Simulation Tool Shows How Current Flows

Engineers at Purdue University (West Lafayette, Ind.) have created a nanotech simulation tool that shows how current flows between silicon atoms and individual molecules. The tool will help researchers design "molecular electronic" devices for future computers and advanced sensors, according to Geng-Chiau Liang, a postdoctoral research assistant in Purdue's School of Electrical and Computer Engineering.

"I believe we might be the first theorists who have created a tool to show how electricity is conducted between molecules and silicon at the atomic level," said Avik Ghosh, a research scientist in electrical and computer engineering who worked with Liang.

The goal is to create sensors and other devices that use molecules, such as proteins and DNA, instead of conventional electronic components. So-called "biochips," for example, could use proteins in sensors for detecting contaminants and pollutants in the air and water and analyzing the blood and biological samples.

"The idea is that molecules might be able to complement or supplement silicon," Ghosh said. "All traditional research in molecular electronics has focused on combining molecules with metal contacts, but we've been studying the interaction of molecules and silicon instead of metals because the computer industry is built on semiconductors, which is silicon."

Liang and Ghosh used the tool to show how current flows between silicon atoms and molecules called buckminsterfullerenes or "buckyballs." Named after the architect R. Buckminster Fuller, who designed the geodesic dome, buckyballs are soccer-ball-shaped molecules containing 60 carbon atoms. A buckyball has a width of ~1 nm, which is roughly 10 atoms wide.

"This paper is a proof of concept showing that our theory is at a point that we can actually look at experiments and explain them quantitatively," Liang said. "We have shown how the conductance of electricity changes when you change the type of bond connecting buckyballs to silicon."

The way in which a new simulation tool predicts current flow between silicon atoms and molecules depends on how the two materials are connected to each other. The graphs on the left are the simulation tool’s predictions, while the graphs on the right are from data collected in experiments performed by other researchers. (Source: School of Electrical and Computer Engineering, Purdue University)

The researchers used their computational model to predict how electricity flows when buckyballs and silicon are connected in three ways. In one case, there is no chemical bond — the buckyball is simply sitting on top of the silicon. In another case, the molecule has been connected to silicon by annealing the silicon. And in the third case, the buckyball is resting inside a tiny pit, a natural defect existing in the silicon.

The model precisely plotted how conduction and voltage changed in the three types of connections, and those predictions agreed with experimental data from other researchers who measured the actual changes in current flow in laboratories.

"Because our predictions agreed with actual experimental data, we know they are accurate," Ghosh said. "This means you can use the model to give theoretical guidance to experiments instead of using strictly a trial-and-error approach."

With Supriyo Datta, the Thomas Duncan Distinguished Professor of Electrical and Computer Engineering at Purdue, and his students, Liang and Ghosh developed the mathematical theory on which the model is based. The researchers used a Purdue supercomputer to develop and test the simulation. The engineers used buckyballs in their simulation because the molecules are well-known in the scientific community and data is readily available. The tool, however, could be used to simulate conduction using any molecule connected to silicon.

"Researchers want to know what kind of molecule can provide specific conduction characteristics, and if they can substitute other molecules for buckyballs," Liang said. "What we can now do is theoretically explain the experiments in quantitative detail, which is really important for any technology. To do this, you must have an atomistic understanding of current flow; basically, how does electricity conduct at the atomic level?"

The research is ongoing and has been supported by the Network for Computational Nanotechnology, funded by the National Science Foundation and directed by Mark Lundstrom, Purdue's Scifres Distinguished Professor of Electrical and Computer Engineering; the Semiconductor Research Corp.; the Defense University Research Initiative on Nanotechnology, which is supported by the U.S. Army Research Office; and the Defense Advanced Research Projects Agency.

For additional information on emerging technologies, go to www.semiconductor.net/emerging.