Molecule can become a transistor
Return to Virginia Tech "On Campus" page
The problem: Smaller, more powerful microprocessors require squeezing more transistors into a chip -- but there's a limit.
Transistors switch current on and off and amplify current. In existing transistors, this is done by applying voltage to a gate electrode between the input (source) and output (drain) electrodes. More transistors in a single chip means more computational speed. Presently, a single chip can hold up to 28 million transistors, but leakage and tunneling are already a problem. Stray current (leakage) causes crossed signals and electrons bypassing gate fields (tunneling) prevents current amplification.
The solution may be molecular electronics.
"We can use molecules as transistors, switches, and memory devices," says Massimiliano Di Ventra of the Virginia Tech physics faculty.
 |
The benzene molecule used as a |
|
"We want to integrate billions of molecules into a single chip," he says. "We will have this technology probably in 15 to 20 years."
Di Ventra is a theoretical physicist whose research focus is to understand how molecular electronic devices work. He studies how a specific molecule will behave under current flow. "I inject current into the molecule to see if it can work like a transistor, a switching device, and the like," he says.
This spring, he demonstrated that a benzene molecule can work as a transistor (published in the June 5, 2000, issue of Applied Physics Letters), acting not only as a switch but also as an amplifier. Nature columnist Philip Ball hailed the work in a "Science Update" article titled "Painless Gain" (Nature, June 2, 2000).
The transistor's role as an amplifier is critical to ensure that signals remain strong as they pass from place to place. "So far, obtaining gain from a single-molecule device has been a big stumbling block for molecular electronics," Ball wrote. "Now Massimiliano Di Ventra . . . and his colleagues have shown that this hurdle is, in principle at least, surmountable."
Benzene is a common molecule made up of six carbon atoms forming a hexagon on a plane. It is abundant and cheap to manufacture. Di Ventra; Sokrates Pantelides, a colleague in physics and astronomy at Vanderbilt University; and Norton Lang of the IBM Research Division in New York, did a computer simulation of a benzene molecule between two electrodes (the source and the drain) and applied an electric field perpendicular to the molecule (the gate field).
In the theoretical simulation with the benzene ring molecule, the drain and source are two gold electrodes connected to the benzene molecule by sulfur atoms. The gate consists of two charged metal disks above and below the molecule and between the electrodes. The electrons flow from source to drain, but the gate field can be adjusted to control the electron flow.
With low voltage at the gate field, there is a very low probability of electrons tunneling across the molecule, Di Ventra says. By increasing the gate field, an "electronic bridge" is formed between source and drain and electrons can tunnel across the molecule easily, allowing a large current flow. This electronic bridge is called "resonant-tunneling." Thus, the molecule acts as a switch, and the signal is amplified by the gate as in conventional field-effect transistors.
"Now, tunneling destroys chips if they are too crowded; but, with molecules we can use the phenomenon to our advantage," Di Ventra says. "We demonstrated that single molecules can do the same job as transistors. But these single elements need to be combined to form molecular chips. This is a major technological problem. It is like in the 1940s when the transistor was invented: It took 25 years before transistors could be put together in integrated circuits.
"The next step in molecular electronics is to develop molecular chips that will replace the ones we use in our computers," he says.
Di Ventra has ongoing collaborations with Mark Reed at Yale University, who is working on molecular devices and molecular memories, and Phaedon Avouris at IBM Research Division in New York, who is working on carbon nanotubes field-effect transistors.
See another article about Di Ventra's work in the National Partnership for Advanced Computation Infrastructure's EnVision Science magazine.
Massimiliano Di Ventra, assistant professor of physics at Virginia Tech, can be reached at diventra@vt.edu.