Hard disk systems have recently encounted a storage density ceiling. Most methods in use today have a limit of a few hundred gigabytes per square inch thanks to perpendicular recording. To try to keep storage density rising, scientists have looked at technologies from holographic storage to molecular polymers, but few have made it past the demonstration stage. In a paper in Nature Photonics this week, researchers describe a way to combine two hard drive writing methods to store data at densities of up to one terabit per square inch, and suggest the media could be stable up to ten terabits per square inch.
Each of the two write methods deals with the issue of writing pieces of data very close together without affecting the bits around it. When bits are tightly packed, an effect called superparamagnetism can kick in, where the tiny amount of heat created by the write head will accidentally flip nearby bits and ruin surrounding data. As researchers attempt to pack data bits tighter into a surface, being able to write to an isolated bit without disturbing surrounding ones has become very challenging.
One of the methods, thermally-assisted magnetic recording (TAR), heats an area of a small-grain surface to write it, and then cools the surface once the writing is done. In TAR, the amount of heat, design of the media, and distance between bits keep superparamagetism at bay. The heat also allows the material to magnetize more quickly, reducing the time it takes to write by a small amount.
The other writing method, bit-patterned recording (BPR), writes to a surface that has "magnetic islands" lithographed in. The islands isolate each write event so that superparamagnetic effects can't bleed into other bits.
Each of the methods alone don't contribute huge improvements to data density, and can only get up to two to three hundred gigabits per square inch. As far as practicality goes, TAR has been limited by the availability of the small grained media it needs, materials that stand up to the heating and cooling, and our ability to control the size of the area we heat. BPR, for its part, needs a write head that specifically matches the size of its magnetic data islands.
When BPR and TAR's powers are combined, though, each solves the other's problem. With BPR's magnetic islands, small-grain media is no longer needed, and TAR ensures that only the bit that is heated is written, eliminating the need for a specific size of write head. Together, they form a writing system that can limit bits to tiny areas on inexpensive surfaces, and don't affect surrounding data bits.
The actual device used here routes lasers through a waveguide to a plasmonic antenna that does the writing. When the light reaches the antenna, it is translated into a charge. The antenna is shaped like 'E', with the two outer prongs serving as grounds, and the middle prong acting as a sort of lightning rod to concentrate the surface charge to a small area.
With a middle prong of 20-25 nanometers and each track separated by 24 nanometers, researchers found they could write to areas as small as 15 nanometers in diameter without affecting surrounding data. The efficiency of the signal from waveguide to antenna was about 40 percent, though the overall error rate was low and the system could write at speeds of 250 megabits per second. Researchers were easily able to obtain a density of one terabit per square inch, and posit that densities of 10 terabits per square inch are theoretically possible using this method.
Overall, the authors of the paper note that the high-density data recordings were high-quality, and that the hardware could be suited for use in lithography, biosensors, and nano-manipulation as well.
The next front runner in data storage density and type is far from clear—for example, a method that involved electron quantum holography was able to store 35 bits per electron, and various solid state technologies continue to vie for attention—but this combined bit-pattern and thermally-assisted magnetic recording seems sufficiently close to current hard disk drives to be viable.
Source: ars technica