Print 20 comment(s) - last by Clauzii.. on Jul 3 at 5:04 PM

A scanning electron microscope image of the germanium compound nanowire. The arrow shows the location of the more complex core.  (Source:
Pour in that data. A new nanowire technology could enable very high density storage devices.

Nanowires and other microscopic particles like them are on the forefront of science these days. Various nanowire constructs have been covered at DailyTech for their properties and the potential they may have to a multitude of technologies. Engineers at the University of Pennsylvania's School of Engineering and Applied Science have taken nanowire research into the direction of physical data storage.

While consumers have seen the price of data storage media of all kinds drop while storage capacity and densities have risen, one thing remains constant for all bit-based mechanisms: binary 0s and 1s represent data. They may be stored and interpreted by different means, but binary is the system that all our personal electronics use from data storage to volatile memory to processing itself.

One way to further increase density could be to add a third state, or a 2. UPenn's nanowire storage medium does just that. The wire itself is a coaxial system, like the cable that carries television into your cable box. The nanowire's shell is composed of germanium telluride (GeTe), while the core is a more complex germanium/antimony/tellurium compound Ge2Sb2Te5 .

Both of these materials are known as phase-change materials. Under the stimulus of an electric field, the materials change from a crystalline, ordered structure to an amorphous, unordered. To supplement this, the core and shell can be separately modified from crystalline to amorphous.

To make this work as a data storage device, picture the crystalline state to be a 0 and the amorphous to be a 1. When the compounds are in a crystalline state, they have a very low resistance to electricity thanks to their crystalline structure. When an electric pulse is applied, the compound heads and becomes amorphous, greatly increasing its resistance to current flow. In this way, measuring the resistance of the nanowire can result in either a 0 or 1.

Where the magic happens is when the shell and core are separately tuned, one crystalline while the other is amorphous. This creates a third level of resistance over the nanowire – the 2.

In addition to a third readable state, UPenn's nanowires have other properties which make them ideal for volatile memory storage. Due to the third state itself, densities become much greater. This could either enable smaller memory devices for portable electronics, or much more storage in current form factors.

Such tiny structures have been known to self-assemble. A bottom up assembly would revolutionize memory production which typically relies on a top down approach. Rather than etching circuits into various materials, the nanotubes could be coaxed to assemble themselves into usable structures. Combined with a crystal's tendency to lack defects, this could enable entirely new production methods which involve less outside manipulation, cutting cost and loss simultaneously.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

But what about...
By Cheapshot on 7/3/2008 10:34:09 AM , Rating: 2
the speed increase? If these are used in Solid state will there be anymore potential for transfer/ read, or write rates?

RE: But what about...
By Digimonkey on 7/3/2008 11:15:53 AM , Rating: 2
Technically yes. More than likely if this technology were going to come out soon...within 10 years, it would probably be paired up with a controller chip to interpret 2's as a 10 or a 01. If the 2 was used often enough, the memory controller could technically access more data even though the width of the bus would remain 32 or 64 bit wide.

RE: But what about...
By masher2 (blog) on 7/3/2008 1:36:28 PM , Rating: 3
That's not quite how it would work. Rather, they would simply encode the entire value in ternary, meaning they'd have to use less bits overall. For instance, a 32-bit binary value can be stored in 21 ternary bits.

RE: But what about...
By Digimonkey on 7/3/2008 2:03:13 PM , Rating: 2
But the processor would have to be designed with the logic to handle ternary bits, along with software having to be rewritten. I was just thinking in intermediate terms.

RE: But what about...
By masher2 (blog) on 7/3/2008 2:24:23 PM , Rating: 3
The processor? just need a little glue logic to convert binary to trinary in whatever controller chip runs the storage system.

RE: But what about...
By Digimonkey on 7/3/2008 3:01:04 PM , Rating: 2
That's true, I was wrong. You'd still need multiples of 7-bits however if my math is correct. It'd take 7-bits with 3 states to represent the number of 255. so we'd access 28 bits and 56 bits at a time. Correct me if I'm wrong.

RE: But what about...
By Digimonkey on 7/3/2008 3:18:36 PM , Rating: 2
Nevermind, my maths wrong too.

RE: But what about...
By Etsp on 7/3/2008 2:27:45 PM , Rating: 2
No, not really, there would simply be some control components that would send the data to the bus as standard binary, there is no need for the processor to know that ternary is being used at the physical level, as the components required to read the data would be able to convert it to binary relatively easily. Data stored in Binary, Ternary, Decimal, and Hexadecimal is all the same data. The only difference is the number of symbolic states used to represent the data.

"There's no chance that the iPhone is going to get any significant market share. No chance." -- Microsoft CEO Steve Ballmer

Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki