The hard drive in your computer stores information in tiny regions that can be flipped from one magnetic polarity to the other to represent the zeroes and ones of binary data. The speed of such devices depends on how quickly these “bits” can be flipped, or “rewritten,” and a team of physicists has now shown experimentally that writing speeds might be made some 30 times faster by fine-tuning the magnetic geometry of the storage device. The demonstration may not lead immediately to faster memories but gives researchers a promising pathway to pursue them.
In most magnetic materials, each atom has a spin that acts like a microscopic magnet. If all these spins point in the same direction, the pattern is called ferromagnetic order, as found, for example, inside an ordinary iron magnet. But magnetic order can be more subtle, with the direction of spins varying systematically across the atomic lattice. Atomic spins in a line can alternately point up and down, for example, or they can gradually rotate to trace out a helix as you move across the lattice. These and other configurations produce no large-scale magnetism because every spin is cancelled out by another. This kind of pattern is known as antiferromagnetic order.
The bits of modern magnetic memories are typically small domains of ferromagnetic order. Flipping a bit requires simultaneously reorienting many atomic spins, limiting the speed of information rewriting. Physicists have long suspected that it should be easier to switch a region of antiferromagnetic order into a non-antiferromagnetic state, because you can use light to knock spin-carrying electrons from one atom to a neighbor, disrupting the geometric pattern. In principle, ones and zeros could be represented by the two states “antiferromagnetic order” and “non-antiferromagnetic order.” Experiments to test the response times of ferromagnetic and antiferromagnetic regions are challenging, however, because most materials have exclusively one type of order, and other physical differences obscure the comparison of magnetic properties.
To overcome this problem, Christian Schüssler-Langeheine of the Berlin Helmholtz Center for Materials and Energy and his colleagues ran experiments using a thin film of the element dysprosium, which is ferromagnetic at temperatures less than 87 K and antiferromagnetic at higher temperatures. “We measure how both kinds of order change in the very same sample,” says team member Nele Thielemann-Kühn, who is now at the Free University of Berlin. “All we had to do was to change the sample temperature.”
For a direct comparison of the switching speeds of ferromagnetic and antiferromagnetic regions, the team studied the rate for disturbing the order in the two cases, rather than looking explicitly at the flipping of spins. They used pulses of infrared laser light to disturb the magnetic order of dysprosium samples by nudging electrons responsible for the magnetism from one atom to another. To monitor how quickly the magnetic order changed, they then scanned the sample with pulses of x rays, whose scattering depends on the magnetic state of dysprosium. The researchers found that the antiferromagnetic zones lost their magnetic order around 30 times faster than the ferromagnetic zones and required much less laser energy to trigger the change.
The difference, says Thielemann-Kühn, can be explained by thinking about the angular momentum associated with each atomic spin. In a ferromagnetic zone, parallel spins add up to create a large total angular momentum. Moving a spin from one atom to another doesn’t affect that total and therefore doesn’t disrupt the ferromagnetic order. Disrupting ferromagnetism requires transporting the angular momentum away from the spin system into other parts of the surrounding environment. In contrast, an antiferromagnetic zone possesses no net angular momentum because the contributions from different atoms cancel out. The magnetic order can be erased merely by shuffling the atomic spins and destroying their delicate spatial arrangement.
“This experiment is very elegant,” says physicist Stefan Eisebitt of the Max Born Institute in Berlin. “It’s also important, as it shows that optical demagnetization can proceed faster if there’s no need to transport angular momentum away.”
“Our work is mainly fundamental research,” says Thielemann-Kühn, but she believes that their results might lead to speedier and more efficient magnetic devices. Although antiferromagnetic materials lack the clear distinction between opposite polarities that make ferromagnetic memories so convenient, she says that other researchers have been working to find ways to store bits of information in antiferromagnetic zones, which could in principle be flipped much faster. Another possibility might be to improve the switching speed of conventional ferromagnetic bits by placing them in close contact with anti ferromagnetic materials that could provide a pathway for moving angular momentum in or out.