Thanks for the Memories

As electronic musicians, we have insatiable appetites for digital storage; no matter how much we have, we always need more. Hard disks and solid-state
Image placeholder title
Image placeholder title

FIG. 1: Molecules of iron terpyridine attach themselves to a nanotube of indium oxide, forming a memory cell capable of storing three bits of data in far less space than silicon memory cells.

As electronic musicians, we have insatiable appetites for digital storage; no matter how much we have, we always need more. Hard disks and solid-state storage (flash memory) keep increasing in capacity while decreasing in cost, but that trend can't continue forever.

Or can it? If research that is being conducted at the University of Southern California's Viterbi School of Engineering ( and the NASA Ames Research Center in Palo Alto, California, comes to fruition, solid-state storage capacity could expand dramatically beyond what is possible with conventional flash memory.

USC associate professor Chongwu Zhou leads a team that has demonstrated the potential of nanotubes as the basis for molecular memory devices. The nanotubes created by Zhou's team are made of indium oxide and are about 10 nanometers (nm) in diameter and 2,000 nm long. To give you a sense of scale, an average bacterium is about 1,000 nm long, and the smallest known virus is about 20 nm long. One nanometer is about the width of ten hydrogen atoms.

In a process called laser ablation, the nanotubes self-assemble in an oxygen-rich environment after indium is vaporized. They are then placed on a thin layer of quartz and submerged in a special material that coats them. Molecules of the material (in this case, iron terpyridine) attach themselves to the nanotubes (see Fig. 1), forming a field-effect transistor that can act as a memory cell.

Perhaps the most important property of the coated nanotubes is that they can assume one of eight charge states, allowing each one to store three bits of digital information (23 = 8). The semiconducting nanotube is exquisitely sensitive to these states, changing its conductivity accordingly. As a result, the data can be read by sampling the resistance of the nanotube.

Not only can each memory cell hold three bits of data, but the cells are also much smaller than conventional silicon cells, leading to an initial projected data density of 40 GB (a little more than a DVD's worth of data) per square centimeter. Compare that with 1.8 GB/cm2 in current silicon memory and 10 to 20 GB/cm2 in today's hard-disk drives. And that's only the beginning, according to Zhou. “If we scale down the nanowire length by a factor of 10 to 200 nanometers,” he says, “the density can approach 400 GB per square centimeter.”

Another breakthrough in Zhou's research is the stability of the memory cells. Whereas most 1-bit-per-cell molecular memory prototypes retain their data for no more than a few hours, Zhou's nanotube cells can hold on to their data for up to 600 hours. Of course, that's not nearly long enough to be commercially useful, but it's certainly a step in the right direction.

Other potential advantages of these nanotubes include low manufacturing cost (the devices assemble themselves) and much lower power consumption than silicon memory. Still, Zhou believes that practical applications are five to ten years away. “Substantial development work is needed to push the performance even further and to develop a fabrication process amenable for mass production,” he says.

The benefit to electronic musicians is obvious. Flash memory cards with 20 to 200 times the storage capacity of today's cards could easily hold many long tracks of high-resolution audio, lots of high-definition video, and endless MIDI data in a portable device with no moving parts. What more could we possibly want? (Ask me again in 20 years!)