New research brings closer a new type of computer memory that would combine the capacity of a magnetic hard disk with the speed, size, and ruggedness of flash memory.
The storage technology, called racetrack memory, was first proposed in 2004 by Stuart Parkin, a research fellow at IBM’s Almaden Research Center in San Jose, California. Now a team led by Parkin has determined exactly how the bits within a racetrack memory system move under the influence of an electrical current. This knowledge will help engineers ensure that data is stored without overwriting previously stored information.
The new work also helps explain a mystery that surrounded the basic physics of the racetrack memory—whether the bits act like particles with mass, accelerating and decelerating, when moved by electric current. “To further develop racetrack memory, we need to understand the physics that makes it possible,” says Parkin.
In racetrack memory, bits of information are represented by tiny magnetized sections called domain walls along the length of a nanowire. These domain walls can be pushed around—to flip a bit from “0” to “1” or vice versa—when electrical current is applied. Unlike current storage technology, racetrack memory has the potential to store bits in three dimensions, if the nanowires are embedded vertically into a silicon chip. The stored information is read magnetically.
In 2008, the journal Science published a paper coauthored by Parkin that showed how multiple domain walls can traverse the length of a nanowire without being destroyed. The new work, also published in Science, specifies the velocity and acceleration of domain walls as they make their way along a nanowire when an electrical current is applied.
“There’s been debate among theorists about how domain walls will respond,” says Parkin. Researchers understood the motion of domain walls when they were exposed to magnetic fields, but they still had questions about how domain walls move in response to an electrical current—a crucial point because an actual memory device would use electrical current to manipulate bits. One important question was whether domain walls would behave like particles with mass, taking time to speed up and slow down.
The new research shows that they do. It took about 10 nanoseconds, and a distance of a micron, for a domain wall to reach its final velocity, about 140 meters per second. It took another 10 nanoseconds, and one micron, for the domain wall to slow to a stop after the current was turned off. Thus domain walls do indeed behave like particles with mass, and move in a predictable way.
The researchers were careful to use electrical pulses that were just a few nanoseconds in length. If they had used pulses whose duration was similar to the time it takes the domain wall to reach its final velocity, then the acceleration of the domain wall wouldn’t have been measurable.
“It is extremely important to account for these effects in clocking schemes of racetrack memory,” says Shan Wang, professor of materials science and engineering and electrical engineering at Stanford, referring to the algorithms that would control the reading and writing of bits in a racetrack memory device. “Otherwise, domain walls … would be written in wrong locations of the nanowire.”
However, Wang says a practical racetrack memory device remains some way off. “This paper only shows the exquisite shifting of bits, but it is not a memory device yet.”
Still to be determined, for instance, says Peter Fischer, a staff scientist at Lawrence Berkeley National Labs, is “how clean and perfect the device needs to be to work billions of times.”
Parkin says racetrack memory’s reliability is likely to depend on the materials used for the nanowires, and the design—which will be worked out as a prototype is developed. “It shouldn’t take too long,” Parkin says. “Maybe in two years we should have this prototype.”
The dark secret behind those cute AI-generated animal images
Google Brain has revealed its own image-making AI, called Imagen. But don't expect to see anything that isn't wholesome.
The hype around DeepMind’s new AI model misses what’s actually cool about it
Some worry that the chatter about these tools is doing the whole field a disservice.
The walls are closing in on Clearview AI
The controversial face recognition company was just fined $10 million for scraping UK faces from the web. That might not be the end of it.
This horse-riding astronaut is a milestone in AI’s journey to make sense of the world
OpenAI’s latest picture-making AI is amazing—but raises questions about what we mean by intelligence.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.