4 Inch Wafer of Trillions of Aligned Carbon Nanotubes

Carbon nanotubes can be five times as energy efficient and five times faster than silicon. engineers at the University of Wisconsin–Madison can align the nanotubes for computer chips by turning them into 2D liquid crystals. They can coat an entire 4-inch wafer with a uniform array of highly aligned carbon nanotubes in 40 seconds.

A purification process produces nanotubes in solvents, or nanotube inks, which they then flow at a constant velocity and thickness over a layer of water. At the interface between the ink and water, the nanotubes begin to concentrate and self-organize, forming a liquid crystal. That liquid crystal is then transferred on a substrate moved through the ink and water interface. The result is a wafer covered in trillions of highly aligned carbon nanotubes.

The nanotube density is near that needed for electronics. They determined the alignment of the tubes to be within 6° locally. This almost ideal nanotube ordering led to excellent electrical properties that were confirmed consistent across the entire wafer.

The process is a big advance for carbon nanotube research. However, it does need some tweaks before nanotube computer processors end up in smartphones and laptops. The industry standard is 12-inch wafers, so the process, patented through the Wisconsin Alumni Research Foundation, needs to be scaled up while maintaining the uniformity of the nanotube alignment.

They can also be deposited in multiple layers, like 3D integrated circuits. That would allow us to increase the number of transistors significantly.

SOURCE – University of Wisconsin–Madison
Written by Brian Wang, Nextbigfuture.com

7 thoughts on “4 Inch Wafer of Trillions of Aligned Carbon Nanotubes”

  1. There may be some synergy between this research and the Nantero memory. Perhaps the yield for their cells would be higher if the CNT's were aligned in one direction only?

  2. Cudos to the researchers. This is a major step forwards; aligning CNTs on wafers has been a long sought, but elusive, goal.

    Now, it has to be noted that the CNTs are aligned, but the density is stochastic. This means that you cannot be certain that a certain surface patch contains a CNT. To be able to use the CNTs as conductors instead of metal layers on a wafer, you would have to use a contact area for the CNT that is sufficiently large so that you can be certain to have a few CNTs in that area. Thus, you cannot use the small diameter of the CNT (for SWCNT, 1-3 nm) to go "below" the lithography limit of current systems (8 nm linewidth for the comming generation of EUV lithography?) and create a conducting line that is extremely narrow.

    Furthermore, you have excellent – provided you make sure to have only metallic SWCNTs on the wafer – conduction along the major axis of the CNTs. But in order to "splice" together several CNTs to span the length of the chip – the CNT's have a finite length – or to make a 90-degree turn you would have to deposit a metal "blob" to act as a "jumper" from a group of CNTs to the next group.

  3. Curiously, I used to be a commercial printer, and watching water do things on surfaces and mix with stuff and ride upwards on metering rollers is far from a science. Also, remember that a practical definition of *turbulence* is water running down a trough. This process seems to rely on g to form the interface, where the tubes settle to the water surface and align by following the flow, which may also use g(?). They "begin to concentrate and self-organize" while floating, then are carried by water surface tension to the eventual permanent substrate.

    My first idea would be to look at the "self-organize" aspect and see if there are similarities to already proven micr0g crystal growing, thus advantages. The uniformity of the product, the size of the crystals, is the need here also. If the water were replaced by a more gel substance, no g would be needed to form the interface. Then the transfer from the gel would also be mechanical rather than floating. Mail me the check!

  4. This process seems to be dependent on the properties of the ink/water interface. In micr0g water will tend to become droplets, wobble around, and the interface will have a small radius of curvature.

    This one may work better in gravity.

Comments are closed.