Physicists Discover Compelling Evidence for Magnetically Bound Excitons

In art, the negative space in a painting can be just as important as the painting itself. Something similar is true in insulating materials, where the empty spaces left behind by missing electrons play a crucial role in determining the material’s properties. When a negatively charged electron is excited by light, it leaves behind a positive hole. Because the hole and the electron are oppositely charged, they are attracted to each other and form a bond. The resulting pair, which is short lived, is known as an exciton [pronounced exit-tawn].

Excitons are a key part of many technologies, including solar panels, photodetectors and sensors, as well as light-emitting diodes found in televisions and digital display screens. In most cases, the exciton pairs are bound by electrical, or electrostatic, forces, also known as Coulomb interactions. Now, in a new study in Nature Physics, Caltech researchers report detecting excitons that are not bound via Coulomb forces but rather by magnetism. This is the first experiment to detect how these so-called Hubbard excitons, named after the late physicist John Hubbard, form in real-time.

“Using an advanced spectroscopic probe, we were able to observe in real time the generation and decay of magnetically bound excitons, the Hubbard excitons,” says study lead author Omar Mehio (PhD ’23), a recent graduate student at Caltech who worked with David Hsieh, the Donald A. Glaser Professor of Physics at Caltech. Mehio is now a postdoctoral fellow at the Kavli Institute at Cornell.

“In most insulators, oppositely charged electrons and holes interact with one another just as an electron and a proton bind to form a hydrogen atom,” Mehio explains. “However, in a special class of materials known as Mott insulators, the photo-excited electrons and holes instead bind through magnetic interactions.”

The results could have applications in the development of new exciton-related technologies, or excitonics, in which the excitons would be manipulated through their magnetic properties. “Hubbard excitons and their magnetic binding mechanism demonstrate a drastic departure from the paradigms of traditional excitonics, creating the opportunity to develop a whole ecosystem of novel technologies that are fundamentally unavailable in conventional excitonic systems,” Mehio says. “Having excitons and magnetism strongly intertwined in a single material could lead to new technologies that harness both properties.”

To create the Hubbard excitons, the researchers applied light to a type of insulating material known as an antiferromagnetic Mott insulator. These are magnetic materials in which the electron spins are aligned in a repeating, stable pattern. The light excites the electrons, which jump to other atoms, leaving holes behind.

“In these materials, when an electron or hole moves through the lattice, they leave in their wake a string of magnetic excitations,” Mehio says. “Imagine you tie one end of an elastic rope around your friend, and the other end around yourself. If your friend runs away from you, you will feel the rope pull you in that direction and you will begin to follow. This scenario is analogous to what happens between a photo-excited electron and the hole it leaves behind in a Mott insulator. With Hubbard excitons, the string of magnetic excitations between the pair serves the same role as the rope connecting you to your friend.”

To demonstrate the existence of the Hubbard excitons, the researchers used a method called ultrafast time-domain terahertz spectroscopy, which allowed them to look for the very short-lived signatures of the excitons at very low-energy scales. “Excitons are unstable because the electrons want to go back into the holes,” Hsieh explains. “We have a way of probing the short time window before this recombination occurs, and that allowed us to see that a fluid of Hubbard excitons is transiently stabilized.”

The study, titled “A Hubbard exciton fluid in a photo-doped antiferromagnetic Mott insulator,” was funded by the Army Research Office, the David and Lucile Packard Foundation, the National Science Foundation, Caltech’s Institute for Quantum Information and Matter (an NSF Physics Frontiers Center), Caltech, the German Research Foundation, the Gordon and Betty Moore Foundation, and the Slovenian Research Agency. Other authors include Xinwei Li, Honglie Ning (PhD ’23), and Nicholas Laurita, all formerly of Caltech; Caltech graduate student Yuchen Han; Zala Lenarčič of the Jozef Stefan Institute in Slovenia and UC Berkeley; Michael Buchhold of the University of Cologne in Germany (and a former Caltech postdoc); and Zach Porter and Stephen Wilson of UC Santa Barbara.