Researchers Explore Alternatives for Future Data Storage

The density of data stored on magnetic hard disks has been increasing at an annual rate of about 60 percent over the past few years. According to Mark Kryder of Carnegie-Mellon University, who spoke in a Tuesday afternoon session at the March Meeting, areal densities of 10 Gbits/in2 could be achieved by the year 2001 if the present rate continues. Beyond 2001, Kryder believes that giant magnetoresistive materials will eventually be used in read heads to provide an adequate signal at such high recording densities.

However, continued progress depends upon improved recording media. For instance, soft magnetic alloys with high magnetization, such as iron nitride, will be needed for the record heads because of the high coercivity of the media. And to enable head-media separations of the order of 10 to 15 nanometers, hard, wear-resistant media such as barium ferrite and diamond-like carbon will be required.

Researchers at IBM Almaden are working to enhance the ability of the tiny domains on the film medium to become magnetized, a parameter called magnetic anisotropy. Working with sandwiches of atom-thin layers of iron and platinum, Dieter Weller has produced a film with an anisotropy 10 times that of materials used in present day hard disks. "Such unprecedentally large and orientable anisotropies would make these materials attractive for future use in longitudinal or vertical magnetic recording--if the practical aspects of economically mass-producing the media on product disks can be overcome," said Weller.

Hundreds of materials have been tested over the last decade in search of higher anisotropies, with the best results arising from artificial multilayers produced by depositing alternate layers of two or more different materials. However, Weller's new film growth technique essentially eliminates the need for artificial layering, since it produces a natural superlattice structure with the highest possible interface density, and therefore the highest possible anisotropy. Although data storage in the film medium is still improving, scientists continue to explore possible alternate methods. One of these is a holographic system, in which data is stored in the form of an optical interference pattern in a three-dimensional photorefractive crystal.

According to Stanford University's Lambert Hesselink, the advantages of such a system would be its high data-transfer rates, due to the fact that data is written not in linear streams of bits, but in whole two-dimensional "pages", a rapid access time of about 100 microseconds, and a high storage density. "Holography is several orders of magnitude higher in areal density than conventional storage systems," Hesselink explained. "While the surface density is about the same, the thickness is about 10,000 times larger in a holographic system."

While Hesselink has successfully transferred data from a computer to such a holographic system and read the data back again, some obstacles remain. For example, it is difficult to prepare the hologram material and maintain data for long periods. Laptop applications might not be feasible, thanks to the bulky apparatus required for holographic systems, but the "page" data format could lead to such services as video-on-demand.

Another alternative form of data storage would employ the needle of a scanning probe microscope to write and read bits of data in the form of tiny atom piles or pits of about 20 nanometers in size. A typical system would employ arrays of thousands of probe tips, with the capability to write data at rates of 108 bits per second.

According to Gary Gibson of Hewlett Packard, who is leading the development of such a system, scanned probe microscopies have already demonstrated sub-angstrom resolution and the ability to manipulate matter at the atomic level. And recent advances in micromachining techniques may make possible the inexpensive manufacture of such devices. However, Gibson added, a practical product is at least 10 years into the future.