to enable compact nuclear fusion. Engineers at Rensselaer Polytechnic Institute (Troy, N.Y.) said they have opposed two oppositely charged centimeter-sized lithium tantalate crystals to create a fusion device that can operate off a battery at room temperature. "In a [conventional] fusion device de- signed to produce energy, the release of high-energy ions further heats the plasma, thereby sustaining the reaction. We get the same amount of energy out of the fusion reaction, but we cannot use it to sustain the reaction," said the technique's inventor, associate professor Yaron Danon. "Instead, we plan to use the energy emitted to create a portable neutron source that has applications in non- destructive testing or, possibly, explosive-mine detection," he said. Indeed, Danon predicts that different application areas will benefit from the four types of high-energy particles that a pyroelectric crystal accelerator can emit: high-energy electrons, ions, neutrons and X-rays. The electrons that pyroelectric crystals produce could be used for therapeutic purposes, such as cancer treatments, he said. With some improvements, the high-energy emissions might be used to inspect cargo or scan luggage. Danon performed his research for the Department of Energy with Jeffrey Geuther and Frank Saglime, doctoral candidates in nuclear engineering.
A researcher at Lawrence Berkeley National Laboratories has demonstrated a fuel cell measuring just 200 nanometers across that potentially can be integrated on-chip to supply power from a hydrogen reservoir for decades. Today, there are only two ways to power remote sensors and similar devices that require little power over years of unattended use. For devices with lifetimes of less than 10 years, the solution is expensive, bulky lithium batteries. For longer lifetimes, the answer is batteries that draw energy from radioactive isotopes. While experimenting with making metallic nanowires at the University of Wisconsin, Madison, Kenneth Lux hit upon a way to build three-dimensional electrodes porous enough for nano fuel cells. By making a nanowire alloy of two metals, he found it was possible to remove the atoms of one metal in the alloy, leaving behind a densely porous 3-D structure that increased the surface area of the electrode by orders of magnitude. Lux discovered the best way to make porous 3-D platinum electrodes: soak copper-platinum alloy nanowires in nitric acid, removing their copper. Later, they found, they could create nano fuel cells by merely laying them out lithographically so their anode and cathode electrodes protruded from the same side, with a liquid electrolyte reservoir that bent to chemically connect them. With concept proven, Lux is trying to replace the liquid electrolyte with a solid-state version, enabling future remote sensor chips to potentially integrate all the components but fuel for arrays of on-chip fuel cells.
Monday, February 13, 2006
Carbon nanotubes are nearly ideal one-dimensional semiconductors, but current fabrication techniques, such as arc discharge, laser ablation and chemical-vapor deposition, leave their ends capped. To attain good wetting with solders, low-contact resistance with the substrate and unimpeded emission from the nanotubes' tips requires opening their ends with an extra etching step. That process invariably damages them, says Georgia Institute of Technology professor Ching-Ping Wong. Sometimes the ones with poor adhesion from the substrate are entirely dislodged, fraying their ends, causing them to age prematurely and overall degrading their electrical performance in circuits. Current carbon nanotube growth methods also can't be combined with inexpensive low-temperature substrates. Wong claims to have a new carbon nanotube growth and fabrication technique that solves all those problems by separating the growth from the assembly of nanotube-based devices. His "carbon nanotube transfer" technology is a two-step process whereby sheets of open-ended carbon nanotubes are separately grown on silicon substrates, then transferred to epoxy substrates in a manner similar to that used for flip-chips.
Posted by R. Colin Johnson at 7:00 AM
Monday, February 06, 2006
Dye-sensitized solar cells present a low-cost option for renewable power generation, but their efficiency has maxed out at a dismal 11 percent. Now a project at Pennsylvania State University suggests that incorporating titania (titanium oxide) nanotube arrays could provide the needed efficiency boost to move the cells toward commercialization, according to Craig Grimes, a professor of electrical engineering and materials science and engineering at Penn State. Solar cells directly convert sunlight into electrical energy by generating free electrons from incident photons. Photovoltaic conversion was discovered as long ago as 1839 by Alexander Bequerel. Dye-sensitized solar cells--invented in the 1990s--depend on a thin-film version of photovoltaic conversion. But, unlike silicon-based thin-film cells, in which light is absorbed by an expensive semiconductor, in dye-sensitized solar cells absorption occurs in an inexpensive thin film comprised of dye molecules attached to titanium oxide nanoparticles in an electrolyte. When the dye cells absorb a photon, the resultant excitation injects electrons into the titanium, which transports them to the negative electrode, with the positive electrode attached to the electrolyte. Today, dye-sensitized solar cells remain a laboratory curiosity, chiefly because no one has come close to achieving their theoretical maximum efficiencies. Many researchers have worked on the problem, because dye-sensitized solar cells could potentially be manufactured very cheaply, but inefficiencies inherent in the transport of electrons to the negative electrode have severely limited their performance. As a result, even the best dye sensitized solar cell prototypes only have about 11 percent photovoltaic conversion efficiency, Grimes noted. To solve the problem, Grimes' research group proposes substituting titania nanotubes for nanoparticles. Consequently, when electrons are liberated by photovoltaic conversion, they can move to the negative electrode via ballistic transport along the titania nanotubes.
Friday, February 03, 2006
A new technology offers advance warning of the ability of a virus to infect humans. The technique could help stem the spread of bird flu, according to the inventors of glycan microarray technology. The Consortium for Functional Glycomics, a project of the National Institute of Health's National Institute of General Medical Sciences, is making the technology widely available to researchers for free. The microarray can pinpoint pathogens in a few hours that can infect humans. The glycan microarray was created by consortium members at The Scripps Research Institute in cooperation with Mount Sinai School of Medicine and the Armed Forces Institute of Pathology. Microarrays work by depositing hundreds, even thousands, of different examples of DNA. In this application, however, carbohydrate molecules called glycans are used. Samples are then spread across the entire array. Finally, a fluorescent-tagged antibody that binds to samples is spread across the array and illuminated with a laser. A photocell detector recognizes the precise site where the sample binded, thereby pinpointing a specific glycan and revealing whether the sample could infect humans. For viruses and many other pathogens, the carbohydrates on human or bird cell surfaces offer a binding site that enables them to burrow through and cause infections. The Consortium for Functional Glycomics was formed to determine which carbohydrates coat which cells as well as catalog their vulnerability to infections from various pathogens. The glycan microarray was developed to house all this knowledge on a single chip so that carbohydrate binding site vulnerabilities can be quickly identified for a variety of maladies.
Posted by R. Colin Johnson at 5:32 AM