Researchers use light to generate nanoscale magnetization

Using an engineered metasurface that traps light, Cornell researchers have demonstrated a new way to generate strong static magnetic fields without using external magnets or magnetic materials - an approach that could advance spintronics, quantum and photonic computing, and data storage.

In an article published in Advanced Science, Shivaksh Rawat, a Ph.D. candidate working with Gennady Shvets, the J. Preston Levis Professor of Engineering in the School of Applied and Engineering Physics, and Samyobrata Mukherjee, a postdoc in the same group, described how a new technique for light manipulation - the so-called "time interface" - can be used to convert part of an optical wave's energy into static magnetization.

When a light wave experiences a spatial interface - for example, when it travels through air and then water - some portion of the light gets reflected off the surface while the rest is transmitted through it. Similarly, when a light wave experiences a time interface - a sudden change in the optical properties of the propagation medium, such as an increase or decrease in the refractive index - it also produces reflected and transmitted waves. However, a time interface can also excite a static zero-frequency mode of the system; in other words, stopping part of the light's rapidly varying magnetic field and converting it into a stationary magnetic field pattern that remains in place instead of continuing to oscillate.

To create a time interface, the researchers utilized a 2D metasurface - a carefully engineered rectangular array of germanium nanostructures designed for trapping mid-infrared light. The researchers wondered what would happen if the metasurface is illuminated with an intense, short burst of higher-energy (near-infrared) photons while the mid-infrared light is still inside the structure. Their modeling revealed that the near-infrared light released electrons from the germanium atoms, leaving "electron holes" and many free electrons. The rapid generation of electron-hole pairs by the near-infrared light can create a time-dependent change in the light's refractive index that acts as a time interface for the trapped mid-infrared light.

"We used an approach known as localized free carrier generation, which has advantages over other methods of nanoscale magnetization," Rawat said. "One of the important contributions of our work is that our approach is material agnostic. Any non-metallic surface will work."

During the time interface, some of the initial energy in the trapped light shifts to new red-shifted light waves, while the rest is trapped as the kinetic energy of circulating free electrons, which support current loops that sustain a persistent magnetic field in the hot spots. In the absence of losses, this magnetization would persist indefinitely, but even under normal conditions, the magnetic field lasted about 300 femtoseconds, or roughly three 10-trillionths of a second. While that seems extraordinarily short, it equals about 20 cycles of the mid-infrared light wave that created it.

"Our work demonstrates a fundamentally new method for creating strong magnetic fields using light rather than conventional magnets," Rawat said. "By rapidly changing the optical properties of an engineered metasurface, we were able to convert part of a passing light wave into a localized magnetic field that remained after the light had passed. We used a more efficient and highly localized approach than many existing methods, and our work helps explain how energy is redistributed in rapidly changing optical materials."

By understanding how light can both transmit information and create magnetism, the work could open new research directions at the intersection of photonics and magnetism. Beyond advancing time-varying photonics, the research could have important implications for spintronics, magnetic data storage, photonic and quantum computing, and precise control of magnetic environments.

This work was supported by the University of Dayton Research Institute and by the Office of Naval Research, and the Army Research Office. Portions of the research were performed using computer resources of the Cornell University Laboratory of Plasma Studies, with assistance from Steven Lantz in the Cornell Center for Advanced Computing.