Keio Research Highlights

The manipulation of magnetic moments has been a subject of human curiosity. More than 200 years ago, Hans Christian Oersted discovered that a charge current, a flow of electrons' charge, can manipulate nearby magnetic compass needles―so-called "magnetic moments". This pioneering work connects electricity and magnetism, providing the foundation of classical electromagnetism. This phenomenon stems from the fact that a charge current produces a magnetic field, which couples the charge current and magnetic moments indirectly.

A breakthrough in the manipulation of magnetic moments was achieved with the prediction of spin torques. In 1996, John Slonczewski and Luc Berger independently predicted that an injection of a spin current, a flow of electrons' spin, into a magnet enables the manipulation of the magnetic moment through the transfer of angular momentum. The transfer of angular momentum from the spin current to the magnetic moment exerts a torque ― a spin torque ― on the magnetic moment, enabling the manipulation of the magnetic moment without using magnetic fields. Since its prediction, the spin torque has attracted substantial attention and ushered in a new era of spintronics, which aims to explore novel phenomena originating from the spin degree of freedom of electrons. The manipulation of magnetic moments using spin currents and spin torques offers a plethora of applications, such as non-volatile magnetic memories, nanoscale microwave sources, and neuromorphic computing devices. In the past two decades, spintronics based on spin currents and spin torques has evolved into a central part of condensed matter physics, materials science, and device physics.

Although charge and spin currents have been the mainstay of the physics and technology of solid-state devices, an electron has yet another property, called the orbital degree of freedom. The orbital degree of freedom is known to be crucial for the equilibrium properties of solids. Corresponding to the spin and orbital degrees of freedom, electrons in solids can carry both spin and orbital angular momentum. Hence, it would be natural to expect the existence of the orbital counterpart of charge and spin currents: orbital currents, or the flow of orbital angular momentum (see Fig. 1). Theoretically, orbital currents have played an important role in understanding nonequilibrium phenomena in magnetic heterostructures; orbital transport has been believed to be more fundamental than spin transport. The physics behind the fundamental role of the orbital dynamics is that in solids, the orbital degree of freedom mediates the coupling between the electrons and lattice, and thus the angular momentum of the lattice can be transferred to the orbital part of the electrons, producing a variety of nonequilibrium phenomena driven by angular momentum transfer. In the presence of spin-orbit coupling ― a relativistic effect that couples the spin and orbital degrees of freedom ― nonequilibrium angular momentum of the orbital part can further be transferred to the spin part of the electrons.

Despite the significance of orbital dynamics, observing orbital currents in experiments has proven difficult. In this work, we experimentally explore orbital transport. Our measurements have revealed that even light metals without strong spin-orbit coupling can generate sizable torques acting on magnetic moments in adjacent magnets. This is clearly different from the prediction of spin torques because spin-orbit coupling plays a crucial role in generating spin currents and spin torques. We also found that the magnitude of the torque is quite sensitive to what magnet is used. Furthermore, the flow of angular momentum responsible for torque generation propagates much longer than the spin decay length of the magnet. These distinct features provide evidence that torque originates from orbital currents.

Although spin and orbital currents are both flows of angular momentum, their transport properties are fundamentally different, especially in magnets. Similar to the spin torque induced by an injection of a spin current into a magnet, an injection of an orbital current exerts a torque on the magnetic moment. Since the generation of torque (orbital torque) requires the transfer of orbital angular momentum of the conduction electrons to the local spins, the sign and magnitude of the torque are quite sensitive to the electronic structure of the magnet. We also note that in a magnet, a spin current decays quickly because of the exchange splitting. In contrast, an orbital current is not directly coupled with exchange splitting, suggesting that the orbital transport length can be much longer than the spin transport length.

In classical electromagnetism, the manipulation of magnetic moments can be achieved using magnetic fields produced by charge currents. The development of spintronics has led to the notion that spin currents and spin torques provide an efficient and versatile way to control magnetic states and dynamics in various classes of materials. Our study has proved that orbital currents and orbital torques provide an exciting new opportunity for efficient control of magnetic moments. The observed remarkable features of orbital currents demonstrate the great potential of orbital currents for device applications. This finding paves the way for the development of a new class of devices that exploit not only charge and spin currents, which have been the mainstay of electronics and spintronics, but also orbital currents, which until now have been unexplored.