June 30, 2008
Stanford, Samsung engineers sort out nanotube puzzle of keeping different types apart
By David Orenstein
Imagine you are a farmer who wants to grow corn and soybeans on different fields, but every bag of seeds you can buy has both plants mixed together. You don't want to separate the seeds by hand. Instead, you engineer a different solution: You change the ground itself so that one field will grow only corn and the other only soybeans.
That, metaphorically speaking, is how a team of Stanford and Samsung engineers has tackled one of the most frustrating problems holding back the progress of research on carbon nanotubes: reliably separating nanotubes that conduct electricity from those that semiconduct electricity. In the July 4 edition of the journal Science, the researchers describe their technique to sort the nanotubes while simultaneously laying them out on a silicon wafer.
Carbon nanotubes, incredibly tiny structures that in powerful microscopes look like rolled up sheets of chicken wire, are the darlings of many electronics researchers. Semiconducting nanotubes can be used for transistors, the tiny switches that allow computer chips to process information. Metallic nanotubes, meanwhile, conduct electricity very well and could act as ultrafine wiring. Both kinds of tubes are so small, on the scale of billionths of a meter in diameter, that they have the potential to become the materials of choice in computer chips of the future. But nanotubes have presented several tricky problems that researchers have been struggling to overcome.
When researchers try to make transistors with nanotubes, the unwanted presence of metallic tubes causes the transistors to conduct current even when they are not supposed to. Conversely, when the researchers attempt to make wiring, the presence of semiconducting tubes sometimes blocks the flow of current. What has been needed is a way to keep the two apart that is reliable but not impossibly tedious. Led by Stanford's Zhenan Bao, the Stanford-Samsung team has found a way to pattern and then coat a 5-inch silicon wafer with chemicals that lock down only the desired type of nanotubes, like the imaginary farmland that would accept only one type of seed.
"During the process of depositing the nanotubes on the silicon substrate, the separation also takes place," said Bao, an associate professor of chemical engineering. "We can place the nanotubes where we want them to be and select the kind of nanotubes we want."
In addition, the nanotubes end up aligned well enough to make connections across electrodes and gather with enough density to ensure those connections, she said.
Capture by chemistry
In the experiments described in the Science article, the team first cleaned the silicon surface to expose a layer of hydroxyl groups, which are simply pairings of single oxygen and hydrogen atoms. Subsequent chemical reactions caused a new layer of molecules to take the place of the hydroxyls. At the tip of each pile in the layer was an amine (nitrogen and hydrogen) that sticks more strongly to semiconducting nanotubes than to metallic ones. Alternatively, if the tip is a benzene ring, metallic nanotubes will stick more strongly to it than semiconducting ones.
The team then spread a solution of nanotube mixture over the substrate by using a machine that spins the solution into a thin film over the whole wafer. Only the more strongly bound semiconducting nanotubes remained stuck in place amid the centrifugal force at play on the spinning wafer. The metallic nanotubes were flung from it. What was left was a meshy network of semiconducting nanotubes that could then be stamped with electrodes to make transistors.
The team verified the results with spectroscopy and by measuring the switching properties of the transistors they had made after depositing metal electrodes on the chip. The transistors had "on-off" ratios of conductivity averaging around 100,000, meaning that there was a very significant difference in the current flow through the transistor when it was "on" and when it was "off," Bao said. In contrast, previously reported transistors made of meshes of semiconducting nanotubes, plagued by the presence of metallic nanotubes, have had ratios of less than 100.
These room temperature solution-deposited transistors also had a high "charge carrier mobility" in the order of 1-6 square centimeters per volt second, a measure of how fast charges can move through them when they are "on."
Bao said her group is now working on the next step, which is to combine the use of strategically positioned amines and benzene rings to create whole circuits of nanotube transistors and the metallic nanotube wiring and electrodes to connect them.
"The method we've developed could be very useful for developing low-cost, transparent electrodes for large area displays and solar panels, or as a means of sorting nanotubes for other applications," Bao said.
Such feats, as yet unrealized, would demonstrate that nanotubes are on their way to being as useful as they have long been presumed to be.
In addition to Bao, the paper's other authors are lead author and postdoctoral researcher Melburne LeMieux and graduate students Mark Roberts and Soumendra Barman, all of Stanford, and Samsung Advanced Institute of Technology researchers Yong Wan Jin and Jong Min Kim.