Arrays of atomic force probe tips are promising nanotech approaches to denser, faster, cheaper memories (see this post from nine months ago). James Tyrrell of nanotechweb.org (requires free registration) explores the latest progress from the IBM “millipede project” and looks at what yet needs to be done. The researchers have completed a fully functional prototype system. They have further demonstrated accurate control of tip position at a scanning speed three to four orders of magnitude faster than a typical scanning probe microscope, and the ability to write more than 108 indents at a density of 1 Tb per square inch with a single tip. From “Nanoindenter passes memory test“
Scanning probe-based data storage is back in the news as IBM researchers turn their attention to the thermo-mechanical indentation of polymer media at high patterning densities and fast transfer rates.
Results published in the journal Nanotechnology [abstract] suggest that there could be more mileage in polymer systems than experts first thought and nanotechweb.org interviews Urs Dürig of IBM’s Zurich Research Laboratory, Switzerland, to find out more.
What are the attractions of polymer probe technology compared with other data storage schemes?
Polymer probe technology offers the potential of extremely high storage density, up to 4 Tb/inch2. The basic principle of hot-embossing the information in the form of indents is conceptually simple and lends itself well to a highly parallel MEMS implementation. In this way, we address several critical issues. The data rate can be boosted by two to three orders of magnitude by operating a corresponding number of indenters in parallel. Small size also means low power consumption, which is critical in particular for portable applications. Small size also means robustness of the device against shock and vibration. Finally, small size also means fast positioning, which translates to worst case seek/access times in the order 1 ms in comparison with 10 ms for hard disk drives.
The researchers note that even with this impressive progress, a “millipede” type memory might not be competitive with flash memory. Regardless, the techniques for rapid and precise control of arrays of AFM tips might be of use to proposals for mechanosynthesis with AFM (see this post from two weeks ago and this one from three weeks ago).