Key takeaways

  • If computers used ripples in magnetic fields, called magnons, to encode and process information, the result would be devices with potential memory speed on the order of billionths of a second.
  • UCLA researchers and partners caused two distinct types of magnons to interact so that output is not directly proportional to input — a crucial step toward computing advances.
  • This multi-institution, long-term research collaboration is studying magnons using a rarely used but promising terahertz laser technology.

One vision for the future of computing involves using ripples in magnetic fields — called magnons — as a basic mechanism. In this application, magnons would be comparable to electricity as the basis for electronics.

In conventional digital technologies, such magnonic systems are expected to be far faster than today’s technologies, from laptops and smartphones to telecommunications. In quantum computing, the advantages of magnonics could include not only quicker speeds but also more stable devices.

A recent study in the journal Nature Physics reports an early-stage discovery along the path to developing magnonic computers. The researchers caused two distinct types of ripples in the magnetic field of a thin plate of alloy, measured the results and showed that the magnons interacted in a nonlinear manner. “Nonlinear” refers to output that is not directly proportional to input — a necessity for any sort of computing application.

To date, most research in this area has focused on one type of magnon at a time, under relatively stable conditions described as equilibrium. Manipulating the magnons, as done in these studies, pushes the system out of equilibrium.

This is one of many investigations underway through a multiyear collaboration between theorists and experimentalists from multiple fields of science and engineering, including a second study that recently appeared in Nature Physics. The project, supported by government and private grantors, brings together researchers from UCLA, MIT, the University of Texas at Austin and the University of Tokyo in Japan.

“With our colleagues, we’ve started what I would call a campaign to spur progress in nonequilibrium physics,” said Prineha Narang, a co-author of the study and professor of physical sciences in UCLA College. “What we’ve done here fundamentally advances the understanding of nonequilibrium and nonlinear phenomena. And it could be a step toward computer memory using ultrafast phenomena that happen on the order of billionths of a second.”

One key technology behind these findings is an advanced technique for adding energy to and evaluating samples using lasers with frequencies in the terahertz range, which sits between the wavelengths of microwave and infrared radiation. Adopted from chemistry and medical imaging, the method is applied only rarely to study magnetic fields.

According to Narang, who is a member of the California NanoSystems Institute at UCLA, the use of terahertz lasers suggests potential synergy with a technology growing in maturity.

“Terahertz technology itself has reached the point where we can talk about a second technology that relies on it,” she said. “It makes sense to do this type of nonlinear control in a band where we have lasers and detectors that can be put on a chip. Now is the time to really push forward because we have both the technology and an interesting theoretical framework for looking at interactions among magnons.”

The researchers applied laser pulses to a 2-millimeter-thick plate made from a carefully chosen alloy containing yttrium, a metal found in LEDs and radar technology. In some experiments, a second terahertz laser was used in a coordinated way that paradoxically added energy but helped stabilize samples. 

A magnetic field was applied to the yttrium in a specific fashion that allowed for only two types of magnon. The investigators were able to drive either type of magnon individually or both at the same time by rotating the sample to certain angles relative to the lasers. They were able to measure the interactions between the two types and found that they could cause nonlinear responses.

“Clearly demonstrating this nonlinear interaction would be important for any sort of application based on signal processing,” said co-author Jonathan Curtis, a UCLA postdoctoral researcher in the NarangLab. “Mixing signals like this could allow us to convert between different magnetic inputs and outputs, which is what you need for a device that relies on manipulating information magnetically.”

Narang said that trainees are vital to the current study, as well as the larger project.

“This is a really hard, multiyear endeavor with a lot of pieces,” she said. “What’s the right system and how do we go about working with it? How do we think about making predictions? How do we limit the system so it’s behaving as we want it to? We wouldn’t be able to do this without talented students and postdocs.”

The study includes MIT chemistry professor Keith Nelson and UT Austin physics professor Edoardo Baldini, along with the UCLA team led by Narang, which was supported by the Quantum Science Center, a Department of Energy National Quantum Information Science Research Center headquartered at Oak Ridge National Laboratory. The study was primarily supported by the Department of Energy as well as the Alexander von Humboldt Foundation, the Gordon and Betty Moore Foundation, the John Simon Guggenheim Memorial Foundation and the Japan Society for the Promotion of Science — all of which provide ongoing support for the collaboration.