Interfacial Spin-to-Charge Conversion in Sputtered MoTe2 Heterostructures Probed by Spin Pumping and Spin-Torque Ferromagnetic Resonance

In the world of modern electronics, the ability to control the flow of information is paramount. For decades, this has been done by moving electric charge, the physical flow of electrons, through wires and chips. However, scientists have long sought a more efficient method by harnessing a different property of the electron: its spin. You can think of spin as a tiny, intrinsic magnetic orientation, like a microscopic compass needle pointing either up or down. The field of spintronics aims to use these magnetic orientations to store and process data, offering the potential for devices that are faster and consume far less energy than current technology. A major hurdle in this field is the conversion between charge and spin. Researchers need materials that can easily turn a flow of electric charge into a stream of aligned spins, and vice versa. While some heavy metals have been known to do this well, a new class of materials called transition metal dichalcogenides has emerged as a promising alternative. These are layered compounds with unique atomic structures that promise even greater efficiency, but a critical question remained unanswered: does this conversion happen deep inside the material's bulk, or does it occur only at the very surface where two different materials touch?

A team of researchers set out to solve this mystery by studying a specific layered material called molybdenum ditelluride, or MoTe2, paired with a magnetic metal known as permalloy. They wanted to determine whether the powerful conversion of spin to charge happened throughout the entire thickness of the MoTe2 layer or if it was a phenomenon restricted strictly to the interface where the two materials met. To find the answer, they built a series of tiny electronic devices, varying the thickness of the MoTe2 layer from just a few nanometers up to twenty nanometers. They then subjected these devices to a technique called spin pumping, where they used magnetic resonance to inject a stream of spins from the magnetic layer into the MoTe2. If the conversion were happening inside the bulk of the material, the strength of the resulting electrical signal should have grown as the MoTe2 layer got thicker, eventually leveling off once the material was thick enough to carry the spin current fully. Instead, the researchers observed something different. The electrical signal generated by the spin-to-charge conversion remained almost exactly the same, regardless of whether the MoTe2 layer was thin or thick. This lack of change indicated that the conversion was not relying on the bulk material at all, but was instead occurring entirely at the boundary between the magnetic layer and the MoTe2.

To confirm this finding and rule out the possibility that the bulk material was simply too efficient to show a difference, the team performed a second set of experiments using a different method called spin-torque ferromagnetic resonance. In this setup, they sent an electrical current through the device to generate a spin current and observed how it affected the magnetic layer. They measured the efficiency of this process across the same range of thicknesses. Once again, the results showed no dependence on how thick the MoTe2 layer was. The efficiency of converting charge to spin remained constant, reinforcing the conclusion that the effect was driven by the interface. The researchers also compared their findings to samples made with platinum, a well-known material where the conversion is known to happen in the bulk. The platinum samples behaved exactly as expected, with the signal growing as the layer thickened, which served as a perfect control to prove that their experimental setup was working correctly and that the MoTe2 was behaving uniquely.

Further investigation revealed that the mechanism at play was a specific type of interaction known as the Rashba-Edelstein effect, which arises from the breaking of symmetry at the interface. To prove that this effect was indeed an interfacial phenomenon and not a bulk property, the team inserted a thin layer of gold between the magnetic metal and the MoTe2. This gold spacer acted as a barrier, separating the two materials so they could no longer interact directly. When they repeated the measurements with this barrier in place, the strong conversion signal vanished, dropping to a level similar to what was seen in the magnetic layer alone. This dramatic reduction confirmed that the powerful conversion effect required the direct contact between the MoTe2 and the magnetic layer. The study also uncovered that the spin currents generated at this interface had unusual properties, creating forces that pushed the magnetization in directions not typically seen in standard metal layers, a sign of the complex atomic interactions occurring right at the surface.

The work demonstrates that for sputtered MoTe2, the key to its high performance lies not in the volume of the material, but in the quality of the interface where it meets the magnetic layer. This distinction is crucial for the future development of spintronic devices. It suggests that engineers do not need to grow thick, perfect crystals of these materials to achieve high efficiency; instead, they can focus on optimizing the interface itself. The researchers found that the efficiency of the spin-orbit torque generated in these structures was remarkably high, comparable to or even exceeding that of the best heavy metals currently used. By showing that these effects are robust even in polycrystalline films made by standard industrial methods, the study opens a clear path toward creating low-power magnetic memory and logic devices based on two-dimensional materials, moving the field closer to practical, energy-efficient applications.