SIDEBAR: Computing a better LED
In the past, scientists and engineers have learned about
properties of materials empirically. They have experimented with composition and processing
and have developed rules of thumb for what to expect when they try something new. But todayÕs
materials are so complex and are asked to do so much that researchers are turning to new,
theory-based methods that can offer more guidance in their development efforts. And some of
the most advanced work along these lines is taking place at IBM's Zurich Research Laboratory.
There, Wanda Andreoni and Alessandro Curioni are studying the organic molecule tris(8-hydroxyquinolate)
aluminum. Alq3, as it is usually abbreviated, is widely used in organic light-emitting devices because
it is an efficient emitter of light. "Our goal," Andreoni says, "is to better understand the functioning
of these materials, and then to suggest new materials, particularly ones that shine brighter and at
different wavelengths".
To understand Alq3, Andreoni and Curioni have modeled the molecule on a computer. Their calculations essentially
start from scratch, using a technique called density functional theory to compute the molecule's shape, movement
and electronic properties. By comparing the results of their calculations with measurements made on Alq3 by
experimentalists, Andreoni and Curioni have found that they can produce a virtual molecule that behaves just
like the real thing.
More recently, the two researchers have focused on how Alq3 interacts with the cathode Ñ the film of metal that
produces electrons that combine with electron holes in the organic molecule to produce photons. Since this interaction
is critical to efficient emission of light, Andreoni and Curioni want to understand it in great detail. To see how an Alq3
molecule behaves when sitting on a surface of aluminum or some other metal, their models follow hundreds of atoms, tracking
the behaviors and interactions of the electrons for all of them. The calculations demand computing power of a magnitude
found on only a few massively parallel computers around the world.
From those calculations, Andreoni and Curioni also hope to learn how to optimize the interface between
the organic molecule and the cathode. To this end, they have been modeling a variety of surfaces and looking
for differences in how Alq3 performs -- and for the reasons behind the differences. The researchers have also
studied the materials' durability and tendency to degrade over time. Their computer simulations have helped
them identify causes of degradation and find ways to monitor it.
Already, Andreoni says, she and Curioni are far beyond the traditional empirical understanding of materials. "We can predict properties of materials from calculations and design new materials on the computer." Over the next few years, as computers become faster and more powerful, she expects that such "ab initio" calculations will become a widespread tool in finding new and better materials.
]
