IBM Shrinks Computer Memory Into Only Twelve Atoms..!

Scanning tunneling microscope image of
twelve iron atoms that were assembled
into an atomically precise
antiferromagnet.
(Credit: IBM)

IBM announced last week that it had successfully created the smallest magnetic memory storage yet – they were able to record a bit into just 12 iron atoms. The results of this research can be found in the current issue of Science.

“It begins with a simple question,How small can you make a magnetic structure, and still act as a classic magnet for data storage?” - IBM researcher Andreas Heinrich. 

The traditional way to approach this is the simple Moore’s Law approach – starting from the top down, keep shrinking storage. The other approach, which is what Heinrich’s team is working on, is to actually build the magnet up, atom by atom.

They did this by using a scanning, tunneling electron microscope, placing iron atoms one at a time on a copper-nitrogen surface. By placing the atoms one at a time, the researchers could start to understand how small it was possible to make the storage of one bit.

As it turns out, the absolute limit is 12 atoms – a magnetic storage unit for data just can’t get any smaller. That’s because when there are fewer than 12 atoms, quantum effects dominate. For example, a structure of six iron atoms switches quantum spin states about 1,000 times per second. At eight atoms, it was significantly less – about once per second. But at 12 atoms, there was no quantum switching at all. In other words, as long as you have at least 12 atoms, the iron structure will behave like a classical magnet, and it’s possible to store a single bit of information in the structure.

Below you can see eight of the 12-atom structures, each storing one bit. This is one byte of information in only 96 atoms.

Another challenge of building on such a small scale is the ferromagnetic properties of most magnetic storage. In a ferromagnetic piece of metal, like a compass needle, all of the atoms have their magnetic moments aligned, producing a large magnetic field that can be sent across large distances. While this works great for storing bits in a hard drive, “it’s bad on the nanoscale,” according to Heinrich. That’s because the atoms are packed so closely together. If the atoms have the same magnetic moment, they’ll repel each other.

Instead, the atoms in this memory system are anti-ferromagnetic. In other words, says Heinrich, “One points north, then south, then north. Locally, the magnetic fields cancel, which means they can be packed closely together. Bits are literally next to each other on the atomic scale.”

Of course, this is just the first step in a long road to practical application. Building bits like this are expensive. But Heinrich is optimistic about the potential. He expects practical applications in the area of spintronics within the next 5-10 years, and is also excited about the possibility of applying his methods to computation – possibly even quantum computation. “What I’m going to be doing is focusing on whether quantum effect can be enhanced to do quantum computation using the scanning, tunneling microscope,” he said. His team is looking into different materials to apply potential quantum computing applications to, including silicon.

It’ll be interesting to see what they come up with.