An experimental "ferroelectric" data-recording device crams 4 trillion bits onto every square inch. This all-electrical method leaves magnetic and flash memory devices scrambling to catch up.

The device is the work of Dr. Yasuo Cho and Kenkou Tanaka at Japan's Tohuku University. The device is built from a ferroelectric material, which is a material where the electrical polarization can be changed by applying an external electrical field. A microscopic stylus rides on the surface of the material, making contact with the countless tiny spots on the structure below. To write data, a pulse goes through the stylus and onto one of these spots, changing the spot's polarization and thus allowing for binary data recording.

These electrically charged 0's and 1's can add up to a lot of data because there are so many spots on the material. Cho and Tanaka were able to place each bit just 12.8 nanometers apart, which adds up to about 4 trillion on each square inch of the material. That far exceeds the highest data density levels that even the most advanced magnetic disk drives can currently attain.


Dr. Cho explains what he sees as the future of the device:

"We expect this ferroelectric data storage system to be a candidate to succeed magnetic hard disk drives or flash memory, at least in applications for which extremely high data density and small physical volume is required."

The engineers had to overcome some technical challenges to accomplish this record-breaking data density. When a lot of bits had to be written on consecutive spots on the material, the polarized regions would expand into each other and coalesce into a single, indistinct region. This made the bits themselves indistinct, and harmed the data storage. They worked around this by figuring out how to anticipate such consecutive data-recording and lower the writing voltage so that the polarization was not quite strong enough to bleed over onto the neighboring spots.


Although the fact that this drive is completely electric means it can avoid some of the natural pitfalls of magnetic or thermal-based data drives, the drive isn't quite ready for preorder just yet. Cho and Tanaka say they need to greatly increase the speed of both reading and writing data, and accuracy is still an issue - the bit error rate is low at 1.2%, but not nearly low enough for commercial sale. They also need to find a way to develop the ferroelectric material for a much lower cost than the materials they're using the lab.

And then, of course, there's the ongoing, exponential increase of data storage elsewhere. Cho and Tanaka have quite possibly already tapped the upper limit of what electrical data storage can do, but magnetic storage might still have plenty of room to go. The disk drive maker Seagate, for instance, has said it's technically feasible to one day cram 50 trillion bits per square inch, a data density that would make this ferroelectric achievement look positively quaint by comparison.

[via Applied Physics Letters]