One method of material memory storage is called return-point memory, which functions similarly to how a combination lock operates. "With a lock, rotating the dial clockwise and counterclockwise in a particular sequence yields a result - the lock opening - that depends on how the dial was moved," explained Nathan Keim, associate professor of physics at Penn State's Eberly College of Science, who led the research team. In the case of materials with return-point memory, alternating between positive and negative deformations can leave a lasting imprint that can either be read or erased.
Keim also noted that the mathematical principles governing this form of memory could apply to a range of systems, from the magnetization of computer hard drives to the damage caused to solid rock. His team had previously demonstrated that these same principles could describe the memory patterns found in disordered solids, where particles may seem randomly arranged but actually contain information about previous forces applied to them.
Return-point memory typically requires alternating forces, such as reversing a magnetic field or pulling a material back and forth. "The mathematical theorems for return-point memory say that we can't store a sequence if we only have this 'asymmetrical' driving in one direction," Keim explained. "If the combination lock dial can't go past zero when turning counterclockwise, it only stores one number in the combination. But we found a special case when this kind of asymmetrical driving can, in fact, encode a sequence."
The research team used computer simulations to test the conditions under which a sequence could be encoded in a material. They varied parameters such as the magnitude and orientation of the external forces, and the way these forces were applied, to examine their impact on memory formation. To better understand how these factors influenced memory storage, they simplified the system's components, like the particles in a solid or the domains in a magnet, into abstract units called "hysterons."
"Hysterons are elements of a system that may not immediately respond to external conditions and can remain in a previous state," said Travis Jalowiec, an undergraduate at the time of the research, who later graduated with a physics degree from Penn State. "Like how parts of a combination lock reflect the previous positions of the dial, and not where the dial is now." The hysterons, Jalowiec noted, could interact in either a cooperative manner, where a change in one triggers a change in the other, or in a non-cooperative, "frustrated" way, where one change prevents another.
Frustrated hysterons, which occur when the hysterons resist one another, were identified as essential for encoding and retrieving sequences in systems driven asymmetrically. Keim used the analogy of a bendy straw with collapsible bellows to explain this concept. "If you pull on the ends of the straw a tiny amount and stop, one will pop open, and it being open means that the others do not. The change in one relieves the stress in the system."
While cooperative hysterons required symmetrical driving (alternating forces) to store sequences, the presence of just one pair of frustrated hysterons enabled a system to store a sequence despite asymmetric forces. "Finding a pair of frustrated hysterons in a real material has been elusive," said Keim. "It's hard to observe because often the signature of frustration is that something doesn't happen. The behavior we found is rare, but it would stand out like a sore thumb in a real material, so it gives us a new way to look for and study materials with frustration."
The researchers are optimistic that their findings could lead to the development of artificial systems capable of such memory storage, starting with simple mechanical systems and advancing toward more complex ones, such as asymmetrical combination locks. These findings could also inspire the creation of new methods to store, recall, and erase information in mechanical systems.
Keim further explained that the memory characteristic they discovered could store both the largest and most recent deformation. "If you can make a system that stores a sequence of memories, you can use it like a combination lock to verify a specific history or recover diagnostic or forensic information about the past," he said. "There is increasing interest in mechanical systems that sense their environments, perform computations, and respond or adapt without ever using electricity. A better understanding of memory expands these possibilities."
In addition to Keim and Jalowiec, the research team includes Chloe Lindeman, a graduate student at the University of Chicago at the time of the research, now a Miller Postdoctoral Fellow at Johns Hopkins University. Funding for this work came from the U.S. Department of Energy, Penn State Schreyer Honors College, and the Penn State Student Engagement Network.
Research Report:Generalizing Multiple Memories from a Single Drive: The Hysteron Latch
Related Links
Penn State
Space Technology News - Applications and Research
| Subscribe Free To Our Daily Newsletters |
| Subscribe Free To Our Daily Newsletters |
