Magnetic media has been the mainstay of computer storage for decades. Just as with processors, shrinking feature size—smaller clusters of magnetic atoms—have allowed huge gains in storage density. Just as with processors, though, these gains are starting to push up against physical limits, as it's getting harder and harder to set the magnetic state of a cluster of atoms without wiping out the information on the neighboring clusters.
Now, researchers at IBM have teamed up with collaborators in Germany and Switzerland to store information using a related phenomenon, antiferromagnetism. And they've shown that it's possible to store a bit in a feature that contains as few as six iron atoms. The downside is that the storage was only stable at extremely low temperatures. If the sample was allowed to heat up to 5K, the information on the bits vanished.
The magnetism we're familiar with from things being stuck on the side of the fridge comes from materials where the magnetic moments of the atoms in the material are all aligned and pointed in the same direction. Antiferromagnetic materials also have aligned magnetic moments, but they exist in alternating stripes. You can think of this in terms of a set of bar magnets, arranged so that the north end of one was always next to the south end of its two closest neighbors. The net result is that, even though the magnetic moments of the material are highly ordered, the sum of those moments is zero, meaning it doesn't generate any external magnetic field.
That makes it rather difficult to read the state of an antiferromagnet with standard magnetic hardware. But since the magnetic moments vary on the atomic level, the researchers reasoned, it should be possible to read them with hardware that can read individual atomic states. To do that, they polarized the tip of a scanning-tunneling microscope by applying a six Tesla external field.
Small rows of iron atoms, they found, spontaneously formed an antiferromagnet. The first atom in the row would adopt a specific orientation, and then the following atoms would take on alternating states. (So, either up-down-up-down... or down-up-down-up...). Applying a larger voltage to the STM tip was sufficient to flip the orientation, changing the first atom from down to up, and forcing the rest of the row to flip accordingly.
The more neighbors an atom had, the more stable its state would be. A single row of atoms would switch states spontaneously at a fairly high rate. Adding a second row of atoms would extend the lifetime, as would making the row of atoms longer. The researchers settled on a cluster of 12 atoms, arranged in two rows of six. At the temperatures the experiments were performed, these clusters retained their states for at least 17 hours.
The combination of stability, a change in orientation on demand, and a readable state (first atom up or down) are all you need for magnetic memory. So, the authors set up an array of 96 iron atoms to form an 8-bit device, with each bit composed of two rows of six atoms. The geometry enables the authors to place each bit in a way that its state is very unlikely to bleed over into the surrounding ones and, in fact, acts to stabilize their state.
Using the STM tip, they were able to read the states in each bit, and write new information into the bits on demand. The density of information there (or, put another way, the atoms needed to store a bit) is 100 times higher than current hard drive technology.
But don't expect to see this in your desktop any time soon. Scanning-tunneling microscopes haven't exactly been miniaturized yet, and the temperatures at which the bits are stable are extremely low, less than 5K. So unless you want your computer hooked up to a high-end piece of refrigeration hardware, that presents a serious issue as well. (Although overclockers would probably relish the chance to hook up a heat sink to the same hardware.)
So, what to think of the paper's last sentence? It concludes by stating, "Our results demonstrate that switchable nanoscale antiferromagnets are candidates for future memory, storage, and spintronic applications." That candidacy depends on two things: getting atomic-level precision in a miniaturized device, and getting antiferromagnetic materials to retain their state at room temperature. Both of those are very large leaps over the technology demonstrated in this paper.
Science, 2012. DOI: 10.1126/science.1214131 (About DOIs).
Listing image by Photograph by IBM
reader comments
18