betaFeedback

WorksPhysics and Astronomy

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

Abstract

Transition metal dichalcogenides (TMDs) and their Weyl semimetal phases, such as MoTe2_2, have attracted significant attention for spin-orbit torque applications due to their efficient charge-to-spin conversion. However, whether this conversion originates predominantly from the bulk or the interface remains unclear. Here, we investigate spin-charge interconversion in MoTe2_2 using spin-pumping and spin-torque ferromagnetic resonance (SP-FMR and ST-FMR). Thickness-dependent measurements reveal large spin-to-charge conversion and spin-torque efficiencies that are essentially independent of MoTe2_2 thickness, indicating that the conversion is predominantly governed by the Rashba-Edelstein effect at the Py/MoTe2_2 interface rather than by bulk spin transport. This behavior contrasts with the characteristic thickness dependence observed in Pt heterostructures and is further supported by bidirectional SP-FMR and interface-separation measurements. Our results highlight the dominant role of the Py/MoTe2_2 interface in enabling efficient spin-charge conversion and spin-orbit torques in TMD-based spintronic devices. These findings highlight the potential of sputtered MoTe2_2/Py heterostructures for low-power spintronic applications, including magnetic memory and logic devices.

Introduction

Since the discovery of two-dimensional (2D) materials, their unique properties have motivated extensive research in both fundamental physics and device applications[1, 2, 3, 4, 5]. Among these, transition metal dichalcogenides (TMDs) have emerged as particularly promising platforms due to their strong spin-orbit coupling (SOC), tunable band structure, and reduced crystalline symmetry[6, 7, 8, 9]. These properties have positioned TMDs as leading candidates for spintronic devices requiring lower power consumption and efficient charge-spin interconversion[10, 11, 12, 13, 14].

Recent research has focused on low-symmetry TMDs such as MoTe2\mathrm{MoTe}_2 and WTe2\mathrm{WTe}_2, where broken inversion symmetry combined with strong SOC gives rise to interfacial Rashba interactions and spin momentum-locking[15, 16, 17, 18, 19, 20, 21, 22, 23]. Interfaces with ferromagnetic materials can further enhance these effects through orbital hybridization and increased structural asymmetry. In this framework, the perpendicular electric field at the interface, together with the strong SOC, produces an effective magnetic field acting on conduction electrons. As a result, the spin orientation becomes locked to the momentum. This interaction can be interpreted as an effective Zeeman-like term that lifts the spin degeneracy of the electronic bands and, under an applied current, generates a non-equilibrium spin accumulation at the interface[24]. This current-induced spin accumulation is known as the spin Rashba-Edelstein effect (SREE), which converts charge currents into spin accumulation, while its reciprocal process, the inverse SREE (ISREE), converts spin current into charge current[24, 25]. These interface-driven mechanisms are fundamentally different from conventional spin Hall effects, in which the charge-to-spin interconversion occurs within the bulk and without spin-momentum locking.

Among TMDs, the identification of materials that combine strong SOC effects with compatibility with scalable fabrication techniques is of particular interest for spintronics applications. MoTe2\mathrm{MoTe}_2 stands out among TMDs due to its polymorphic phases (2H, 1T′, Td), temperature- and strain-dependent topological transitions, and large reported spin-torque efficiencies, even for sputtered samples[15, 16, 17, 18, 19, 20, 21, 26, 27, 28]. For example, MoTe2\mathrm{MoTe}_2-based heterostructures have demonstrated larger spin-torques efficiencies than those reported for heavy metals with strong SOC, such as Pt, Ta and W[29, 30, 31]. Additionally, MoTe2\mathrm{MoTe}_2 has been reported to exhibit unconventional spin-orbit torques in some heterostructures, a phenomenon associated with broken mirror symmetry and therefore not generally expected in polycrystalline samples[32, 33, 34]. The observation of such effects in sputtered MoTe2\mathrm{MoTe}_2-based systems is particularly relevant because it suggests that high spin-orbit torque efficiencies can be achieved without relying exclusively on growth techniques that produce high-quality single-crystalline TMDs, such as CVD and MBE[35, 36, 37]. From both scientific and technological perspective, these results highlight sputter deposition as a promising route toward the scalable fabrication of spintronics devices.

Despite the high efficiencies reported for MoTe2\mathrm{MoTe}_2, the relative contributions of bulk versus interfacial mechanisms remain a subject of intense debate. This is particularly true for polycrystalline sputtered films, where the conversion efficiency is highly sensitive to growth parameters, substrate choice, and phase stoichiometry[33, 34, 38]. In this work, we address this issue through a comparative study of sputtered MoTe2\mathrm{MoTe}_2 and Pt heterostructures using spin pumping driven by FMR SP-FMR) and spin-torque ferromagnetic resonance (ST-FMR). We first investigate spin-to-charge interconversion by SP-FMR and find that the conversion efficiency in MoTe2\mathrm{MoTe}_2 remains nearly independent of thickness, in strong contrast to the bulk-dominated behavior observed in Pt reference samples. This interpretation is further supported by bidirectional spin pumping measurements, which provide a direct means to distinguish between bulk and interface conversion mechanisms. We then investigate the reciprocal charge-to-spin conversion using ST-FMR. Angular-dependent measurements reveal signatures of unconventional out-of-plane torques associated with symmetry-breaking interfacial effects. By combining SP-FMR, ST-FMR, and interface-engineering experiments, our results demonstrate that the dominant conversion mechanism in these heterostructures is governed primarily by Rashba-type interfacial effects rather than bulk spin Hall transport.

Experimental results and theoretical discussion

SP-FMR in SiO2_2/Py/MoTe2_2 heterostructures

To investigate the spin-to-charge conversion in MoTe2\mathrm{MoTe}_2, we initially employed the SP-FMR technique, which allows us to detect the inverse spin Hall effect (ISHE) or ISREE[24, 25, 39, 40, 41, 42, 43, 44, 45]. The detailed experimental setup and parameters are explained in appendix § 7. Appendix § 6 shows the characterization of our MoTe2\mathrm{MoTe}_2 films. In the SP-FMR, the magnetization precession of a ferromagnetic (FM) layer transports spin angular momentum to the adjacent layer (generally non-magnetic, NM), creating a non-equilibrium spin accumulation that undergoes diffusion across the bulk of the adjacent material, thereby generating a spin current[34, 35, 36, 37, 38]. The spin current can be converted into a measurable charge current in the transversal direction, either through the bulk of the NM layer or through the interface NM/FM between the two layers, as described by the equations

J⃗cISHE=(2eℏ)θSH(σ^s×J⃗s),bulk interconversion(1)\vec{J}_{c}^{ISHE} = \left(\frac{2e}{\hbar}\right)\theta_{SH}\left(\hat{\sigma}_{s} \times\vec{J}_{s}\right), \qquad\text{bulk interconversion} \tag*{(1)}
J⃗cIREE=(eαRℏ)(z^×S⃗neq),interface interconversion(2)\vec{J}_{c}^{IREE} = \left(\frac{e\alpha_{R}}{\hbar}\right)\left(\hat{z} \times\vec{S}_{neq}\right), \qquad\text{interface interconversion} \tag*{(2)}

where σ^S\hat{\sigma}_S is the spin polarization of the spin current injected by the FM layer, θSH\theta_{SH} is the spin Hall angle, ee is the elementary charge and S⃗neq\vec{S}_{neq} is the non-equilibrium spin density, which is polarized perpendicular to the charge current direction. In our experiments, the bulk interconversion corresponds to the ISHE, while the interface interconversion corresponds to the ISREE.

The investigation of the spin-to-charge conversion in MoTe2_2 by the spin-pumping technique was initially done using two types of magnetic materials, YIG (Y3_3Fe5_5O12_{12}) and Py (Ni81_{81}Fe19_{19}). Despite the ferrimagnet YIG (Y3_3Fe5_5O12_{12}) being the standard material for spin pumping experiments in ferromagnet/non-magnet (FM/NM) heterostructures, we were unable to obtain consistent SP-FMR signals in YIG/MoTe2_2 bilayers, in agreement with previous reports on this system[46]. This behavior is likely related to the inefficient spin injection from YIG into two-dimensional van der Waals materials, which has also been observed in other YIG/2D heterostructures. Therefore we employed Py as the spin injector. Although Py-based spin pumping experiments are accompanied by additional contributions, such as galvanomagnetic effects (anomalous Hall effect, anisotropic magnetoresistance, etc.) and spin-to-charge self-conversion within the ferromagnetic layer[43, 47, 48, 49, 50], Py provides a robust and reliable source of spin current. Accordingly, spin pumping measurements were performed on a series of SiO2_2/MoTe2_2(t)/Py(7) samples, where the 7-nm-thick Py layer serves both as the spin injector and as a protective capping layer that prevents the oxidation of MoTe2_2. Figure 1 (a) schematically illustrates the SP-FMR geometry employed in this work. Under ferromagnetic resonance, a spin current J⃗S\vec{J}_S is pumped from the Py layer into the underlying MoTe2_2 layer along z-direction. The spin polarization σ^S\hat{\sigma}_S, which follows the magnetization direction, lies in the film plane, so that the spin-to-charge conversion generates a transverse charge current, J⃗C∝σ^S×J⃗S\vec{J}_C \propto\hat{\sigma}_S \times\vec{J}_S, flowing along y direction. Figure 1(b)-(Figure 1d) show the SP-FMR signals measured for the SiO2_2/MoTe2_2(3)/Py(7) sample with the magnetic field applied at ϕ=0∘\phi= 0^\circ, 180∘180^\circ, and 90∘90^\circ, respectively. To directly quantify the spin-to-charge conversion, the electrical signal is expressed as the charge current IC=V/RI_C = V/R, where VV is the voltage measured by a nanovoltmeter and RR is the electrical resistance between the voltage probes. The solid black curves are Lorentzian fits to the experimental data. Reversing the magnetic field from ϕ=0∘\phi= 0^\circ [Figure 1(b)] to ϕ=180∘\phi= 180^\circ [Figure 1(c)] reverses the spin polarization, and consequently the sign of the converted charge current, as expected from J⃗C∝σ^S×J⃗S\vec{J}_C \propto\hat{\sigma}_S \times\vec{J}_S. In contrast, for ϕ=90∘\phi= 90^\circ [Figure 1(d)], the charge current along the y direction is expected to vanish because σ^S×J⃗S\hat{\sigma}_S \times\vec{J}_S is perpendicular to the voltage probes, in agreement with the negligible measured signal. Since ferromagnetic resonance in Py also produces intrinsic electrical signals, including galvanomagnetic effects and spin-to-charge self-conversion, the measured current contains both the contribution associated with MoTe2_2 and the intrinsic response of the Py layer. Therefore, the measured current can be modeled as the sum of symmetric and antisymmetric Lorentzian functions,

SP-FMR experimental geometry and measured signals

Figure 1. (a) Schematic illustration of the SP-FMR experimental geometry, showing the heterostructure stack and the detected signal. (b - d) Measured SP-FMR signal as a function of the in-plane applied magnetic field for the azimuthal angles ϕ=0∘\phi= 0^\circ, ϕ=180∘\phi= 180^\circ and ϕ=90∘\phi= 90^\circ, defined as the angle between the magnetic field and the current direction. The solid lines correspond to the best fits to the data using Eq. (3). The fits yield IS=−216.8I_S = -216.8 nA (ϕ=0∘\phi= 0^\circ), IS=213.9I_S = 213.9 nA (ϕ=180∘\phi= 180^\circ), confirming the spin-pumping origin, while IA=−46.1I_A = -46.1 nA and IA=10.7I_A = 10.7 nA, respectively. At ϕ=90∘\phi= 90^\circ, the symmetric signal vanishes as expected. All measurements were performed at an applied RF power of 110 mW.

I=ISΔH2(H−Hr)2+ΔH2+IAΔH(H−Hr)(H−Hr)2+ΔH2,(3)I=I_{\mathrm{S}}\frac{\Delta H^{2}}{(H-H_{\mathrm{r}})^{2}+\Delta H^{2}}+I_{\mathrm{A}}\frac{\Delta H(H-H_{\mathrm{r}})}{(H-H_{\mathrm{r}})^{2}+\Delta H^{2}}, \tag*{(3)}

where HrH_{\mathrm{r}} is the resonance field, ΔH\Delta H is the half-width at half-maximum, and ISI_{\mathrm{S}} and IAI_{\mathrm{A}} are the amplitudes of the symmetric and antisymmetric Lorentzian components, respectively. The sign of the fitted signal follows the direction of the spin-to-charge conversion predicted by Eq. (1) or Eq. (2), providing a direct consistency check with the expected polarity of the bulk ISHE and the interfacial ISREE.

To investigate the spin-to-charge conversion mechanisms in Py/MoTe2_2 bilayers, we systematically studied the dependence of the SP-FMR signals on the MoTe2_2 thickness by fabricating SiO2_2/MoTe2_2(tMoTe2_{\mathrm{MoTe_2}})/Py(7) heterostructures with tMoTe2_{\mathrm{MoTe_2}} = 3, 4, 8, 10, 15, and 20 nm. To account for the self-conversion signal of the Py layer, a 7 nm Py reference sample was also measured under identical conditions. The corresponding SP-FMR spectra are presented in Figure 2(a)-(Figure 2h), together with the fitting curves obtained using Eq. (3). Measurements were performed for two opposite magnetic-field orientations (ϕ=0∘\phi= 0^\circ, and ϕ=180∘\phi= 180^\circ), allowing the separation of the spin-pumping contribution from spurious voltage signals. As expected for a genuine spin-pumping induced charge current, reversing the magnetic field orientation results in a complete inversion of the signal polarity while preserving both the resonance field and the spectral lineshape. The excellent agreement between the experimental data and the fitting curves confirms that the spectra are well described by the adopted fitting model.

SP-FMR spectra measured for the reference SiO$_2$/Py(7) film and SiO$_2$/MoTe$_2$/Py(7) bilayers

Figure 2. SP-FMR spectra measured for the reference SiO2_2/Py(7) film (a) and SiO2_2/MoTe2_2(tMoTe2t_{MoTe_2})/Py(7) bilayers with tMoTe2t_{MoTe_2} = 3 (b), 4 (c), 5 (d), 8 (e), 10 (f), 15 (g) and 20 nm (h). The blue and pink symbols correspond to measurements with ϕ=0∘\phi= 0^\circ and ϕ=180∘\phi= 180^\circ, respectively. The solid black curves are fits to Eq. (3). The measurements were carried out at an rf frequency of 9.4 GHz and an rf power of 110 mW.

A comparison between the reference Py film [Figure 2(a)] and the Py/MoTe2_2 bilayers [Figure 2(b)-(Figure 2h)] reveals a pronounced enhancement of the detected current upon the introduction of the MoTe2_2 layer. While the self-conversion current of Py reference is approximately Is=110I_{\mathrm{s}}=110 nA, the heterostructures exhibit currents ranging from about 190 to 230 nA, nearly doubling the signal intensity. This enhancement clearly indicates an additional spin-to-charge conversion contribution associated with the MoTe2_2 layer. Furthermore, despite the MoTe2_2 thickness varying from 3 to 20 nm, the signal amplitude remains nearly constant, exhibiting only moderate sample-to-sample variations without any systematic thickness dependence. This behavior suggests that the dominant conversion mechanism is largely thickness independent over the investigated range, consistent with a process governed primarily by the interface or by a characteristic length shorter than the investigated thickness range. Additionally, both the resonance field and linewidth remain essentially unchanged for all samples, indicating that the magnetic properties of the 7 nm Py layer are only weakly affected by the presence and thickness of the MoTe2_2 layer.

To better visualize these results, the magnitude of the symmetric SP-FMR components (ISI_{\mathrm{S}}) is plotted as a function of MoTe2_2 thickness in Figure 3(a). The values of ISI_{\mathrm{S}} were obtained by averaging the absolute values extracted from the measurements performed with ϕ=0∘\phi= 0^\circ and ϕ=180∘\phi= 180^\circ. The error bars include both the fitting uncertainty and variance between the two field orientations. For comparison, analogous measurements were carried out on SiO2_2/Pt(tPtt_{\mathrm{Pt}})/Py(7) bilayers, which exhibits the thickness dependence expected for a bulk spin-to-charge conversion process [Figure 3(b)]. The complete SP-FMR data for the Pt/Py samples are presented in Fig. A.3 of Appendix C. Fitting the data in Figure 3(b) with IPeak∝tanh⁡(tPt/2λS)I_{\mathrm{Peak}} \propto\tanh(t_{\mathrm{Pt}}/2\lambda_{\mathrm{S}}) yields a spin diffusion length of λS=(1.1±0.1)\lambda_{\mathrm{S}} = (1.1 \pm0.1) nm. The data point at tPt=0t_{\mathrm{Pt}} = 0 corresponds to the self-conversion signal of the Py(7) reference layer. In contrast, the SP-FMR signal measured in the MoTe2_2/Py bilayer remains nearly constant over the entire investigated thickness range, showing only small sample-to-sample variations around an average value of approximately 210 nA, well above the Py self-conversion signal. The absence of a systematic thickness dependence is consistent with a spin-to-charge conversion process dominated by the MoTe2_2/Py interface, as expected for the ISREE.

Dependence of the SP-FMR peak current on the thickness of the spin-to-charge conversion layer

Figure 3. Dependence of the SP-FMR peak current (IPeakI_{\mathrm{Peak}}) on the thickness of the spin-to-charge conversion layer. (a) SiO2_2/MoTe2_2(tMoTe2t_{\mathrm{MoTe_2}})/Py(7) heterostructures exhibit an approximately thickness-independent response, consistent with an interface-dominated spin-to-charge conversion process. The dashed horizontal line indicates the average IPeakI_{\mathrm{Peak}}, and the red symbol at tMoTe2=0t_{\mathrm{MoTe_2}}=0 represents the self-conversion signal Py(7) reference film. (b) SiO2_2/Pt (tPtt_{\mathrm{Pt}})/Py(7) heterostructures display the characteristic thickness dependence expected for a bulk spin-diffusion process. The solid line is a fit to IPeak∝tanh⁡(tPt/2λS)I_{\mathrm{Peak}} \propto\tanh(t_{\mathrm{Pt}}/2\lambda_{\mathrm{S}}), yielding a spin diffusion length of λS=(1.1±0.1)\lambda_{\mathrm{S}} = (1.1 \pm0.1) nm.

To further verify the interfacial origin of the spin-to-charge conversion inferred from the thickness-independent SP-FMR amplitudes, we performed SP-FMR measurements in a trilayer geometry that enables spin-current injection into the nonmagnetic layer from opposite directions. This configuration is realized in the Co/MoTe2_2/Py heterostructure, where spin pumping from either the Py or Co layer injects spin current into the MoTe2_2 layer. For Rashba-type interfacial conversion, described by Eq. (2), the generated charge current is independent of the spin-current injection direction[51, 52, 53, 54, 55]. Consequently, spin currents injected from opposite sides produce electrical signals with the same polarity. As a control experiment, we also investigated Co/Pt/Py, where the dominant bulk spin-to-charge conversion in Pt is expected to follow Eq. (1), leading to opposite signal polarities for opposite spin-current injection directions. Figure 4 summarizes the results obtained for Co(7)/Pt(10)/Py(7) [Figure 4(a–Figure 4c)] and Co(7)/MoTe2_2(3)/Py(7) [Figure 4(d–Figure 4f)]. Since the measured SP-FMR signal contains contributions from both the spin-to-charge conversion in the nonmagnetic layer and the self-conversion in the ferromagnetic layers, the latter was subtracted using independent measurements on single Co and Py reference samples, as shown in the insets of Figure 4(b). Two main conclusions emerge from these measurements. First, the spin-to-charge conversion at the Co/Pt and Co/MoTe2_2 interfaces exhibit a relatively weaker spin-to-charge conversion than the Co/Py and MoTe2_2/Py interfaces is weaker than at the Pt/Py and MoTe2_2/Py interfaces, as evidenced by the smaller IPeakI_{\mathrm{Peak}} amplitudes. This behavior is likely associated with a lower spin-mixing conductance at the Co interface, possibly resulting from the larger roughness of the sputtered Co films. Second, in the Co/Pt/Py trilayer the two resonances exhibit opposite polarities, as expected for bulk spin-to-charge governed by Eq. (1). In contrast, in the Co/MoTe2_2/Py trilayer both resonances have the same polarity, despite the opposite spin-current injection directions. This behavior is precisely that predicted by Eq. (2) for the ISREE, providing direct evidence that spin-to-charge conversion in MoTe2_2 is dominated by interfacial Rashba states rather than by bulk states.

Schematic illustrations and SP-FMR spectra of SiO2/Co(7)/MoTe2(3)/Py(7) and SiO2/Co(7)/Pt(10)/Py(7) heterostructures

Figure 4. (a, d) Schematic illustration of the SiO2/Co(7)/MoTe2(3)/Py(7) and SiO2/Co(7)/Pt(10)/Py(7) heterostructures, respectively, indicating the directions of charge current generated by the (a) ISHE and (d) ISREE. (b, e) SP-FMR spectra of the corresponding trilayers. The insets show the self-conversion signals measured independently in the Co(7) and Py(7) reference samples. (c, f) Spin-to-charge conversion signals obtained after subtracting the Co and Py self-conversion contributions from the total SP-FMR spectra in (b) and (e), respectively, thereby isolating the contributions from Pt and MoTe2. All measurements were performed with an rf power of 110 mW.

ST-FMR in SiO2/Py/MoTe2 heterostructures

Figure 5 shows a schematic representation of the ST-FMR experimental setup and the measured signals. As illustrated in Figure 5(a), an rf current, IrfI_{rf}, is applied to a FM/NM bilayer, where the NM in the present work is MoTe2. The current flowing through the NM layer generates an rf Oersted field (H⃗oe\vec{H}_{oe}) that drives the magnetization dynamics of the adjacent FM layer. When the excitation frequency matches the natural precession frequency of the FM magnetization, the system reaches FMR condition. Simultaneously, IrfI_{rf} can be converted into a spin current (J⃗s\vec{J}_{s}) either within the NM layer via the SHE or at the FM/NM interface through the spin Rashba effect (SREE), generating additional spin-orbit torques that modify the magnetization dynamics. These torques can be incorporated into the Landau-Lifshitz-Gilbert equation[56, 57, 58, 59, 60, 61]

ST-FMR geometry, setup, representative signals, and angular dependence

Figure 5. (a) Schematic illustration of the ST-FMR geometry for the SiO2_2/NM/Py bilayer. An rf current IrfI_{\rm rf} flowing through to the NM layer generates an Oersted field (H⃗oe\vec{H}_{\rm oe}) and a spin current J⃗S\vec{J}_{S} via the bulk spin Hall effect or the interfacial Rashba-Edelstein effect (REE). The spin current exerts damping-like (τ⃗DL\vec{\tau}_{DL}) and field-like (τ⃗FL\vec{\tau}_{FL}) torques on the magnetization m^\hat{m}. (b) Schematic of the ST-FMR setup with lock-in detection used to measure the mixing voltage (VmixV_{\rm mix}) attributed to the combined action of τ⃗DL\vec{\tau}_{DL} and τ⃗FL\vec{\tau}_{FL}. (c) Representative ST-FMR signal measured in SiO2_2/MoTe2_2(10)/Py(7) measured for opposite magnetic field directions. (d) Angular dependence of the symmetric (VSV_S) and antisymmetric (VAV_A) components of VmixV_{\rm mix}. Measurements were performed at a microwave frequency of 6.0 GHz and an rf power of 8.0 mW.

dm^dt=−γm^×H⃗eff+α(m^×dm^dt)+γMs(τ⃗FL+τ⃗DL)(4)\frac{d\hat{m}}{dt}=-\gamma\hat{m}\times\vec{H}_{eff}+\alpha\left(\hat{m}\times\frac{d\hat{m}}{dt}\right)+\frac{\gamma}{M_{s}}\left(\vec{\tau}_{FL}+\vec{\tau}_{DL}\right) \tag*{(4)}

where γ\gamma is the gyromagnetic ratio, α\alpha is the gilbert damping parameter, H⃗eff=−∇⃗ME\vec{H}_{eff}=-\vec{\nabla}_{M}E is the effective magnetic field, and τ⃗FL=τFL(m^×σ^)\vec{\tau}_{FL}=\tau_{FL}(\hat{m}\times\hat{\sigma}) and τ⃗DL=τDL[m^×(m^×σ^)]\vec{\tau}_{DL}=\tau_{DL}[\hat{m}\times(\hat{m}\times\hat{\sigma})] are the field-like (FL) and damping-like (DL) spin-orbit torques, respectively, with σ^\hat{\sigma} denoting the spin current polarization direction. As illustrated in Figure 5(a), τ⃗FL\vec{\tau}_{FL} is normal to the plane defined by m^\hat{m} and σ^\hat{\sigma}. It has the same symmetry as an effective magnetic field acting on the magnetization and therefore contributes to its precessional motion. In contrast, the damping-like torque lies within the plane defined by m^\hat{m} and σ^\hat{\sigma}, and has the same symmetry as the Gilbert damping torque. Consequently, it can either increase or decrease the effective magnetic damping, depending on its sign.

Since FM materials exhibit anisotropic magnetoresistance (AMR), magnetization precession under FMR produces an oscillating electrical resistance. The mixing of this time-dependent resistance with IrfI_{\rm rf} generates a rectified voltage, detected as the ST-FMR signal using the lock-in technique illustrated in Figure 5(b). The damping-like spin-orbit torque typically gives rise to a symmetric Lorentzian contribution, whereas the Oersted field and field-like torque mainly contribute to an antisymmetric Lorentzian component. Consequently, the measured ST-FMR spectrum consists of a superposition of symmetric and antisymmetric Lorentzian line shapes and can be described by[56, 57, 58, 59, 60, 61]

Vmix=VSΔH2(H−Hr)2+ΔH2+VAΔH(H−Hr)(H−Hr)2+ΔH2(5)V_{\rm mix}=V_S\frac{\Delta H^2}{(H-H_r)^2+\Delta H^2}+V_A\frac{\Delta H(H-H_r)}{(H-H_r)^2+\Delta H^2} \tag*{(5)}

where HrH_r is the resonance field, ΔH\Delta H is the half-width at half-maximum, and VSV_S and VAV_A are the amplitudes of the symmetric and antisymmetric Lorentzian components, respectively. A representative ST-FMR signal for the SiO2_2/MoTe2_2/Py sample is shown in Figure 5(c) (green symbols). The dark solid curves represent the addition of symmetric VSV_S and antisymmetric VAV_A components of the signal, obtained by fitting the experimental data using Eq. (5), with the corresponding fitted values indicated in Figure 4c). Notably, VAV_A exhibits an asymmetry under a 180° rotation of the magnetic field: its magnitude differs significantly between ϕ=45∘\phi=45^\circ and 225∘225^\circ. This behavior is not accounted for by the conventional angular dependence shown by the solid curve in Figure 5(d), indicating the presence of an additional contribution to the ST-FMR signal, as discussed below.

The extracted values of VSV_S and VAV_A are essential for determining the effective spin-torque efficiency, (ξFMR\xi_{FMR}), which is calculated from [56, 57, 58, 59, 60, 61]

ξFMR=VSVAeℏμ0MstFMtNM1+4πMeffHr,(6)\xi_{FMR} = \frac{V_S}{V_A}\frac{e}{\hbar}\mu_0 M_s t_{FM}t_{NM}\sqrt{1+\frac{4\pi M_{eff}}{H_r}}, \tag*{(6)}

where MsM_s is the saturation magnetization, tNMt_{NM} and tFMt_{FM} are the thicknesses of the NM and FM layers, respectively, and 4πMeff4\pi M_{eff} is the effective magnetization determined from the Kittel equation ω=γH(H+4πMeff)\omega= \gamma\sqrt{H(H+4\pi M_{eff})}. Assuming that the damping-like and field-like torque efficiencies, ξDL\xi_{DL} and ξFL\xi_{FL}, independent of the FM layer thickness, they can be extracted from the linear dependence of 1/ξFMR1/\xi_{FMR} as a function of 1/tFM1/t_{FM}[56, 57, 58, 59, 60, 61],

1ξFMR=1ξDL(1+ℏeμ0MstNMtFMξFL).(7)\frac{1}{\xi_{FMR}} = \frac{1}{\xi_{DL}}\left(1+\frac{\hbar}{e\mu_0 M_s t_{NM}t_{FM}}\frac{\xi_{FL}}{}\right). \tag*{(7)}

The thickness dependence of ξFMR\xi_{FMR} can be used to qualitatively assess whether the generated spin torque originates predominantly from bulk or interfacial mechanisms. For torques generated in the NM bulk via SHE, ξFMR\xi_{FMR} is expected to increase with NM thickness and eventually saturate once the thickness exceeds the spin diffusion length. In contrast, for torques generated predominantly by interfacial mechanisms, such as the Rashba-Edelstein effect, ξFMR\xi_{FMR} is expected to remain nearly constant with increasing NM thickness. To investigate this behavior, we fabricated two sets of samples. In the first set, the FM layer thickness was varied while keeping the MoTe2_2 thickness constant, enabling a detailed analysis of the DL and FL torque efficiencies. In the second set, the MoTe2_2 thickness was varied while keeping the magnetic layer thickness fixed, allowing us to evaluate whether the spin-to-charge interconversion originates predominantly from the MoTe2_2 bulk or from the FM/MoTe2_2 interface.

To corroborate the thickness-dependent SP-FMR results, we performed ST-FMR measurements on SiO2_2/MoTe2_2(tMoTe2t_{\mathrm{MoTe}_2})/Py(7) heterostructures with tMoTe2=2t_{\mathrm{MoTe}_2}=2 - 20 nm. The Py layer was deposited on top of the MoTe2_2 layer to protect it from oxidation. In addition to field-swept measurements, angle-dependent ST-FMR experiments were carried out by varying the in-plane angle ϕ\phi from 0 to 360°, where ϕ\phi is the angle between the external magnetic field HH and the rf current IrfI_{rf}. Such measurements are essential for identifying unconventional torque components that cause angular asymmetries, as previously reported[32, 31, 61].

Figure 6 summarizes the experimental measurements. The left column shows the ST-FMR spectra (VmixV_{\mathrm{mix}}) for fixed angles of ϕ=225∘\phi=225^\circ and ϕ=45∘\phi=45^\circ. The spectra exhibit pronounced symmetric Lorentzian contributions, particularly for thinner MoTe2_2 layers. As the MoTe2_2 thickness increases, the antisymmetric component becomes progressively larger, mainly because a larger fraction of the rf current flows through the MoTe2_2 layer, thereby enhancing the Oersted field. In addition to these measurements, we performed in-plane angular measurements, as shown in. The right column of Figure 6 shows the angular dependences of symmetric (VSV_S, blue symbols) and antisymmetric (VAV_A, pink symbols) components extracted from the ST-FMR spectra. Clear angular asymmetries are observed in both VSV_S and VSV_S, which cannot be explained solely by the conventional torque components. For example, the absolute value of VSV_S (and VAV_A) at ϕ=45∘\phi=45^\circ differs from that at ϕ=225∘\phi=225^\circ, demonstrating that the response is not symmetric under a 180° rotation. This behavior is reproduced by the fitting curves and reveals the presence of an additional out-of-plane torque component.

Image not extracted

Figure 6. ST-FMR measurements for SiO₂/MoTe₂(tMoTe2t_{\mathrm{MoTe}_2})/Py(7) heterostructures with MoTe₂ = 2, 4, 8, 15 e 20 nm. Left panels: VmixV_{\mathrm{mix}} signals acquired at ϕ=225∘\phi= 225^\circ and ϕ=45∘\phi= 45^\circ, fitted with symmetric (VSV_{\mathrm{S}}) and antisymmetric (VAV_{\mathrm{A}}) Lorentzian components. Right panels: angular dependence of VSV_{\mathrm{S}} and VAV_{\mathrm{A}} components, fitted using Eqs. (6) and (7). The dashed horizontal lines highlight the angular asymmetries. All measurements were performed at a microwave frequency of 6.0 GHz and an RF power of 8.0 mW.

Phenomenologically, in addition to the conventional cos⁡ϕsin⁡2ϕ\cos\phi\sin2\phi dependence of VSV_S and VAV_A, the experimental data require an additional sin⁡2ϕ\sin2\phi term. This contribution naturally arises by writing the damping-like (DL) and field-like (FL) torques as τDL(ϕ)=τDL0cos⁡ϕ+τDL⊥\tau_{DL}(\phi)=\tau_{DL}^{0}\cos\phi+\tau_{DL}^{\perp}, and τFL(ϕ)=τFL0cos⁡ϕ+τFL⊥\tau_{FL}(\phi)=\tau_{FL}^{0}\cos\phi+\tau_{FL}^{\perp}, where τDL⊥\tau_{DL}^{\perp} and τFL⊥\tau_{FL}^{\perp} are angle-independent out-of-plane contributions. Since VSV_S and VAV_A are proportional to τDL(ϕ)\tau_{DL}(\phi) and τFL(ϕ)\tau_{FL}(\phi), respectively, and also dR/dϕdR/d\phi, with the AMR resistance following R∝cos⁡2ϕR\propto\cos^2\phi, the angular dependence can be written as[29, 32, 58, 59, 61, 62]

VS(ϕ)∝(τDL0cos⁡ϕ+τDL⊥)sin⁡2ϕ(8)V_S(\phi)\propto\left(\tau_{DL}^{0}\cos\phi+\tau_{DL}^{\perp}\right)\sin2\phi \tag*{(8)}
VA(ϕ)∝(τFL0cos⁡ϕ+τFL⊥)sin⁡2ϕ(9)V_A(\phi)\propto\left(\tau_{FL}^{0}\cos\phi+\tau_{FL}^{\perp}\right)\sin2\phi \tag*{(9)}

The observation of out-of-plane torque components indicates an additional symmetry breaking beyond that expected for ideal isotropic metallic bilayers. Such asymmetries are typically attributed to mirror symmetry breaking, structural inversion asymmetry, or the generation of spin currents with out-of-plane spin polarization[29, 32, 58, 59, 61, 62]. In the present MoTe2\mathrm{MoTe}_2 system, this symmetry reduction likely originates at the FM/MoTe2\mathrm{MoTe}_2 interface, where the structural inversion asymmetry, interfacial strain, or local disorder can break the spatial symmetry thereby leading to the generation of a spin current with out-of-plane polarization.

To further distinguish between bulk and interface charge-spin conversion, we used Eq. (6) to calculate the parameter ξFMR\xi_{\mathrm{FMR}}, for the SiO₂/MoTe₂(tMoTe2t_{\mathrm{MoTe}_2})/Py(7) bilayers and compared the results with those obtained for the control SiO₂/Pt(tPtt_{\mathrm{Pt}})/Py(7) heterostructure. For each sample ξFMR\xi_{\mathrm{FMR}} was calculated using the fitted amplitudes at the angle corresponding to maximum ST-FMR signal (ϕ=45∘\phi= 45^\circ). Figure 7(a) summarizes the results. The black symbols correspond to the MoTe₂/Py bilayers, whereas the red symbols represent the Pt/Py samples (multiplied by a factor of two for clarity). In the MoTe₂/Py heterostructures, ξFMR\xi_{\mathrm{FMR}} remains nearly independent of the MoTe₂ thickness, fluctuating around an average value. In contrast, the Pt/Py samples exhibit the expected asymptotic increase with Pt thickness, followed by saturation at the spin diffusion length, characteristic of a bulk spin Hall effect. The distinct thickness dependences therefore indicate that charge-to-spin conversion in the MoTe₂/Py bilayers is predominantly interfacial, consistent with a Rashba-Edelstein mechanism, whereas the Pt/Py bilayers are dominated by the conventional bulk spin Hall effect. These results are fully consistent with the SP-FMR measurements and further support the existence of a Rashba interface in the sputtered MoTe₂/Py heterostructures.

Image not extracted

Figure 7. (a) Extracted values of ξFMR\xi_{FMR}, as a function of the NM layer thickness, tNMt_{NM}, for SiO2_2/MoTe2_2(tMoTe2t_{MoTe_2})/Py(7) and SiO2_2/Pt(tMoTe2t_{MoTe_2})/Py(7) heterostructures. The Pt samples (red symbols) exhibit the characteristic asymptotic thickness dependence expected for a bulk SHE, whereas the MoTe2_2 samples (black symbols) remain nearly independent of thickness, fluctuating around a constant average value, consistent with an interface-dominated charge-to-spin conversion mechanism. The solid red curve is the fit to the bulk SHE model, while the dashed blue line indicates the average value of ξFMR\xi_{FMR} for the MoTe2_2 samples. (b) Linear fit of 1/ξFMR1/\xi_{FMR} versus 1/tFM1/t_{FM} for SiO2_2/Py(tt)/Pt(8), yielding a damping-like torque efficiency of ξDL=0.022\xi_{DL}=0.022 calculated for ϕ=45∘\phi=45^\circ (c) Linear fit of 1/ξFMR1/\xi_{FMR} versus 1/tFM1/t_{FM} for SiO2_2/MoTe2_2(3)/Py(tt), yielding a significantly larger damping-like torque efficiency of ξDL=0.13\xi_{DL}=0.13 calculated for ϕ=45∘\phi=45^\circ.

To quantify the damping-like torque efficiency ξDL\xi_{\mathrm{DL}}, the FM layer thickness was varied while keeping the adjacent layer thickness constant, as described by Eqs. (7). The saturation magnetization was assumed to remain approximately constant (μ0≈0.8\mu_0 \approx0.8 T) over the investigated thickness range. The values of 1/ξFMR1/\xi_{FMR} were then plotted as a function of 1/tFM1/t_{FM}, and ξDL\xi_{DL} was extracted from the corresponding linear fits. Figure 7(b) shows the linear fit for obtained for the Py(tt)/Pt(8) heterostructure, yielding ξDL=0.022\xi_{DL}=0.022, good agreement with previously reported values for Pt/Py. This value is consistent with the results reported in the literature for Pt/Py[59, 61]. The analysis of the MoTe2_2(3)/Py(tt) samples is more complex because the angular asymmetry discussed in Figure 6 implies that the torque efficiency depends on the in-plane angle. Thus, ξDL\xi_{DL} was evaluated at two angles corresponding to the maximum ST-FMR amplitudes, ϕ=45∘\phi=45^\circ and 225∘225^\circ. Figure 7(c) shows the linear fit for obtained for the MoTe2_2(3)/Py(tt) heterostructure for ϕ=45∘\phi=45^\circ, yielding ξDL=0.13\xi_{DL}=0.13 whereas the analysis at ϕ=225∘\phi=225^\circ gives a substantially larger value of ξDL=0.39\xi_{DL}=0.39. The strong angular dependence of ξDL\xi_{DL} reflects the presence of additional non-conventional damping-like torque components arising from the reduced symmetry of the MoTe2_2 interface. Similar angular asymmetries have been reported for both single-crystalline and polycrystalline MoTe2_2-based heterostructures, and are in agreement with the values found in recent work for single-crystal and polycrystalline MoTe2_2[31, 33, 34]. In particular, the value ξDL=0.39\xi_{DL}=0.39 is remarkably close to the efficiency of ξDL=0.35\xi_{DL}=0.35 reported from field-free spin-orbit-torque magnetization switching in optimized MoTe2_2 device[31].

Investigation of ST-FMR and SP-FMR in heterostructures with an Au spacer

To further probe the interfacial origin of the spin-charge conversion, we inserted a 1.5-nm-thick Au spacer layer between the MoTe2_2 and Py layers and compared the results with those of the corresponding MoTe2_2/Py bilayer. The Au spacer is expected to suppress the strong Rashba interaction at the MoTe2_2/Py interface while contributing negligibly to the spin-orbit torques because of its small spin Hall angle. Therefore, if the unconventional out-of-plane torque and the enhanced spin-to-charge conversion originate primarily at the MoTe2_2/Py interface, both effects should be strongly reduced after insertion of the Au layer.

Figure 8(a) and Figure 8(b) show the ST-FMR spectra and angular dependence for the SiO2_2/MoTe2_2(3)/Py(7) bilayer, while Figure 8(c) and Figure 8(d) present the corresponding results for SiO2_2/MoTe2_2(3)/Au(1.5)/Py(7). The angular asymmetry observed in the MoTe2_2/Py disappears after introducing the Au spacer, indicating that the unconventional out-of-plane torque originates at the MoTe2_2/Py interface. In addition, a pronounced reduction of the symmetric ST-FMR component is observed. Comparing Figure 8(b) and Figure 8(d), the symmetric voltage decreases from approximately 0.8 μ\muV to 0.3 μ\muV, corresponding to a reduction of about 63%. This substantial suppression indicates that the interfacial contribution to the spin-charge conversion has been largely eliminated. This interpretation is independently confirmed by the SP-FMR measurements shown in Figure 8(e) and Figure 8(f). The peak spin-pumping current decreases from approximately 215 nA for the MoTe2_2/Py bilayer to about 129 nA after insertion of the Au spacer. Moreover, the remaining current is very close to the self-conversion signal measured for the Py(7) reference sample, indicating that the additional contribution associated with the MoTe2_2/Py interface has been effectively removed. The excellent agreement between the ST-FMR and SP-FMR results provides compelling evidence that both the unconventional torque components and the enhanced spin-to-charge conversion originate predominantly at the MoTe2_2/Py interface rather than in the bulk of the MoTe2_2 layer. These findings are fully consistent with the thickness-dependent measurements and the bilateral spin-injection experiments, further establishing the Rashba-Edelstein effect as the dominant conversion mechanism in the sputtered MoTe2_2/Py heterostructures.

Effect of an Au spacer layer on the ST-FMR and SP-FMR response of MoTe2/Py heterostructures

Figure 8. Effect of an Au spacer layer on the ST-FMR and SP-FMR response of MoTe2_2/Py heterostructures. (a) ST-FMR mixing voltage (VmixV_{\mathrm{mix}}) and (b) angular dependence of the symmetric VSV_{\mathrm{S}} and antisymmetric VAV_{\mathrm{A}} components for the SiO2_2/MoTe2_2(3)/Py(7) heterostructure. (c) ST-FMR mixing voltage (VmixV_{\mathrm{mix}}) and (d) angular dependence of the symmetric VSV_{\mathrm{S}} and antisymmetric VAV_{\mathrm{A}} components for the SiO2_2/MoTe2_2(3)/Au(1.5)/Py(7) heterostructure. The ST-FMR measurements were performed at a microwave frequency of 6.0 GHz. (e) SP-FMR signals for the SiO2_2/MoTe2_2(3)/Py(7) heterostructure, showing a large spin-to-charge conversion generated at the MoTe2_2/Py interface. (f) SP-FMR signals for the SiO2_2/MoTe2_2(3)/Au(1.5)/Py(7) heterostructure, where the insertion of a 1.5 nm Au spacer strongly suppresses the SP-FMR signal while leaving the ST-FMR response nearly unchanged, demonstrating that the spin-to-charge conversion originates predominantly at the MoTe2_2/Py interface rather than from the bulk of MoTe2_2. The rf power was 8 mW for the ST-FMR measurements and 110 mW for the SP-FMR measurements.

Conclusions

In this work, we investigated spin-to-charge and charge-to-spin interconversion in sputtered MoTe2_2/Py heterostructures using ST-FMR, SP-FMR, thickness-dependent measurements, angular analysis, and interface-engineering experiments. Our results consistently demonstrate that the dominant conversion mechanism in these heterostructures originates from interfacial Rashba-Edelstein effects rather than from bulk spin Hall transport. The thickness dependence of the ST-FMR and SP-FMR differs markedly from that observed in Pt reference samples. While Pt exhibits the expected asymptotic behavior associated with bulk spin diffusion and the spin Hall effect, MoTe2\mathrm{MoTe}_2 shows nearly thickness-independent conversion efficiencies, supporting an interface-dominated mechanism. Thus, the spin accumulation responsible for the observed torques is primarily generated at the MoTe2/Py\mathrm{MoTe}_2/\mathrm{Py} interface through Rashba-Edelstein mechanisms. Angular-dependent ST-FMR measurements further revealed unconventional torque symmetries, including out-of-plane torque components absent from the conventional angular dependence of polycrystalline metallic bilayers. These components indicate additional symmetry breaking at the interface, potentially associated with structural inversion asymmetry, interfacial disorder, strain, or Rashba-induced spin textures. The extracted damping-like torque efficiencies reached values as high as ξDL≈0.39\xi_{\mathrm{DL}} \approx0.39, demonstrating that sputtered MoTe2\mathrm{MoTe}_2 can generate highly efficient spin-orbit torques comparable to or exceeding those reported for conventional heavy metals and high-quality crystalline TMD systems. The interfacial origin of the conversion was further supported by Au spacer-layer experiments. Inserting a thin Au layer between MoTe2\mathrm{MoTe}_2 and Py strongly suppressed the angular asymmetry and reduced the enhanced conversion signals, indicating that the unconventional torques and large spin-charge conversion efficiencies are intrinsically linked to the direct MoTe2/Py\mathrm{MoTe}_2/\mathrm{Py} interface. Moreover, trilayer SP-FMR measurements using opposite spin-current injection directions revealed voltage polarities consistent with the inverse Rashba-Edelstein effect, distinguishing the interfacial conversion mechanism in MoTe2\mathrm{MoTe}_2 from the conventional inverse spin Hall effect observed in Pt-based systems. Overall, these results demonstrate that sputtered MoTe2\mathrm{MoTe}_2 heterostructures provide an efficient and scalable platform for interface-driven spin-orbit torques. Beyond clarifying the microscopic origin of spin-charge conversion in polycrystalline MoTe2\mathrm{MoTe}_2, this work highlights the importance of interface engineering in TMD-based spintronic devices and opens promising perspectives for low-power magnetic memory and logic technologies based on two-dimensional materials.

Data Availability Statement

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

ACKNOWLEDGMENTS

This research was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) (Grant No. 0041/2022), Financiadora de Estudos e Projetos (FINEP), Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE) – (Grant No. BFP-0345 1.05/24), Universidade Federal de Pernambuco, Multiuser Laboratory Facilities of DF-UFPE, Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) - Rede de Pesquisa em Materiais 2D and Rede de Nanomagnetismo, INCT of Spintronics and Advanced Magnetic Nanostructures (INCT-SpinNanoMag), CNPq 406836/2022-1, and INCT Nanocarbono e Materiais 2D, CNPq 408649/2024-0.

SAMPLE FABRICATION AND CHARACTERIZATION

All thin films were fabricated by DC magnetron sputtering under a base pressure 1.5×10−71.5 \times10^{-7} torr and an argon working pressure 2.7×10−32.7 \times10^{-3} torr. The MoTe2\mathrm{MoTe}_2 and Py thin films were grown with applied DC power 10 W and 23 W, respectively. For the heterostructures SiO2/MoTe2/Py\mathrm{SiO}_2/\mathrm{MoTe}_2/\mathrm{Py}, the silicon substrate wafer was the same, but the sample’s geometry was relatively different depending on the employed technique. For ST-FMR, the sample’s dimensions were 0.150 x 2 mm, while in SP-FMR were 2 x 3 mm. For the electrical voltage detection, conductive silver ink was used to perform the electrical contacts. The Py layers were deposited onto the MoTe2\mathrm{MoTe}_2 without breaking the vacuum chamber. The MoTe2\mathrm{MoTe}_2 samples were characterized by grazing incidence X-ray diffraction (XRD), Raman spectroscopy, resistivity measurements and atomic force microscopy (AFM) images. The results are present in Figure A1. The XRD pattern is shown in Figure A1, and it was measured using KαK_{\alpha} radiation in a 50 nm thick MoTe2\mathrm{MoTe}_2, capped by 1 nm of Au to avoid oxidation. The data reveals the polycrystalline growth of sputtered MoTe2\mathrm{MoTe}_2. The Raman spectrum (Figure A1 (b)) was measured using a 532 nm laser radiation. From the data, we observe that our MoTe2\mathrm{MoTe}_2 films are in the 1T′1\mathrm{T}' structural phase, characterized by the presence of Ag\mathrm{A}_g vibrational modes with narrow peaks, indicating high crystallinity and large grain sizes. The resistivity measurement was acquired using a four-point probe system with Au contacts electrodes, each one 2.5 mm apart. The obtained resistivity (ρ=9.69⋅10−4 Ω⋅m\rho= 9.69 \cdot10^{-4}\,\Omega\cdot\mathrm{m}) is compatible with the resistivity of 1T′1\mathrm{T}' MoTe2\mathrm{MoTe}_2 samples, with minor differences attributed to film thickness, substrate growth and surface roughness. To verify the film quality, we probe the MoTe2\mathrm{MoTe}_2 surface using AFM in tapping mode. The surface topography image shows a granular morphology with large grain sizes (Figure A1(d)). The mean surface roughness is 1.87 nm.

X-ray diffraction pattern, Raman spectroscopy, resistivity curve, and AFM images of MoTe2 samples

Figure A1. (a) X-ray diffraction pattern for SiO2/MoTe2(20)/Au(1)\mathrm{SiO}_2/\mathrm{MoTe}_2(20)/\mathrm{Au}(1). (b) Raman spectroscopy data for the sample SiO2/MoTe2(20)/Py(1)\mathrm{SiO}_2/\mathrm{MoTe}_2(20)/\mathrm{Py}(1), where the numbers are the layer thickness in nm. The Raman data reveals the presence of 1T′1\mathrm{T}' structural phase (c) I x V characteristic curve measured through the four-point probe method. The resistivity from the I x V curve gives compatible resistivity values for the 1T′1\mathrm{T}' MoTe2\mathrm{MoTe}_2. (d) AFM surface topography image showing a granular morphology with uniform grain distribution. (e) Three-dimensional (3D) AFM surface representation.

Experimental Setups and Measurements

The ST-FMR signal is a rectified DC voltage that can be measured directly by a nanovoltmeter or indirectly by lock-in detection. We used lock-in detection because it provides a larger signal-to-noise ratio. The two ends of the NM/FM layer channel are connected to electrodes in a Signal-Ground-Signal coplanar waveguide (CPW), as previously illustrated in Figure 1(b), having a symmetric configuration and impedance of 50Ω50 \Omega. The electric connection is made by using conductive silver ink. Before passing through the sample, the rf current is amplitude modulated by a low frequency sinusoidal (7.3 kHz) signal provided by a signal generator. This signal also works as a reference signal for the lock-in. The ST-FMR signal originates from the oscillating AMR caused by the magnetization precession induced by the oscillating torques. Thus, the voltage is rectified because of the combination of oscillating AMR and oscillating rf current. The resulting signal is detected by the lock-in amplifier using the phase locking technique, having symmetric and antisymmetric Lorentzian lineshapes.

References

  1. [1]K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva, and A. A. Firsov. Electric Field Effect in Atomically Thin Carbon Films. Science, 306, Issue 5696 (2004). DOI: 10.1126/science.1102896DOI
  2. [2]S. Z. Butler, S. M. Hollen, L. Cao, et al. Progress, Challenges, and Opportunities in Two-Dimensional Materials Beyond Graphene. ACS Nano 7 (4): 2898-2926 (2013). DOI: 10.1021/nn400280cDOI
  3. [3]G. F. Fiori, F. Bonaccorso, G. Iannaccone, et al. Electronics based on two-dimensional materials. Nature Nanotech 9, 768-779 (2014). DOI: 10.1038/nnano.2014.207DOI
  4. [4]K. S. Kim, J. Kwon, H. Ryu, et al. The future of two-dimensional semiconductors beyond Moore’s law. Nat. Nanotechnol. 19, 895-906 (2024). DOI: 10.1038/s41565-024-01695-1DOI
  5. [5]M. C. Lemme, D. Akinwande, C. Huyghebaert, et al. 2D materials for future heterogeneous electronics. Nat. Commun. 13, 1392 (2022). DOI: 10.1038/s41467-022-29001-4DOI
  6. [6]S. Manzeli, D. Ovchinnikov, D. Pasquier, et al. 2D transition metal dichalcogenides. Nat. Rev. Mater. 2, 17033 (2017). DOI: 10.1038/natrevmats.2017.33DOI
  7. [7]Y. Wang, S. Sarkar, H. Yan, and M. Chhowalla. Critical challenges in the development of electronics based on two-dimensional transition metal dichalcogenides. Nat. Electron. 7, 638-645 (2024). DOI: 10.1038/s41928-024-01210-3DOI
  8. [8]D. Voiry, A. Mohite, and M. Chhowalla. Phase engineering of transition metal dichalcogenides. Chem. Soc. Rev. 44 (9): 2702-2712 (2015). DOI: 10.1039/c5cs00151jDOI
  9. [9]A. Kuc, T. Heine, and A. Kis. Electronic properties of transition-metal dichalcogenides. MRS bulletin, 40(7), 577-584 (2015). DOI: 10.1557/mrs.2015.143DOI
  10. [10]H. Yang, S. O. Valenzuela, M. Chshiev, et al. Two-dimensional materials prospects for non-volatile spintronic memories. Nature 606, 663-673 (2022). DOI: 10.1038/s41586-022-04768-0DOI
  11. [11]N. Zibouche, A. Kuc, J. Musfeldt, and T. Heine. Transition-metal dichalcogenides for spintronic applications. Ann. Phys. 526 395-401 (2014). DOI: 10.1002/andp.201400137DOI
  12. [12]S. Husain, R. Gupta, A. Kumar, et al. Emergence of spin–orbit torques in 2D transition metal dichalcogenides: A status update. Appl. Phys. Rev. 7, 041312 (2020); DOI: 10.1063/5.0025318DOI
  13. [13]Y. Liu, and Q. Shao. Two-dimensional materials for energy-efficient spin-orbit torque devices. ACS nano, 14(8), 9389-9407 (2020). DOI: 10.1021/acsnano.0c04403DOI
  14. [14]X. Lin, W. Yang, K. L. Wang, and W. Zhao. Two-dimensional spintronics for low-power electronics. Nat. Electron. 2, 274-283 (2019). DOI: 10.1038/s41928-019-0273-7DOI
  15. [15]Y. Deng, X. Zhao, C. Zhu, P. Li, R. Duan, G. Liu, and Z. Liu. MoTe₂: semiconductor or semimetal? ACS nano, 15(8), 12465-12474 (2021). DOI: 10.1021/acsnano.1c01816DOI
  16. [16]Z. Wang, D. Gresch, A. A. Soluyanov, et al. MoTe₂: A Type-II Weyl Topological Metal. Phys. Rev. Letts. 117, 056805 (2016). DOI: 10.1103/PhysRevLett.117.056805DOI
  17. [17]H.-J. Kim, S.-H. Kang, I. Hamada, and Y.-W. Son. Origins of the structural phase transitions in MoTe₂ and WTe₂. Phys. Rev. B 95, 180101(R). DOI: 10.1103/PhysRevB.95.180101
  18. [18]M. J. Mleczko, A. C. Yu, C. M. Smyth, et al. Contact engineering high-performance n-type MoTe₂ transistors. Nano letters, 19(9), 6352-6362 (2019). DOI: 10.1021/acs.nanolett.9b02497DOI
  19. [19]Q. Shao, G. Yu, Y.-W. Lan, et al. Strong Rashba-Edelstein effect-induced spin-orbit torques in monolayer transition metal dichalcogenide/ferromagnet bilayers. Nano letters, 16(12), 7514-7520 (2016). DOI: 10.1021/acs.nanolett.6b03300DOI
  20. [20]S. Song, D. H. Keum, S. Cho, et al. Room Temperature Semiconductor-Metal Transition of MoTe2 Thin Films Engineered by Strain. Nano Lett. 16 (1): 188-193 (2016). DOI: 10.1021/acs.nanolett.5b03481DOI
  21. [21]Y. Wang, J. Xiao, H. Zhu, et al. Structural phase transition in monolayer MoTe2 driven by electrostatic doping. Nature 550, 487-491 (2017). DOI: 10.1038/nature24043DOI
  22. [22]P. Li, W. Wu, Y. Wen, et al. Spin-momentum locking and spin-orbit torques in magnetic nano-heterojunctions composed of Weyl semimetal WTe2. Nat. Commun. 9, 3990 (2018). DOI: 10.1038/s41467-018-06518-8DOI
  23. [23]Q.-F. Yao, J. Cai, W. Tong, et al. Manipulation of the large Rashba spin splitting in polar two-dimensional transition-metal dichalcogenides. Phys. Rev. B, 95(16), 165401 (2017). DOI: 10.1103/PhysRevB.95.165401DOI
  24. [24]A. Manchon, H. C. Koo, J. Nitta, S. M. Frolov, and R. A. Duine. New perspectives for Rashba spin-orbit coupling. Nature materials, 14(9), 871-882 (2015). DOI: 10.1038/nmat4360DOI
  25. [25]M. Yama, M. Matsuo, and T. Kato. Theory of inverse Rashba-Edelstein effect induced by spin pumping into a two-dimensional electron gas. Phys. Rev. B, 108(14), 144430 (2023). DOI: 10.1103/PhysRevB.108.144430DOI
  26. [26]S. Paul, S. Talukdar, R. S. Singh, and S. Saha. Topological phase transition in MoTe2: a review. physica status solidi (RRL)-Rapid Research Letters, 17(6), 2200420 (2023). DOI: 10.1002/pssr.202200420DOI
  27. [27]J. Jiang, Z. K. Liu, Y. Sun, et al. Signature of type-II Weyl semimetal phase in MoTe2. Nature commun., 8(1), 13973 (2017). DOI: 10.1038/ncomms13973DOI
  28. [28]A. Zhang, X. Ma, C. Liu, et al. Topological phase transition between distinct Weyl semimetal states in MoTe2. Phys. Rev. B, 100(20), 201107 (2019). DOI: 10.1103/PhysRevB.100.201107DOI
  29. [29]S. T. Chyczewski, H. Lee, S. Li, M. Eladl, J.-F. Zheng, A. Hoffmann, and W. Zhu. Strong Damping-Like Torques in Wafer-Scale MoTe2 Grown by MOCVD. ACS Applied Materials & Interfaces, 17(14), 21996-22003 (2025). DOI: 10.1021/acsami.4c21247DOI
  30. [30]L. Zhu, D. C. Ralph, and R. A. Buhrman. Spin-orbit torques in heavy-metal-ferromagnet bilayers with varying strengths of interfacial spin-orbit coupling. Phys. Rev. Lett., 122(7), 077201 (2019). DOI 10.1103/PhysRevLett.122.077201DOI
  31. [31]S. Liang, S. Shi, C.-H. Hsu, et al. Spin-orbit torque magnetization switching in MoTe2/permalloy heterostructures. Advanced Materials, 32(37), 2002799 (2020). DOI: 10.1002/adma.202002799DOI
  32. [32]S. Li, J. Gibbons, S. Chyczewski, et al. Unconventional spin-orbit torques from sputtered MoTe2 films. Phys. Rev. B 110, 024426 2024). DOI: 10.1103/PhysRevB.110.024426DOI
  33. [33]D. MacNeill, G. M. Stiehl, M. H. D. Guimaraes, R. A. Buhrman, J. Park, and D. C. Ralph. Control of spin-orbit torques through crystal symmetry in WTe2/ferromagnet bilayers. Nature Physics, 13(3), 300-305 (2017). DOI: 10.1038/nphys3933DOI
  34. [34]G. M. Stiehl, R. Li, V. Gupta, et al. Layer-dependent spin-orbit torques generated by the centrosymmetric transition metal dichalcogenide β-MoTe2. Phys. Rev. B, 100(18), 184402 (2019). DOI: 10.1103/PhysRevB.100.184402DOI
  35. [35]W. Choi, N. Choudhary, G. H. Han, J. Park, D. Akinwande, and Y. H. Lee. Recent development of two-dimensional transition metal dichalcogenides and their applications. Materials Today, 20(3), 116-130, (2017). DOI: 10.1016/j.mattod.2016.10.002DOI
  36. [36]S. Xu, J. Zhang, Y. Xia, et al. Single-Site-Directed Unidirectional Epitaxy of Large-Scale 2D Materials. Advanced Materials 38, 13: e20837 (2026). DOI: 10.1002/adma.202520837DOI
  37. [37]C. Bian, Y. Zhao, R. Guzman, et al. Atomically precise synthesis and simultaneous heterostructure integration of 2D transition metal dichalcogenides through nano-confinement. Nat. Mater. 25, 1321-1328 (2026). DOI: 10.1038/s41563-026-02495-9DOI
  38. [38]D. S. Cabeda, A. V. da Silva, and C. Radtke. Influence of the Magnetron Sputtering Power Source Type on the Formation of TMD Thin Films. Crystal Growth & Design 26(4): 1555-1561 (2026). DOI: 10.1021/acs.cgd.5c01221DOI
  39. [39]Y. Tserkovnyak, A. Brataas, and G. E. W. Bauer. Enhanced Gilbert damping in thin ferromagnetic films. Phys. Rev. Letts. 88 (11), 117601 (2002). DOI: 10.1103/PhysRevLett.88.117601DOI
  40. [40]Y. Tserkovnyak, A. Brataas, and G. E. W. Bauer. Spin pumping and magnetization dynamics in metallic multilayers. Phys. Rev. B 66, 224403 (2002). DOI: 10.1103/PhysRevB.66.224403DOI
  41. [41]A. Azevedo, L. H. Vilela Leão, R. L. Rodriguez-Suarez, A. B. Oliveira, and S. M. Rezende. dc effect in ferromagnetic resonance: Evidence of the spin-pumping effect? J. Appl. Phys. 97, 10C715 (2005). DOI: 10.1063/1.1855251DOI
  42. [42]E. Saitoh, M. Ueda, H. Miyajima, and G. Tatara. Conversion of spin current into charge current at room temperature: Inverse spin-Hall effect. Applied physics letters, 88(18) (2006). DOI: 10.1063/1.2199473DOI
  43. [43]A. Azevedo, L. H. Vilela-Leão, R. L. Rodríguez-Suárez, A. F. L. Santos, and S. M. Rezende. Spin pumping and anisotropic magnetoresistance voltages in magnetic bilayers: Theory and experiment. Phys. Rev. B, 83, 144402 (2011). DOI:10.1103/PhysRevB.83.144402DOI
  44. [44]J. C. Sánchez, L. Vila, G. Desfonds, et al. Spin-to-charge conversion using Rashba coupling at the interface between non-magnetic materials. Nat. Commun. 4, 2944 (2013). DOI: doi.org/10.1038/ncomms3944
  45. [45]G. Bihlmayer, P. Noёl, D. V. Vyalikh, et al. Rashba-like physics in condensed matter. Nat. Rev. Phys. 4, 642-659 (2022). DOI: 10.1038/s42254-022-00490-yDOI
  46. [46]J. L. Costa, E. Santos, G. R. Gallo, G. Rodrigues-Junior, et al. Phase-selective orbital-charge conversion in MoTe₂. arXiv:2607.01623 (2026)arxiv.org/abs/2607.01623
  47. [47]Y. S. Gui, L. H. Bai, and C. M. Hu. The physics of spin rectification and its application. Sci China-Phys Mech Astron, 56: 124-141, (2013). DOI: 10.1007/s11433-012-4956-6.
  48. [48]A. Azevedo, R. O. Cunha, F. Estrada, et al. Electrical detection of ferromagnetic resonance in single layers of permalloy: Evidence of magnonic charge pumping. Phys. Rev. B, 92, 024402 (2015). DOI: 10.1103/PhysRevB.92.024402DOI
  49. [49]A. Tsukahara, Y. Ando, Y. Kitamura, et al. Self-induced inverse spin Hall effect in permalloy at room temperature. Phys. Rev. B, 89, 235317 (2014). DOI: 10.1103/PhysRevB.89.235317DOI
  50. [50]L. Bai, P. Hyde, Y. S. Gui, et al. Universal Method for Separating Spin Pumping from Spin Rectification Voltage of Ferromagnetic Resonance. Phys. Rev. Lett. 111, 217602 (2013). DOI: 10.1103/PhysRevLett.111.217602DOI
  51. [51]J. Shen, Z. Feng, P. Xu, D. Hou, Y. Gao, and X. Jin. Spin-to-Charge Conversion in Ag/Bi Bilayer Revisited. Phys. Rev. Lett. 126, 197201 (2021). DOI: 10.1103/PhysRevLett.126.197201.DOI
  52. [52]J. Cheng, B. F. Miao, Z. Liu, et al. Coherent Picture on the Pure Spin Transport between Ag/Bi and Ferromagnets. Phys. Rev. Lett. 129, 097203 (2022). DOI: 10.1103/PhysRevLett.129.097203DOI
  53. [53]J. E. Abrão, E. G. da Silva, G. Rodrigues-Junior, J. B. S. Mendes, and A. Azevedo. Probing the spin-momentum locking on Rashba surfaces via spin current. ACS Applied Materials & Interfaces, 17(9), 13162-13169 (2024). DOI: 10.1021/acsami.4c06090.DOI
  54. [54]E. C. Souza, J. D. M. de Lima, J. L. Costa, G. Rodrigues-Junior, R. O. Cunha, J. B. S. Mendes, and S. M. Rezende. Spin-to-charge conversion mediated by the inverse Rashba–Edelstein effect in polycrystalline thin films of the Weyl semimetal TaP. Journal of Applied Physics, 138(6) (2025). DOI: 10.1063/5.0281343DOI
  55. [55]E. G. da Silva, J. E. Abrão, E. Santos, S. Bedanta, H. F. Ding, J. B. S Mendes, and A. Azevedo. Appl. Phys. Lett. 123, 202402 (2023). DOI: 10.1063/5.0169242.DOI
  56. [56]A. Manchon, J. Železný, I. M. Miron, et al. Current-induced spin-orbit torques in ferromagnetic and antiferromagnetic systems. Rev. Mod. Phys. 91, 035004 (2019). DOI: 10.1103/RevModPhys.91.035004DOI
  57. [57]M. B. Jungfleisch, W. Zhang, J. Sklenar, et al. Interface-driven spin-torque ferromagnetic resonance by Rashba coupling at the interface between nonmagnetic materials. Phys. Rev. B, 93, 224419 (2016). DOI: 10.1103/PhysRevB.93.224419DOI
  58. [58]S Karimeddiny, and D. C. Ralph. Resolving discrepancies in spin-torque ferromagnetic resonance measurements: Lineshape versus linewidth analyses. Phys. Rev. Applied, 15, 064017 (2021). DOI: 10.1103/PhysRevApplied.15.064017.DOI
  59. [59]L. Liu, T. Moriyama, D. C. Ralph, and R. A. Buhrman. Spin-Torque Ferromagnetic Resonance Induced by the Spin Hall Effect. Phys. Rev. Lett., 106, 036601 (2011). DOI: 10.1103/PhysRevLett.106.036601DOI
  60. [60]M.-H. Nguyen, and C.-F. Pai. Spin-orbit torque characterization in a nutshell. APL Mater., 9, 030902 (2021). DOI: 10.1063/5.0041123.DOI
  61. [61]J. L. Costa, E. Santos, A. Y. M. Tani, et al. Investigating spin and orbital effects via spin-torque ferromagnetic resonance. J. Appl. Phys., 139, 243904 (2026). DOI: 10.1063/5.0320961.DOI
  62. [62]N. Soya, S. Yoshikawa, T. Katase, and K. Ando. Spin-torque ferromagnetic resonance in SrTiO₃-based systems: impact of out-of-phase Oersted field torque. npj Spintronics 3, 38 (2025). DOI: 10.1038/s44306-025-00102-2.DOI

Paper details

vlnc:W_q98eeewn
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents