🔬 materials science

Challenges in orbital current-driven domain wall motion in light metal/ferrimagnet heterostructures

This study demonstrates that while orbital currents from light metals (Mn, Ti) can transfer angular momentum to a ferrimagnetic Gd-Fe-Co alloy, they fail to drive domain wall motion due to weak torque conversion and insufficient interfacial stabilization, a limitation that is overcome only by inserting a thin platinum layer to facilitate orbital-to-spin conversion.

Original authors: Min-Gu Kang, Jaerin Kim, Benjamin J. Jacot, Laura Van Schie, Pietro Gambardella

Published 2026-09-07

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Min-Gu Kang, Jaerin Kim, Benjamin J. Jacot, Laura Van Schie, Pietro Gambardella

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the world of modern electronics, the ability to control magnetism with electricity is the engine behind data storage and processing. For decades, scientists have relied on a specific trick to move magnetic information: they use heavy metals like platinum to generate a flow of spin, a tiny form of angular momentum that pushes magnetic regions called domain walls along a wire. This method works well, but it requires heavy, expensive materials. Recently, a new idea emerged suggesting that lighter, cheaper metals might do the same job. Instead of generating spin directly, these light metals could generate a flow of orbital angular momentum, a related but distinct property of electrons, which could then be converted into the spin needed to move magnetic walls. If this worked, it would open the door to faster, more energy-efficient devices without the need for heavy elements. However, while theory predicted this was possible, direct proof that these orbital currents could actually push magnetic walls in a real device had remained elusive.

A team of researchers at ETH Zurich set out to test this hypothesis using a specific magnetic alloy made of gadolinium, iron, and cobalt. They built tiny tracks on silicon chips, layering the magnetic alloy with either heavy platinum or lighter metals like manganese and titanium. Their goal was simple: send an electric current through the top metal layer and see if it could push a magnetic wall along the track. When they used platinum, the result was immediate and robust. The magnetic walls moved smoothly and predictably, confirming that the heavy metal was successfully generating the necessary force. This established a reliable baseline for how the system should behave when the physics works as expected.

When the researchers swapped the platinum for the light metals, manganese or titanium, the story changed completely. Despite sending strong electric currents through these layers, the magnetic walls refused to budge. They remained stuck in place, no matter how the current was adjusted. This was a surprising result, given that theory suggested these light metals should be capable of generating the required orbital currents. To understand why the walls didn't move, the team looked deeper. They measured the tiny electrical signals generated by the current and found that a small amount of force was indeed being transferred from the light metal to the magnetic layer. The force was real, but it was far too weak to overcome the friction and pinning that held the magnetic walls in place. It was as if someone was pushing a heavy door with just enough strength to make it creak, but not enough to swing it open.

The researchers realized that the problem lay in how the force was generated and transmitted. In the light metal layers, the orbital currents were created, but when they reached the magnetic alloy, they faced a complex internal structure. The magnetic alloy contains two different types of atoms that interact with the incoming force in opposite ways, effectively canceling each other out. Furthermore, the interface between the light metal and the magnetic layer lacked a specific stabilizing force that helps organize the magnetic walls into a shape that is easy to push. Without this organization, the weak force that did arrive was scattered and ineffective.

To solve this, the team introduced a very thin layer of platinum, just one nanometer thick, between the light metal and the magnetic alloy. This small addition acted as a bridge. It captured the orbital flow from the light metal and efficiently converted it into the spin force needed to move the walls. More importantly, this thin platinum layer helped organize the magnetic walls into the correct shape, making them much easier to push. With this single nanometer layer in place, the magnetic walls began to move again, responding to the current with the same speed and reliability as they did with the heavy platinum alone.

The study concluded that while light metals can generate the initial orbital currents, they cannot drive magnetic motion on their own. The process requires a specific interface to convert that orbital flow into useful spin force and to stabilize the magnetic structure. Simply injecting orbital current is not enough; the entire system must be tuned to ensure the force is strong and the magnetic walls are ready to move. This finding clarifies the path forward for developing new spintronic devices, showing that the key to success lies not just in the materials used to generate the current, but in the precise engineering of the interfaces where that current meets the magnetism.

Technical Summary: Challenges in Orbital Current-Driven Domain Wall Motion in Light Metal/Ferrimagnet Heterostructures

Problem Statement
Recent theoretical developments suggest that the orbital Hall effect (OHE) in light metals (e.g., Mn, Ti) could generate transverse orbital currents capable of driving magnetic dynamics without relying on the strong spin-orbit coupling (SOC) characteristic of heavy metals. However, experimental evidence for orbital-current-driven domain wall (DW) motion remains scarce, particularly in ferrimagnetic alloys. A critical unresolved question is whether orbital currents injected into rare-earth transition-metal ferrimagnets can generate sufficient torque to overcome pinning and drive DW motion. This is complicated by the complex sublattice structure of ferrimagnets, where rare-earth (Gd) and transition-metal (FeCo) sublattices possess opposite signs of SOC, potentially leading to partial cancellation of the effective torque. Furthermore, the role of interfacial Dzyaloshinskii–Moriya interaction (DMI) in stabilizing the chiral Néel-type DWs required for efficient torque coupling in these specific heterostructures is unclear.

Methodology
The authors investigated current-induced DW motion in amorphous ferrimagnetic Gd-rich Gd $_{25}$ (Fe $_9$ Co $_1$ ) $_{75}$ (GFC) thin films (15 nm thick) with perpendicular magnetic anisotropy (PMA). The study employed a comparative approach using three distinct heterostructure configurations:

Reference: Pt/GFC bilayers, utilizing the well-established spin Hall effect (SHE) in Pt.
Direct Injection: Light metal (LM) / GFC bilayers, where LM = Mn or Ti, to test direct orbital-current injection.
Conversion Layer: LM / Pt(1 nm) / GFC trilayers, incorporating an ultrathin Pt spacer to facilitate orbital-to-spin conversion and introduce interfacial SOC.

Experimental characterization included:

Real-space imaging: Polar magneto-optical Kerr effect (MOKE) microscopy to visualize DW nucleation and displacement under current pulses.
Quantitative torque measurement: Second-harmonic Hall resistance ( $R_{xy}^{2\omega}$ ) measurements to extract the damping-like effective field ( $B_{DL}$ ) and Hall conductivity ( $\xi_{DL}^E$ ).
Magnetic characterization: SQUID magnetometry and anomalous Hall effect measurements to confirm magnetic states and compensation temperatures.
Parameter extraction: Systematic variation of light metal thickness ( $t_{LM}$ ) to analyze dependencies of DW mobility ( $\mu_{DW}$ ), effective DMI constant ( $D$ ), and domain wall width ( $\Delta_{DW}$ ).

Key Results

Reference Performance: Pt/GFC bilayers exhibited robust, reproducible, and polarity-dependent DW motion driven by spin-orbit torque (SOT), with a mobility of $\mu_{DW} \approx 6.0 \times 10^{-10}$ m $^3$ /As. The DWs were confirmed to be left-handed Néel-type, stabilized by interfacial DMI.
Failure of Direct Orbital Injection: In Mn/GFC and Ti/GFC bilayers, no deterministic, directional DW motion was observed across a wide range of current densities and pulse durations, even up to thermal nucleation thresholds. While second-harmonic Hall measurements detected finite but extremely small damping-like torques ( $\xi_{DL}^E \approx -0.15$ and $-0.05$ for Mn and Ti, respectively), these were insufficient to depin DWs. The negative sign of the torque in LM/GFC contrasts with the positive sign in Pt/GFC, suggesting the torque is dominated by the Gd sublattice's negative SOC, likely partially cancelled by the FeCo sublattice.
Restoration via Pt Interlayer: Inserting a 1-nm Pt layer (LM/Pt/GFC) dramatically enhanced the damping-like torque (positive sign, $\xi_{DL}^E \approx +1.0$ to $+1.6$ ) and restored robust, deterministic DW motion.
Mechanism of Recovery: The recovery of DW motion in trilayers is attributed to two factors: (1) efficient orbital-to-spin conversion within the Pt layer, governed by its strong SOC, and (2) the induction of interfacial DMI at the Pt/GFC interface, which stabilizes chiral Néel DWs.
Thickness Dependence and Limitations: In LM/Pt/GFC trilayers, DW mobility did not scale simply with the thickness of the light metal (which would be expected if bulk orbital current generation were the sole driver). Instead, $\mu_{DW}$ correlated strongly with the effective DMI constant and domain wall width. For Mn/Pt/GFC, mobility peaked at 12 nm, matching the peak in DMI. For Ti/Pt/GFC, a sharp drop in mobility between 8 and 12 nm coincided with a reduction in domain wall width.

Significance and Claims
The paper concludes that while orbital currents can be generated in light metals and transferred to ferrimagnets, they are insufficient on their own to drive domain wall motion in LM/GFC bilayers. The primary challenges identified are:

Weak Net Torque: The net torque is significantly reduced due to competing orbital-to-spin conversion channels in the Gd and FeCo sublattices and potentially weak orbital accumulation in the light metals.
Lack of Chiral Stabilization: The absence of strong interfacial SOC in direct LM/GFC bilayers results in vanishing DMI, failing to stabilize the Néel-type DWs necessary for efficient coupling with damping-like torques.

The study demonstrates that realizing orbital-torque-driven DW motion requires the concurrent optimization of three elements: torque generation, efficient orbital-to-spin conversion (facilitated by a heavy-metal interlayer like Pt), and the stabilization of chiral domain walls via interfacial DMI. The results highlight that in light-metal/ferrimagnet heterostructures, interfacial spin-orbit effects and domain wall structural parameters are more critical determinants of DW dynamics than the magnitude of the bulk orbital current alone.

Technical Summary: Challenges in Orbital Current-Driven Domain Wall Motion in Light Metal/Ferrimagnet Heterostructures

More like this