Abstract
Twodimensional (2D) ferromagnetic materials with unique magnetic properties have great potential for nextgeneration spintronic devices with high flexibility, easy controllability, and high heretointegrability. However, realizing magnetic switching with low power consumption at room temperature is challenging. Here, we demonstrate the roomtemperature spinorbit torque (SOT) driven magnetization switching in an allvan der Waals (vdW) heterostructure using an optimized epitaxial growth approach. The topological insulator Bi_{2}Te_{3} not only raises the Curie temperature of Fe_{3}GeTe_{2} (FGT) through interfacial exchange coupling but also works as a spin current source allowing the FGT to switch at a low current density of ~2.2×10^{6} A/cm^{2}. The SOT efficiency is ~2.69, measured at room temperature. The temperature and thicknessdependent SOT efficiency prove that the larger SOT in our system mainly originates from the nontrivial topological origin of the heterostructure. Our experiments enable an allvdW SOT structure and provides a solid foundation for the implementation of roomtemperature allvdW spintronic devices in the future.
Similar content being viewed by others
Introduction
Spintransfer torque^{1,2,3,4,5} magnetic random access memory (STTMRAM) is an appealing alternative to overcome the performance bottleneck encountered in traditional semiconductorbased memory, offering superior performance in terms of nonvolatility, high density, and lowpower dissipation. However, the initiative parallel or antiparallel collinear magnetic configuration would lead to an incubation delay when using STT for magnetic switching, and the large writing current could break down the tunneling barrier. In comparison, spin–orbit torque (SOT)^{6,7,8,9,10} could eliminate such performance drawbacks and allow for faster operation, better endurance, and higher energy efficiency. Therefore, it is of fundamental and technical importance to use SOT for switching the magnetization, which is expected to become the major competitor for nextgeneration memories^{11,12,13,14,15}.
To date, great efforts have been devoted to exploring new principles and materials for realizing highperformance SOT devices^{16,17,18,19,20}. Usually, heavy metals that include W, Ta, Pt, etc., were employed as the spin current sources through chargespin conversion, which could exert a torque on the adjacent ferromagnetic layer for magnetization switching^{21,22,23,24}. For higher SOT efficiency, van der Waal (vdW) topological insulators (TIs) were recently suggested as a replacement for heavy metals due to their unique feature of spinmomentum locking in the nontrivial topological surface state (TSS). This has been demonstrated to allow for highefficiency SOTdriven magnetic switching in threedimensional (3D) ferromagnet at room temperature with low critical switching current^{17,25,26,27,28}. However, 3D ferromagnets would limit the size scaling and lower the spin transparency due to the dangling bonding interface^{29}. Therefore, there is a need to develop new material systems with lower dimensions and superior interfaces for higher SOT efficiency^{30,31,32}, which may bring new opportunities to break the power consumption bottleneck of integrated circuits^{33}.
The recentlydiscovered vdW 2D ferromagnetic materials offer an atomic flat surface and can maintain their magnetic ordering down to the 2D limit, which would satisfy such demand. The Mermin–Wagner–Hohenberg (MWH) theorem predicted that thermal fluctuations in a 2D magnetic system^{34,35} forbade the longrange magnetic order at finite temperature because the continuous symmetry could not be spontaneously broken in a 2D system. However, recently it has been discovered that 2D intrinsic ferromagnetic materials could exist through breaking the MWH theorem by magnetic anisotropy such as Fe_{3}GeTe_{2} (FGT)^{36}, CrI_{3}^{37}, and Cr_{2}Ge_{2}Te_{6}^{38}, among others. FGT has received extensive attention by virtue of its hard magnetic properties, Kondo lattice behavior, itinerant ferromagnetism, and other fascinating characteristics^{39,40,41,42,43}. Remarkably, intercalating lithium ions into the interlayer gap of FGT can change the density of states on the Fermi surface and successfully raise the T_{c} to room temperature^{44}. These inspiring results suggest FGT is an ideal 2D candidate for exploring SOTdrivenmagnetic switching. Recently, the SOTdriven magnetization switching of FGT has been demonstrated using Pt as a spin current source through the spin Hall effect or interfacial Rashba–Edelstein effect^{45,46}. However, these devices only work at low temperatures (<200 K). Furthermore, much higher SOT efficiency could be envisaged through constructing the allvdW heterostructure, which can provide a clean interface and thus support high interfacial spin transparency. Therefore, there is an urgent need to design allvdW heterostructures to achieve energyefficient SOT switching that can operate at room temperature for future 2D spintronic applications.
Here, we realize SOTdriven magnetic switching in an MBEgrown allvdW Bi_{2}Te_{3}/FGT heterostructure at room temperature. The SOTinduced magnetization switching is achieved with a critical switching current density of ~2.2 × 10^{6} A/cm^{2}. The dampinglike SOT efficiency was calculated to be about ~2.69 at room temperature. The high efficiency proves the superior characteristics of allvdW heterostructures constructed from 2D ferromagnetic materials. We analyze the difference between the largefield powerlaw fitting and the smallfield derivation fitting from the harmonic measurements. In particular, the weak vdW interactions between adjacent layers make it possible to combine atomic layers with different matching degrees, thereby getting rid of lattice matching and compatibility restrictions. The highquality heterostructure interface is one of the most important factors for achieving high spin transmissivity. Our results provide a paradigm for the construction of allvdW SOT devices at room temperature and promote the development of 2D ferromagnets for practical applications.
Results
Magnetotransport measurements in Bi_{2}Te_{3}, FGT, and Bi_{2}Te_{3}/FGT heterostructures
In this work, we deposited thin films on a (0001) sapphire substrate by MBE, combined with the reflection highenergy electron diffraction (RHEED) to in situ monitor the surface structure of the film during the preparation, and analyzed the surface morphology by atomic force microscopy (AFM). When preparing the waferscale allvdW heterostructure, it is very critical to maintain the surface flatness of the bottom layer to ensure the optimal lattice matching and compatibility of the two layers. Therefore, after growing topological insulators on (0001) sapphire substrates, the growth temperature needs to be slowly increased in the growth chamber to maintain a Terich environment, which will ensure an excellent single crystallinity of the heterostructure. To better understand the sample quality, RHEED was in situ rotated during the growth process to check the inplane crystallinity. During the rotation, RHEED stripes changed regularly and coherently, which could exclude the presence of multidomain. To prevent its degradation, we covered the top surface of FGT with a protective layer. Micrometersized Hallbar devices were fabricated by the standard photolithography combined with ion beam etching. The schematic of the device and the measurement setup are shown in Fig. 1a. The Hallbar structure was patterned with the dimensions of 100 μm (length) × 30 μm (width) for electrical transport measurements, as shown in Fig. 1b, where V_{xy} and V_{xx} represent the Hall and longitudinal voltage, respectively. As an emergent quantum matter, TIs attract a lot of interest due to the bulk gap and the spinmomentumlocked Dirac fermions on the surface. Hence, for these types of materials, such as Bi_{2}Te_{3} and Bi_{2}Se_{3}, it has been proved by both the theory and experiment that surface states consist of a single Dirac cone at the Γ point, and its simplicity has become an ideal object for studying the spintronics and electronics simultaneously. In the following, we grew 8 nm Bi_{2}Te_{3} on the sapphire substrate and performed the magnetotransport measurement. By applying an outofplane magnetic field, the Hall resistance shows a negative slope, as shown in Fig. 1c, which features the ntype Bi_{2}Te_{3}. The right inset in Fig. 1c presents the temperaturedependent 2D carrier density (n_{2D})^{47} obtained from the Hall data, which reflects the conduction dominantly from the bulk state. The left inset of Fig. 1c elucidates the band structure of Bi_{2}Te_{3}, and the position of the Fermi level (E_{F}) determines the spinmomentum locking properties. As a layered vdW crystal, FGT has metallic ferromagnetism^{36,41}. Each vdW layer is composed of five atomic sublayers with the lattice constants of a = b = 3.9536 (7) Å and c = 16.396 (2) Å. During the growth of FGT in the MBE chamber, the stripelike RHEED pattern was captured, reflecting an atomic smooth interface, and its crystal structure was further characterized by Xray diffraction (XRD) (Supplementary Fig. S1). To clarify the magnetic behavior of FGT, we conducted magnetotransport measurements by a physical property measurement system (PPMS) on a 30 nm FGT thin film. Figure 1d clearly shows the temperaturedriven transition from a ferromagnetic to a paramagnetic state with the T_{c} around 220 K, which is the same as that in the previous report^{44}. Figure 1e shows the hysteresis loops between 80 K and 300 K in Bi_{2}Te_{3}(8)/FGT(3) heterostructure with an inplane magnetic field, which verified its perpendicular magnetic anisotropy (PMA) feature. The number enclosed in brackets denotes the thickness of the individual layer in nanometers. Furthermore, the saturation magnetization (M_{s}) and the magnetic properties in Bi_{2}Te_{3}(8)/FGT(3) were characterized by a superconducting quantum interference device (SQUID) in Fig. 1f, clearly showing the room temperature ferromagnetism. The fascinating phenomenon of the combination of topological insulators and magnetic 2D materials lays the foundation of our current research. The highangle annular darkfield scanning transmission electron microscopy (HAADFSTEM) in the inset further confirms the atomic structure of Bi_{2}Te_{3} and FGT.
Currentinduced SOT switching in Bi_{2}Te_{3}/FGT heterostructure
Here, Bi_{2}Te_{3}(8)/FGT(3) was taken as the research object for SOT switching. Figure 2a shows the geometric diagram of SOTdriven magnetic switching dynamics in the vdW heterostructure of Bi_{2}Te_{3} and FGT. By injecting sufficient spin current density (J_{spin}) from Bi_{2}Te_{3}, SOT enables the magnetization (M) switching in the adjacent ferromagnetic layer above the critical writing current density (J_{write}). It is worth noting that the injected J_{write} is orthogonal to the accumulated spin polarization direction and the generated J_{spin} direction. Here, J_{write} can be determined as \({J}_{{{{{\rm{write}}}}}}={I}_{{{{{\rm{write}}}}}}/\left[w*\left({t}_{{{{{\rm{{Bi}}}}}}_{2}{{Te}}_{3}}+{t}_{{{{{\rm{FGT}}}}}}\right)\right]\), where w = 30 μm is the width of the Hallbar^{25}. The effective spin–orbit field (H_{so}) induced by the spin current is along the tangential of M and could tilt M up or down to get the positive or negative z component (M_{z}). Usually, for PMA samples, an additional external magnetic field (H_{ext}) needs to be applied during the measurement process to break the mirror symmetry for the deterministic SOT switching^{48}. Thus, when sweeping the applied charge current with an external inplane field, the SOT from the chargespin conversion in Bi_{2}Te_{3} would induce the magnetization reversal in the ferromagnetic layer.
To demonstrate the SOT switching in the Bi_{2}Te_{3}/FGT heterostructure, a series of inplane magnetic fields were applied with a 10ms pulse current along the Hall bar to obtain a deterministic switch polarity. We observe that the M changes steadily as the applied J_{write} increases, and a complete magnetic switching is achieved when the J_{write} reaches approximately 4 × 10^{6} A/cm^{2} at 200 K, as shown in Fig. 2b. When the current density is greater than 2.5 ×10^{6} A/cm^{2}, Hallresistance (R_{H}) begins to decrease after reaching the maximum, which is related to a Joule heating effect. Our explanation for this case is that the magnetic interactions of FGT are not sufficient to fight against the thermal fluctuations, resulting in a decrease of M^{46}. Interestingly, the critical J_{write} for SOT switching is much smaller than the values reported in FGT/Pt heterostructure, probably due to the high efficiency of chargespin conversion in TI nontrivial origin. As the applied magnetic field is reversed, the opposite chirality of the SOT switching curve demonstrates the typical characteristics of SOT in the PMA sample, as shown in Fig. 2c. Moreover, the device was measured at different temperatures of 210 K and 190 K. We found that as the temperature decreases, the range of the applied magnetic field to achieve the SOT switch gradually increases (Supplementary Fig. S2 and Fig. S3). For an indepth understanding of the switching behavior, we summarize the dependence of the current density on the applied magnetic field for SOT switching at different temperatures in the phase diagram of Fig. 2d. Here, the critical switching current density (J_{sw}) that is defined as the sign change in R_{H} is gradually reduced at the higher magnetic field. The deterministic switching happens in the large field and current region, while both up and down magnetization states are possible in the intermediate region associated with a small field and current region. Additionally, the switching current decreases with increasing temperature, which is attributed to the simultaneous decrease in M_{s}, as already proved in Fig. 1f.
Harmonic Hall measurements in Bi_{2}Te_{3}/FGT heterostructure
To quantitatively evaluate the SOT efficiency, we use the harmonic Hall measurement to characterize the effective field of SOT, which could provide a solid understanding of each SOT component, as well as its influencing factors. We apply a small sinusoidal current (J_{a.c.}) to the channel of the device and then generate a SOT in the ferromagnetic layer, which will be decomposed into two mutually orthogonal vector components: dampinglike torque \({\tau }_{{{{{\rm{DL}}}}}} \sim m\times \left(\sigma \times m\right)\) and fieldlike torque \({\tau }_{{{{{\rm{FL}}}}}} \sim \sigma \times m\)^{49}. In the measurement, the frequency is fixed at 133.33 H_{z} through the lockin amplification, and the magnetization oscillation of M around the equilibrium position generates the harmonic Hall signals, including the inphase first harmonic Hall voltage (V_{1ω}) and outofphase second harmonic Hall voltage (V_{2ω}). We analyzed the secondharmonic anomalous Hall resistance \(({R}_{{{{{\rm{AHE}}}}}}^{2{{{{{\rm{\omega }}}}}}})\) and planar Hall resistance \(({R}_{{{{{\rm{PHE}}}}}}^{2{{\omega }}})\) to determine the currentinduced SOT effective field. After applying an external magnetic field (H_{x}) to the xaxis, the second harmonic Hall resistance \({R}_{{xy}}^{2{{{{{\rm{\omega }}}}}}}\) could be obtained by the following equation for values of H_{x} larger than the magnetic anisotropy field H_{k}:^{27}
where H_{DL} and H_{FL} are the dampinglike effective field (\(\propto m\times \sigma\)) and fieldlike effective field (\(\propto \sigma\)), respectively. R_{p} and R_{A} are the planar Hall resistance and anomalous Hall resistance, respectively. R_{offset} is the resistance offset. R_{ANE} is the transverse resistance contributed by the anomalous Nernst effect and other spinrelated thermoelectric effects^{50}. For the dampinglike effective term, it decreases as the external field increases. For the thermalrelated term, its sign changes as the external field direction reverse, while its magnitude keeps constant. Usually, the R_{p} is extremely small compared to the anomalous Hall counterpart and thus \({R}_{{xy}}^{2{{{{{\rm{\omega }}}}}}}\) mainly originates from the dampinglike effective field term and thermaleffect term. Figure 3a displays a series of \({R}_{{xy}}^{2{{{{{\rm{\omega }}}}}}}{H}_{x}\) curves under different applied J_{a.c.} at 200 K. It demonstrates a distinct field dependence, while a step function could also be observed, which means that in addition to the contribution of the dampinglike Hall signal, it also has a thermal contribution in our sample. With increasing the J_{a.c.}, both signals are enhanced. The inset in Fig. 3a schematically illustrates the second harmonic Hall signal that comes from the SOTinduced magnetization oscillation around the equilibrium position. For quantitatively characterizing the thermal signal, we carried out the temperaturedependent \({R}_{{xy}}^{2{{{{{\rm{\omega }}}}}}}{H}_{x}\) at a fixed J_{a.c.} to provide further evidence. Here, we defined \({R}_{{{{{\rm{ANE}}}}}}=({R}_{{xy}({{{{\rm{sat}}}}}\_{max} )}{R}_{{xy}({{{{\rm{sat}}}}}\_{min} )})/2\) to express the thermal contribution, where \({R}_{{xy}({{{{\rm{sat}}}}}\_{{{{{\mathrm{max}}}}}} )}\) and \({R}_{{xy}({{{{{\rm{sat}}}}}}\_{{{{{\rm{min}}}}}} )}\) are defined as the maximum and minimum values of secondharmonic Hall resistance under a saturated magnetic field^{27}. As temperature decreases, the R_{ANE} becomes much larger, which implies thermal contribution is more pronounced at low temperatures. To understand the origin of the thermalrelated effect, it is worth noting that the metallic and topological nature of FGT could cause a large anomalous Nernst effect (ANE)^{43}. In our sample, the top layer above FGT is air with ambient temperature, while the bottom layer is Bi_{2}Te_{3} with a large current. The vertical thermal gradient from the asymmetric structure may contribute to the thermal current, thus inducing the ANE. Nevertheless, we could differentiate the ANE and the SOTinduced secondharmonic Hall resistance through their magnetic field dependence. Figure 3b displays the influence of the FGT’s large ANE on the heterostructure, and the left inset is a schematic diagram of the step function of the thermal contribution^{51}.
Relying on the above analysis, we extract H_{DL} and display the dependence of H_{DL} on corresponding J_{a.c.} at 200 K, as shown in Fig. 3c. The resistivities of the Bi_{2}Te_{3} and FGT layers of different thicknesses are evaluated (Supplementary Fig. S4). By fitting the process in the large inplane magnetization region with the formula (1), \({H}_{{DL}}/{J}_{{write}}\) is ~160.2 Oe per MA/cm2 in the inset of Fig. 3c. The SOT efficiency (ξ_{DL}) can be obtained using^{52},
where e and ℏ are the electron charge and reduced Plank constant, respectively, t represents the ferromagnetic layer thickness. Accordingly, the value of the ξ_{DL} is determined to be ~5.3 in Bi_{2}Te_{3}(8)/FGT(3) structure at 200 K.
To eliminate the thermal contribution caused by the ANE of FGT on the SOT efficiency, we adjusted the thickness of FGT to manipulate the shunting current in the Bi_{2}Te_{3} for lowering the thermal gradient in the Bi_{2}Te_{3}/FGT heterostructure^{53}. Moreover, we conduct the measurements with different I_{dc} while sweeping H_{x} to observe the variation of R_{xy}, and find that the DC of 0.5 mA to 1.5 mA has no significant effect on the heterostructure, which further verifies that this thickness of the heterostructure has better thermal stability (Supplementary Fig. S5). Figures 4a, b displays the outofplane external magnetic fielddependent R_{xy} on Bi_{2}Te_{3}/FGT heterostructures with varying thicknesses of FGT (3 nm and 4 nm) at 100 K, 150 K, and 200 K. We normalize its R_{xy} to facilitate comparison. It is worth noting that the R_{xy} of Bi_{2}Te_{3}(8)/FGT(3) has an obviously negative ordinary Hall slope in the saturated magnetic field region, which is similar to that from the Bi_{2}Te_{3} Hall signal, indicating that Bi_{2}Te_{3} in the heterostructure has a large shunting effect. In contrast, the R_{xy} of Bi_{2}Te_{3}(8)/FGT(4) shows only the anomalous Hall signal from FGT, which well proves the shunting effect in Bi_{2}Te_{3} has been significantly reduced due to more conducting in FGT after increased thickness. The PMA feature was further verified by performing firstharmonic Hall measurement with an inplane magnetic field, and the results under different temperatures are shown in Fig. 4c. Subsequently, we conducted the secondharmonic Hall measurements and displayed R_{2ω} signals as a function of H_{x} under different J_{write} in Fig. 4d. Interestingly, the step function arising from ANE disappears, which well matches our above prediction. Followed by Eqs. (1) and (2), the room temperature ξ_{DL} is estimated to be ~0.7, which indicates the strong SOC characteristics of TI at room temperature.
To verify the conjecture and understand the related mechanism in our sample, we give a systematic discussion about the temperature dependence of ξ_{DL}. Unlike traditional heavy metals, TI exhibits a topologicallyprotected nontrivial surface state, which is composed of a single massless Dirac fermion with two spinsplitting bands on the surface. When the timereversal symmetry is broken, the surface state will open a gap. The bulk Hamiltonian projected onto the surface state is described as^{54,55}
where \(\upsilon\) is the velocity of the surface state and \(k\) is the Dirac electron momentum. When the J_{write} is applied to TI, the spin of the Dirac electron is locked, and the movement of the Fermi surface in the kspace will produce controllable spin polarization. Another important origin of SOT is the spin Hall effect (SHE) of the bulk state, which utilizes the bulk SOC in TI to convert nonpolarized write current into the spin current. Due to the asymmetric scattering of conductive electrons, the spinup and spindown are deflected in opposite directions, forming a transverse spin current.
Figure 5a displays the schematic spinrelated band structure of the TSS and bulk state. Both of them coexist in the film^{56}, and either the surface or bulk state would provide a contribution to the final SOT. To gain insights into how large surface contribution to the SOT, temperaturedependent SOT efficiency and its relation to the E_{F} were carried out for analysis. Figure 5b shows the precise SOT efficiency results through harmonic Hall measurements in the Bi_{2}Te_{3}(8)/FGT(4) heterostructure. Here, we examined the accuracy of different fitting methods on the calculation results from the perspective of the extended Landau–Lifshitz–Gilbert equation and anisotropy (more details in Supplementary Fig. S6). We found that ξ_{DL} exhibited a drastic nonlinear growth with a decrease in temperature^{57}. At room temperature, the E_{F} predominantly resides within the highly conductive bulk state, but as the temperature decreases, it shifts downwards toward the Dirac cone with a reduced bulk state (Supplementary Fig. S7). The fact that TI with reduced bulk conductance leads to a higher SOT efficiency suggests that the TSS renders significant contributions to the efficient SOT. Furthermore, additional heterostructures with different TI thicknesses (Bi_{2}Te_{3}(6)/FGT(4) and Bi_{2}Te_{3}(10)/FGT(4)) were prepared for comparison with previous samples. The SOT efficiency from the harmonic measurements has undergone a dramatic increase to ~2.69, which further proves the substantial surface contribution of Bi_{2}Te_{3} at room temperature (Supplementary Fig. S8). Besides, it is worth noting that the Rashba spinsplitting surface state in the twodimensional electron gas (2DEG) may coexist with the TSS in Bi_{2}Te_{3} due to the band bending and structural inversion asymmetry^{5,57}. However, the Rashba effective field is expected to increase gradually as the temperature rises in the semiconductor system^{58,59}, different from our experimental results^{60}. Hence, we conclude that the Rashbasplit surface state is not the primary physical mechanism for SOT switching^{26}.
For the chirality of SOT, the relationship between the J_{spin} and the J_{write} can be expressed by the following formula:^{61}
where θ_{SH} is the spin Hall angle, σ is the polarization of the spin, and its direction is orthogonal to the direction of the J_{write}. For nonferromagnetic materials that provide spin currents, the spin direction of the top surface and the bottom surface is opposite, and its chirality is defined by the sign of θ_{SH}. Compared with our results, the SOT switching in FGT/Pt heterostructures shows the same chirality, further accurately confirming our conclusion^{45,46}. As reported previously, the chirality from TSS is the same as that from the bulk state with positive spin Hall angle^{27,62}. Finally, the SOT switching of the FGT layer was successfully demonstrated in the Bi_{2}Te_{3}(8)/FGT(4) heterostructure at room temperature when 10ms pulse currents were applied to the Hall bar with an \({H}_{{{{{\rm{ext}}}}}}=\pm 2{{{{{\rm{kOe}}}}}}\) under several consecutive sweeps, as shown in Fig. 5c. It sets a new stage for exploring allvdW SOT devices.
For clarity, we summarize the switching write current density, SOT efficiency, and its realized maximum temperature of several representative heterostructures for comprehensively understanding the SOT feature in the Bi_{2}Te_{3}/FGT heterostructure, and the results are presented in Table 1. The heavy metal Pt is generally used as the preferred material to achieve SOT switching of FGT at low temperatures. It is worth noting that the minimum ξ_{DL} in the FGT/Pt heterostructure reported by Alghamdi et al. is as large as the maximum of the CoFeB/Pt structure^{45}, demonstrating the vdW FGT superiority. In comparison with our sample, the large ξ_{DL} value well proves the TI of Bi_{2}Te_{3} is superior for chargespin conversion with 2D vdW ferromagnet. Compared to previously reported Bi_{2}Te_{3}based heterostructures, our sample also has a significant advantage in SOT efficiency. Recently, WTe_{2}/FGT heterostructures have been found to achieve SOT properties and relatively excellent performance, but still at low temperature^{63,64}. These results obtained with the same characterization method may provide evidence that the interfacial spin transparency could be significantly enhanced by the vdWgapped interface between Bi_{2}Te_{3} and FGT due to the optimized growth method. In such a case, the ξ_{DL} is related to the internal θ_{SH} of TI and the interfacial spin transparency T_{int}^{29,46}. The interfacial spin transparency could be mainly determined by the mechanisms of spin backflow and spin memory loss, which could be characterized by the effective spinmixing conductance and the spin conductance of the nonferromagnetic layer^{29}. A good interface contributes to the transparency during spin transport at room temperature, which is one of the most important factors in achieving energyefficient SOT switching in an allvdW heterostructure and highlights the strong SOC characteristics of TI.
To summarize, the waferscale vdW Bi_{2}Te_{3}/FGT heterostructure prepared by MBE has successfully realized roomtemperature ferromagnetism and currentdriven SOT switching. We employed the harmonic Hall signals to accurately estimate the SOT efficiency, which was as high as ~0.7 at room temperature, and this value could be further increased to ~2.69 with decreasing TI thickness. Together with the temperaturedependent measurement, the high chargespin conversion efficiency is mainly attributed to the improved interfacial spin transparency and nontrivial topological origin of the allvdW Bi_{2}Te_{3}/FGT heterostructure. The realization of roomtemperature ferromagnetism and SOT switching together in Bi_{2}Te_{3}/FGT heterostructure establishes a promising route for the development of allvdW heterostructures and lays the foundation for implementation of roomtemperature 2D vdW spintronic devices in the future.
Methods
Sample growth
The (0001) sapphire substrate was used to grow the sample. Highpurity Bi, Fe, Ge, and Te were evaporated from Knudsen effusion cells in the MBE system with a base vacuum of 10^{−10} Torr. After degassing at high temperature, the substrate was cooled down to 300 °C for growing both the FGT thin film and Bi_{2}Te_{3}/FGT heterostructure with a growth rate of ~0.05 Å/s, and the sample quality was monitored by an in situ RHEED system.
Characterization
The morphologies of the samples were investigated by AFM. The microstructure and composition were comprehensively characterized by XRD and HAADF in STEM mode. The crosssection TEM sample was prepared by a focused ion beam. MOKE and SQUID were employed to measure their magnetic properties. Furthermore, the magnetotransport studies were carried out in the Quantum Design physical property measurement system.
Data availability
All data generated in this study are provided in the paper and Supplementary Information/Source Data file. Additional data related to this study are available from the corresponding author upon reasonable request.
References
Yoda, H. et al. High efficient spin transfer torque writing on perpendicular magnetic tunnel junctions for high density MRAMs. Curr. Appl. Phys. 10, e87–e89 (2010).
Min, T. et al. A study of write margin of spin torque transfer magnetic random access memory technology. IEEE Trans. Magn. 46, 2322–2327 (2010).
Kawahara, T., Ito, K., Takemura, R. & Ohno, H. Spintransfer torque RAM technology: Review and prospect. Microelectron. Reliab. 52, 613–627 (2012).
Chen, E. et al. Advances and future prospects of spintransfer torque random access memory. IEEE Trans. Magn. 46, 1873–1878 (2010).
Mellnik, A. R. et al. Spintransfer torque generated by a topological insulator. Nature 511, 449–451 (2014).
Cubukcu, M. et al. Spinorbit torque magnetization switching of a threeterminal perpendicular magnetic tunnel junction. Appl. Phys. Lett. 104, 42406 (2014).
Miron, I. M. et al. Perpendicular switching of a single ferromagnetic layer induced by inplane current injection. Nature 476, 189–193 (2011).
Mihai Miron, I. et al. Currentdriven spin torque induced by the Rashba effect in a ferromagnetic metal layer. Nat. Mater. 9, 230–234 (2010).
Liu, L., Lee, O. J., Gudmundsen, T. J., Ralph, D. C. & Buhrman, R. A. Currentinduced switching of perpendicularly magnetized magnetic layers using spin torque from the spin Hall effect. Phys. Rev. Lett. 109, 96602 (2012).
Liu, L. et al. Spintorque switching with the giant spin hall effect of tantalum. Science 336, 555–558 (2012).
Manipatruni, S., Nikonov, D. E. & Young, I. A. Beyond CMOS computing with spin and polarization. Nat. Phys. 14, 338–343 (2018).
Ramaswamy, R., Lee, J. M., Cai, K. & Yang, H. Recent advances in spinorbit torques: moving towards device applications. Appl. Phys. Rev. 5, 31107 (2018).
Patnaik, S., Rangarajan, N., Knechtel, J., Sinanoglu, O. & Rakheja, S. Spinorbit torque devices for hardware security: from deterministic to probabilistic regime. IEEE Trans. Comput. Des. Integr. Circuits Syst. 39, 1591–1606 (2020).
Kim, Y., Fong, X., Kwon, K.W., Chen, M.C. & Roy, K. Multilevel spinorbit torque MRAMs. IEEE Trans. Electron Devices 62, 561–568 (2015).
Demidov, V. E., Urazhdin, S., Anane, A., Cros, V. & Demokritov, S. O. Spin–orbittorque magnonics. J. Appl. Phys. 127, 170901 (2020).
Fan, Y. et al. Magnetization switching through giant spinorbit torque in a magnetically doped topological insulator heterostructure. Nat. Mater. 13, 699–704 (2014).
Kurebayashi, H. et al. An antidamping spin–orbit torque originating from the Berry curvature. Nat. Nanotechnol. 9, 211–217 (2014).
Lin, P.H. et al. Manipulating exchange bias by spin–orbit torque. Nat. Mater. 18, 335–341 (2019).
Safeer, C. K. et al. Spin–orbit torque magnetization switching controlled by geometry. Nat. Nanotechnol. 11, 143–146 (2016).
Fukami, S., Zhang, C., Duttagupta, S., Kurenkov, A. & Ohno, H. Magnetization switching by spinorbit torque in an antiferromagnetferromagnet bilayer system. Nat. Mater. 15, 535–541 (2016).
Zhang, C., Fukami, S., Sato, H., Matsukura, F. & Ohno, H. Spinorbit torque induced magnetization switching in nanoscale Ta/CoFeB/MgO. Appl. Phys. Lett. 107, 1–5 (2015).
Qiu, X., Narayanapillai, K., Wu, Y. & Deorani, P. Spinorbit torque engineering via oxygen manipulation. Nat. Nanotechnol. 10, 333–338 (2015).
Safranski, C., Montoya, E. A. & Krivorotov, I. N. Spin–orbit torque driven by a planar Hall current. Nat. Nanotechnol. 14, 27–30 (2019).
Lee, J. W. et al. Enhanced spinorbit torque by engineering Pt resistivity in Pt/Co/AlO_{x} structures. Phys. Rev. B 96, 64405 (2017).
Khang, N. H. D., Ueda, Y. & Hai, P. N. A conductive topological insulator with large spin Hall effect for ultralow power spin–orbit torque switching. Nat. Mater. 17, 808–813 (2018).
Wang, Y. et al. Room temperature magnetization switching in topological insulatorferromagnet heterostructures by spinorbit torques. Nat. Commun. 8, 1364 (2017).
Wu, H. et al. Roomtemperature spinorbit torque from topological surface states. Phys. Rev. Lett. 123, 207205 (2019).
DC, M. et al. Roomtemperature high spin–orbit torque due to quantum confinement in sputtered Bi_{x}Se_{(1–x)} films. Nat. Mater. 17, 800–807 (2018).
Ryu, J., Lee, S., Lee, K.J. & Park, B.G. Currentinduced spin–orbit torques for spintronic applications. Adv. Mater. 32, 1907148 (2020).
Gao, T. et al. Semimetallic bulk generated spinorbit torques in disordered topological insulator. Preprint at arXiv https://doi.org/10.48550/arXiv.1911.00413 (2013).
Chen, T.Y. et al. Efficient spin–orbit torque switching with nonepitaxial chalcogenide heterostructures. ACS Appl. Mater. Interfaces 12, 7788–7794 (2020).
Chi, Z. et al. The spin Hall effect of BiSb alloys driven by thermally excited Diraclike electrons. Sci. Adv. 6, eaay2324 (2020).
Wang, H. et al. Above roomtemperature ferromagnetism in waferscale twodimensional van der Waals Fe_{3}GeTe_{2} tailored by a topological insulator. ACS Nano 14, 10045–10053 (2020).
Hohenberg, P. C. Existence of longrange order in one and two dimensions. Phys. Rev. 158, 383–386 (1967).
Mermin, N. D. & Wagner, H. Absence of ferromagnetism or antiferromagnetism in one or twodimensional isotropic Heisenberg models. Phys. Rev. Lett. 17, 1133–1136 (1966).
Liu, S. et al. Waferscale twodimensional ferromagnetic Fe_{3}GeTe_{2} thin films grown by molecular beam epitaxy. npj 2D Mater. Appl. 1, 30 (2017).
Huang, B. et al. Layerdependent ferromagnetism in a van der Waals crystal down to the monolayer limit. Nature 546, 270–273 (2017).
Gong, C. et al. Discovery of intrinsic ferromagnetism in twodimensional van der Waals crystals. Nature 546, 265–269 (2017).
Tan, C. et al. Hard magnetic properties in nanoflake van der Waals Fe_{3}GeTe_{2}. Nat. Commun. 9, 1554 (2018).
Zhang, Y. et al. Emergence of Kondo lattice behavior in a van der Waals itinerant ferromagnet, Fe_{3}GeTe_{2}. Sci. Adv. 4, eaao6791 (2018).
Fei, Z. et al. Twodimensional itinerant ferromagnetism in atomically thin Fe_{3}GeTe_{2}. Nat. Mater. 17, 778–782 (2018).
Park, S. Y. et al. Controlling the magnetic anisotropy of the van der Waals ferromagnet Fe_{3}GeTe_{2} through hole doping. Nano Lett. 20, 95–100 (2020).
Xu, J., Phelan, W. A. & Chien, C.L. Large anomalous Nernst Effect in a van der Waals ferromagnet Fe_{3}GeTe_{2}. Nano Lett. 19, 8250–8254 (2019).
Deng, Y. et al. Gatetunable roomtemperature ferromagnetism in twodimensional Fe_{3}GeTe_{2}. Nature 563, 94–99 (2018).
Alghamdi, M. et al. Highly efficient spinorbit torque and switching of layered ferromagnet Fe_{3}GeTe_{2}. Nano Lett. 19, 4400–4405 (2019).
Wang, X. et al. Currentdriven magnetization switching in a van der Waals ferromagnet Fe_{3}GeTe_{2}. Sci. Adv. 5, eaaw8904 (2019).
Ni, Y., Zhang, Z., Nlebedim, I. C. & Jiles, D. C. Ultrahigh sensitivity of anomalous hall effect sensor based on Crdoped Bi_{2}Te_{3} topological insulator thin films. IEEE Trans. Magn. 52, 1–4 (2016).
Yu, G. et al. Switching of perpendicular magnetization by spin–orbit torques in the absence of external magnetic fields. Nat. Nanotechnol. 9, 548–554 (2014).
Wang, H. et al. Large spinorbit torque observed in epitaxial SrIrO_{3} thin films. Appl. Phys. Lett. 114, 232406 (2019).
Avci, C. O. et al. Interplay of spinorbit torque and thermoelectric effects in ferromagnet/normalmetal bilayers. Phys. Rev. B 90, 224427 (2014).
Shao, Q. et al. Strong RashbaEdelstein effectinduced spin–orbit torques in monolayer transition metal dichalcogenide/ferromagnet bilayers. Nano Lett. 16, 7514–7520 (2016).
Hayashi, M., Kim, J., Yamanouchi, M. & Ohno, H. Quantitative characterization of the spinorbit torque using harmonic Hall voltage measurements. Phys. Rev. B 89, 144425 (2014).
Shao, Y. et al. The current modulation of anomalous Hall effect in van der Waals Fe_{3}GeTe_{2}/WTe_{2} heterostructures. Appl. Phys. Lett. 116, 92401 (2020).
Fan, Y. & Wang, K. L. Spintronics based on topological insulators. SPIN 06, 1640001 (2016).
Qi, X.L. & Zhang, S.C. Topological insulators and superconductors. Rev. Mod. Phys. 83, 1057–1110 (2011).
Liu, X. et al. Temperature dependence of spinorbit torquedriven magnetization switching in in situ grown Bi_{2}Te_{3}/MnTe heterostructures. Appl. Phys. Lett. 118, 112406 (2021).
Wang, Y. et al. Topological surface states originated spinorbit torques in Bi_{2}Se_{3}. Phys. Rev. Lett. 114, 257202 (2015).
Leontiadou, M. A. et al. Experimental determination of the Rashba coefficient in InSb/InAlSb quantum wells at zero magnetic field and elevated temperatures. J. Phys. Condens. Matter 23, 35801 (2011).
Eldridge, P. S. et al. Alloptical measurement of Rashba coefficient in quantum wells. Phys. Rev. B 77, 125344 (2008).
Ye, X.G. et al. Orbittransfer torque driven fieldfree switching of perpendicular magnetization. Chin. Phys. Lett. 39, 37303 (2022).
Ramaswamy, R., Lee, J. M., Cai, K. & Yang, H. Recent advances in spinorbit torques: moving towards device applications. Appl. Phys. Rev. 5, 1–50 (2018).
Xue, H. et al. Temperature dependence of spin–orbit torques in nearly compensated Tb_{21}Co_{79} films by a topological insulator Sb_{2}Te_{3}. J. Phys. Chem. Lett. 12, 2394–2399 (2021).
Kao, I. et al. Deterministic switching of a perpendicularly polarized magnet using unconventional spin–orbit torques in WTe_{2}. Nat. Mater. 21, 1029–1034 (2022).
Shin, I. et al. Spinorbit torque switching in an allVan der Waals heterostructure. Adv. Mater. 34, 2101730 (2022).
Acknowledgements
This work was supported by the National Key R&D Program of China (2022YFB4400200 and 2018YFB0407602), the National Natural Science Foundation of China (62274009 and 61774013), the International Collaboration Project (B16001), and the National Key Technology Program of China (Grant No. 2017ZX01032101).
Author information
Authors and Affiliations
Contributions
T.N. and W.Z. conceived the ideas. T.N. designed the experiments. H. Wang and Y.L. contributed to the MBE growth. C.P. and D.C. fabricated devices. H. Wang and J.Y. performed electrical measurements. T.N., H. Wang, H. Wu, J.Z., D.W., N.L., S.S., H.L., P.L., A.F., and K.L.W. discussed and analyzed the data. T.N. and H. Wang wrote the paper with help from all the authors.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks anonymous reviewer(s) for their contribution to the peer review of this work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Wang, H., Wu, H., Zhang, J. et al. Room temperature energyefficient spinorbit torque switching in twodimensional van der Waals Fe_{3}GeTe_{2} induced by topological insulators. Nat Commun 14, 5173 (2023). https://doi.org/10.1038/s4146702340714y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s4146702340714y
This article is cited by

2D Magnetic heterostructures: spintronics and quantum future
npj Spintronics (2024)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.