We're currently going through a transition period where the storage we've relied on for decades is being replaced by much faster solid-state drives. However, one feature neither of these storage technologies enjoys is longevity. A hard drive can last five years under normal use where as an SSD's lifetime is typically much shorter. Optical discs last much longer, but have very limited storage capacity. What if you needed massive amounts of storage that lasts 1,000 years or more?
A Japanese research team working out of Kobe University believes it may have the answer to that very long-term storage problem, and it comes in the form of metal nanodots.
According to Nikkei Technology, the team, led by Makoto Nagata and Noriyuki Miura, positioned metal nanodots on a silicon wafer. Each nanodot is used to form lines of bits by placing them at specific locations on the wafer to represent a binary 0 or 1. Once the data has been laid out, the wafer is sealed using an insulating film. Reading of the data is then done wirelessly.
As this storage format uses existing semiconductor manufacturing techniques, it can easily be scaled down. The test chip created by the team used a 180nm CMOS process and offered 10 gigabits of storage per square inch. However, upgrading to a 14nm process, which is now common for processor manufacturing, sees the data density offered increase to one terabit per square inch, which is roughly equivalent to today's hard drives. The recording layers can also be stacked so as to increase the storage space within each sealed wafer. The test chip included four recording layers.
Related
- Seagate's 1TB Duet Hard Drive Includes Unlimited Amazon BackupsSeagate's 1TB Duet Hard Drive Includes Unlimited Amazon Backups
Reading speeds on the test chip were limited to 40kb per second, but this should also increase drastically along with the storage capacity.
As for how long such a storage format lasts, the combination of metal nanodots and silicon wafer in a sealed unit means that even after 1,000 years each chip is expected to be viable. This was tested using a pressure cooker and accelerated stressing, which allows one hour to represent each year of life. After 1,000 hours, the test chip still worked flawlessly.
With further research I expect the storage density and data read speeds to increase significantly, and they should continue to do so as manufacturing lines embrace ever smaller nanometer processes. The one question that then remains is whether the people of 3017 will have the means to continue reading these, by then, ancient chips.