| NEWS & IDEAS |
January 1998
INTEGRATED ELECTRONICS:
Door Opens for Blue Lasers
INTEGRATED ELECTRONICS:
Aerogels Could Speed Computers
MANAGEMENT AND TECHNOLOGY:
'Live and Don't Learn' Costs Billions
GAS, LIQUID INTERFACES:
Pinning Down an Elusive Number
GEOLOGY:
Crystal Diary Holds Eons of History
INTEGRATED ELECTRONICS:
Door Opens for Blue Lasers
Scientists at Rensselaer Polytechnic Institute have found a way to grow aluminum nitride crystals that are large enough to slice into semiconductor substrates. No one else has been able to do that successfully.
These crystals can be used to fabricate blue and ultraviolet lasers and blue and green light emitting diodes (LEDs).
Semiconductor light sources have always been very attractive because of their ruggedness and economy, says Leo Schowalter, professor and chair of physics at Rensselaer. But the color of LEDs has been pretty much limited to red. Green and blue LEDs are also needed if we are to create traffic signals, automobile lighting, flat-screen television sets, and other applications where long life and high efficiency is important.
In addition, blue and ultraviolet semiconducting lasers would make it possible to squeeze as much as 30 times more material onto a compact disk than can be done with the infrared lasers that are currently used.
As a substrate, aluminum nitride is also ideal for semiconductors in wireless communications and power industry applications. Because aluminum nitride endures extreme heat, it can be used for microelectronic devices on jet engines. But growing aluminum nitride crystals is very, very difficult, says Schowalter.
"Glen Slack, one of our research professors, demonstrated that you can grow aluminum nitride crystals in a tungsten crucible at 2300¡C. But at that temperature, the aluminum attacks the grain boundaries in the tungsten and the crucible doesn't survive very long."
Schowalter and Slack have now solved the problem and formed a company, Crystal I.S., to create the aluminum nitride crystals.
Contact: Contact: Leo Schowalter (518) 276-6435, schowalt@unix.cie.rpi.edu, or http://www.rpi.edu/~schowl/schowalt.htm
INTEGRATED ELECTRONICS:
Aerogels Could Speed Computers
Rensselaer researchers are creating and studying aerogels, substances so porous they are more air than solid material. When used as insulators on computer chips, these porous materials could more than double computing speeds.
Researchers in the Semiconductor Research Corp. Center for Advanced Interconnect Science and Technology (CAIST) at Rensselaer are producing detailed studies of the aerogels and the processes by which they are created, as well as computer models to predict their performance and final properties.
Texas Instruments announced this month that it has demonstrated the successful combination of copper wiring with a similar substance, xerogel. Illustrating the importance of this new type of insulator, the company estimated that within a decade, combining xerogel with copper wires and new designs could result in devices that are 10 times faster than today's best chips.
The Rensselaer aerogels were on display in November when CAIST members gathered on campus to hear reports. CAIST, an interdisciplinary university consortium, was established by industry to improve interconnects, the minuscule system of wires and insulation that carries messages on a chip.
Good insulators - materials with a low dielectric constant - let designers place lines close together and do not slow down the signal. Silicon dioxide, the material now used on most chips, has a dielectric constant of about 4. Decreasing that number to 2 could at least double the speed of computers.
CAIST researchers are also studying polymers that could bring the number down to about 2.5. Air, the perfect insulator, is rated at 1.0. You can't hold chips together with air, but three members of Rensselaer's Isermann Chemical Engineering Department are learning how to solve the problem with aerogels.
Joel Plawsky, Peter Wayner Jr., and William Gill have successfully created highly porous silica films that are between 65 and 90 percent air, with a dielectric constant ranging from 2.3 to 1.4. The team has shown that it can control porosity and thickness. The new films apparently do not create problems by absorbing water during processing, and they stand up well to high temperature.
There still are numerous technical questions to be answered, but the new materials could be used by industry within five years, Plawsky says.
Contact: Joel Plawsky (518) 276-6049, plawsky@rpi.edu; Peter Wayner Jr. (518) 276-6199, wayner@rpi.edu; William Gill (518) 276-2880, gillw@rpi.edu
MANAGEMENT AND TECHNOLOGY:
'Live and Don't Learn' Costs Billions
Billions of dollars are now lost in shortsighted systems development projects because companies don't learn from experience, says Thiagarajan Ravichandran, assistant professor in Rensselaer's Lally School of Management and Technology.
Experience gained in the development of information systems is not preserved and passed on for use in other projects, says Ravichandran, who surveyed procedures at 123 Fortune 1,000 companies and large federal and state agencies.
"The problem is compounded by the fact that information systems people change jobs every 1.5 years," says Ravichandran.
Unlike manufacturing processes, learning in the context of systems development cannot be gained through massive repetition. It must be acquired from other projects. Individual and team experience must be processed, packaged, and passed on in revised policies and procedures.
"To assure quality products you must have a quality process," says Ravichandran. "Systems development must change from being a chaotic, ad hoc process to a very well-defined, well-managed, and controlled process that is continually improved based on evaluation and experience.
"Learning and knowledge creation should be the focus of process management. Practices such as fact-based management, fostering participation of key stakeholders, and debriefing of project leaders and system developers at regular periods throughout the project are required to generate learning. The new knowledge must then be incorporated into new procedures."
The reuse of knowledge is as important as the reuse of systems components, which many experts already stress as an additional key in successful systems development, Ravichandran says.
Contact: Thiagarajan Ravichandran (518) 276-2035, ravit@rpi.edu
GAS, LIQUID INTERFACES:
Pinning Down an Elusive Number
An interdisciplinary team of researchers has found a way to measure a key property that is needed to accurately model problems ranging from the ability of the ocean to absorb carbon dioxide to the best processes for producing plastics.
These problems all involve the interface of a gas and a liquid. Since at least the first century B.C., man has known that small amounts of contaminants (surfactants) on the liquid's surface - oil on troubled waters - can greatly complicate the picture by limiting the liquid's ability to move.
For more than 150 years, Navier-Stokes equations have been used to describe motion in fluids, but until recently only the simplest flows could be solved. Computing power has now increased sufficiently to make many solutions possible - if you can get accurate numbers for key parameters.
To create accurate computer simulations of processes at the gas-liquid interface, researchers need accurate numbers for a property known as interfacial dilational viscosity. At present, they have no direct way of measuring this property. Instead, it is calculated by models that can produce answers that vary by as much as a factor of 100,000.
"Suppose you have two methods of calculating the weight of a book and you don't know which is best," says Amir Hirsa. "The first says it weighs one pound and the second says 100,000 pounds. You'd know you have a problem."
Instead of relying on models, Hirsa, associate professor of mechanical and aeronautical engineering at Rensselaer Polytechnic Institute, Gerald M. Korenowski, a Rensselaer chemist, and Juan Lopez, a mathematician at Penn State, have found a way to measure interfacial dilatational viscosity directly.
Funded by the Office of Naval Research, Korenowski and Hirsa use laser techniques to separately determine two processes: changes in movements of particles at the surface and variations in surfactant concentration. Working with Lopez, they then use these numbers to calculate the sum of interfacial dilatational viscosity and another, more easily determined property known as interfacial shear viscosity when an insoluble surfactant lies on the liquid.
At the November meeting of the American Physical Society, Division of Fluid Dynamics, in San Francisco, the researchers reported that they now have developed approaches to come up with separate measurements for the two properties and to extend their calculations to soluble as well as insoluble monolayers.
Contacts: Amir Hirsa (518) 276-6997, hirsaa@rpi.edu; Gerald Korenowski (518) 276-8480, Koreng@rpi.edu
GEOLOGY:
Crystal Diary Holds Eons of History
Pinned to ties, dangled from ears, or set in friendship rings, January's birthstone contains key chapters in the history of Earth's evolution.
Long popular for both jewelry and abrasives, garnet crystals keep their own private diaries in which they record their experience of millions of years of changing pressure, temperature, and other forces.
"Each crystal is really a chemical tape recorder," says Frank Spear, professor of earth and environmental sciences at Rensselaer and recipient of the 1997 N.L. Bowen Award presented this month by the American Geophysical Union's Volcanology, Geochemistry, and Petrology Section.
Spear found that the distribution of various minerals in garnet and the concentrations of iron, manganese, magnesium, calcium, and other elements reflect what each crystal experienced throughout millions of years of growth.
That growth began at a subduction zone where ocean mud was shoved miles beneath the Earth. There the mud eventually crystallized into a stony batter of mica, feldspar, quartz, garnet, and other ingredients. Eventually, the rock was thrust up into mountain ranges, which explains why garnet deposits are found from the Alps to the Adirondacks.
Microprobe "maps" of a slice of garnet make the chemical changes as easy to see as the rings on a tree stump. But the interpretation is far from simple. For that task, Spear created thermodynamic algorithms that scientists everywhere now use in determining how hot the garnet got, how deep it went, and how high it rose throughout its long, rocky career.
Contact: Frank Spear (518) 276-6103, spearf@rpi.edu, or http://www.geo.rpi.edu/facstaff/spear/spear.html
