WorksPhysics and Astronomy
Nonlinear spin and orbital Edelstein effect in WTe2
Abstract
In materials with spin-momentum locked spin textures, such as Rashba states and topological surface states, the current-induced shift of the Fermi contour in the k space leads to spin polarization, known as the Edelstein effect, which depends linearly on the applied current. However, its nonlinear counterpart has not yet been discovered. Here, we report the observation of the nonlinear Edelstein effect in few-layer WTe2. Under a current bias, an out-of-plane magnetization is induced in WTe2, which is electrically probed using an Fe3GeTe2 electrode, a van der Waals ferromagnet with perpendicular magnetic anisotropy. Notably, with an applied ac at frequency ω, an induced magnetization with second-harmonic response at frequency 2ω is observed, and its magnitude demonstrates a quadratic dependence on the applied current, characteristic of the nonlinear Edelstein effect. This phenomenon is well explained by the current-induced orbital magnetization via the Berry connection polarizability tensors in WTe2. The orbital degree of freedom plays the primary role in the observed nonlinear Edelstein effect, that is, the nonlinear orbital Edelstein effect. This can, in turn, give rise to a nonlinear spin Edelstein effect through spin-orbit coupling.
202 Nonlinear spin and orbital Edelstein effect in WTe
202 Nonlinear spin and orbital Edelstein effect in WTe2.pdf
The efficient conversion of charge to spin is crucial for developing next-generation spintronic devices [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. This conversion is particularly notable in materials exhibiting spin-momentum locked spin textures, such as those with Rashba spin-orbit coupling (SOC) or topological surface states. In Rashba systems, the breaking of inversion symmetry leads to the splitting of spin bands into inner and outer Fermi circles with opposite spin helicities [5]. Applying a current induces an imbalance in the density of spin-up and spin-down electrons, resulting in net spin polarization. Similarly, topological insulators and topological semimetals, with their unique topological surface states, also exhibit efficient charge-to-spin conversion due to their intrinsic spin-momentum locking [6, 7, 8]. This phenomenon, generally known as the Edelstein effect, typically exhibits a linear relationship between the induced spin polarization and the applied current. The linear Edelstein effect has been used to generate spin torques to drive the magnetization switching of ferromagnetic materials, significantly impacting the development of magnetic random-access memory [11].
Despite extensive research on the Edelstein effect, the exploration of its nonlinear counterpart has remained elusive until now. A nonlinear Edelstein effect, where the induced spin polarization depends quadratically on the applied current, would have profound implications. In this study, we present an experimental observation of the nonlinear Edelstein effect in few-layer WTe. When subjected to an ac bias, WTe exhibits an out-of-plane magnetization that displays a second-harmonic response at twice the frequency of the driving current, with a magnitude that scales quadratically with the current—a hallmark of the nonlinear Edelstein effect. This phenomenon is elucidated through the current-induced orbital magnetization due to the Berry connection polarizability (BCP) tensors in WTe.
Orbital magnetization and nonlinear Edelstein effect induced by BCP tensors. Recently, orbital magnetization [13, 14, 15, 16, 17], orbital torques [18, 19, 20], orbital Hall effect [21, 22, 23, 24, 25], and the kinetic magnetoelectric effect [26, 27, 28, 29, 30, 31, 32, 33], which are related to the orbital magnetic moment, have garnered significant research interest. Within the framework of semiclassical Boltzmann transport equation along with constant relaxation time approximation, the current-induced orbital magnetization is given by , where and label the spatial directions, is the electric field, is the scattering time, and is the kinetic magnetoelectric (ME) susceptibility. The can be notable for an asymmetric orbital texture, where is approximately valid at low temperatures ( is the Fermi temperature), is the th component of orbital magnetic moment , is the electron velocity, and the integral is over the Brillouin zone with summation over the band index [34]. Moreover, the nonlinear magnetoelectric coupling can also lead to current-induced magnetization [35].
In WTe with the Td phase, the orbital magnetization can be generated by applying current. Td-phase WTe is classified as a type-II Weyl semimetal, showcasing Fermi arc surface states [36]. Remarkably, at the monolayer level, WTe exhibits the quantum spin Hall effect [37] and superconductivity [38, 39]. The crystal structure of WTe belongs to the space group , with mirror symmetry and glide mirror symmetry [40]. These two mirror symmetries with perpendicular mirror planes together generate a twofold rotational symmetry in the plane through a half-cell translation. By applying a current to break this symmetry, the nonlinear Edelstein effect can be induced through the mechanism of BCP tensors, defined as , where is the interband Berry connection and is the cell-periodic part of the Bloch eigenstate with eigenenergy [41, 42, 43, 44]. The BCP tensors are gauge-invariant band geometric quantities associated with the electric-field induced positional shift of Bloch electrons [43]. The BCP is regarded as the intrinsic origin of the third-order nonlinear Hall effect [44]. Upon applying an electric field, the BCP tensor can lead to Berry connection linearly proportional to the applied current, denoted as . Here, is the BCP tensor, is the applied electric field, and the superscript “(1)” indicates the first-order term of the electric field. Note that the orbital magnetic moment involves the matrix elements , with as the position operator, as the velocity operator, as the Hamiltonian, and as the eigenstate [45]. In the Bloch representation, the position operator is formulated as [46]. Therefore, the BCP tensors can give rise to a field-correction term of through the field-induced Berry connection , that is, , which is linearly proportional to .
Figures 1(a) and 1(b) illustrate the electric-field induced orbital magnetic moments in the Td monolayer WTe by applying an electric field along the axis and axis, respectively, exhibiting antisymmetric distributions in momentum space, where for , and for . This dipolelike distribution forms an exotic orbital texture with orbit-momentum locking [23, 32], giving rise to the kinetic magnetoelectric coefficient , where an in-plane electric field is able to generate out-of-plane orbital magnetization (). Figure 1(c) illustrates the as applying an electric field along the axis in WTe (see Supplementary Material, Figs. S1 and S2 [47] for model calculations), giving . Utilizing this orbital texture, the electric field additionally triggers orbital magnetization, denoted as [55], constituting a nonlinear Edelstein effect. While the nonlinear orbital Edelstein effect analysis provided relies on the Boltzmann equation method, a more comprehensive theoretical framework, starting from nonlinear transport theory [56], remains desirable. Furthermore, it is noteworthy that with the presence of SOC, the orbital magnetization gives rise to spin polarization, leading to the emergence of nonlinear spin Edelstein effect. However, it is important to note that the orbital degree of freedom is more fundamental here, as the nonlinear orbital dynamics are independent of the strength of SOC, whereas the nonlinear spin Edelstein effect only appears when SOC is activated [55].
Figure 1. (a), (b) Illustration of asymmetric distribution of orbital magnetic moment in the conductance band when applying an electric field along (a) axis or (b) axis in Td monolayer WTe, respectively. (c) The coefficient as a function of the in-plane electric field at chemical potential 0.1 eV. (d) Illustration of nonlinear Edelstein effect in WTe. The applied current induces an orbital texture, further leading to orbital magnetization, which exhibits a quadratic dependence on the current.
The nonlinear Edelstein effect is characterized by (, , , , in WTe, with the axis as the out-of-plane orientation) [48]. Here, we only consider the tensor component , specifically for , , , representing the in-plane field-induced out-of-plane magnetization. The is intricately linked to the BCP tensors, allowed by inversion symmetry (refer to the symmetry constraints detailed in Supplementary Material, Table S1 [47]). Since WTe bulk holds both and mirror symmetries [44], the two mirrors enforce , and only the components and are permissible. Therefore, applying an electric field or along the crystalline axis cannot induce out-of-plane magnetization [44, 48]. However, deviating the from the crystalline - or axis allows the generation of out-of-plane magnetization due to the nonzero and .
| ✔ | ✔ | 🚫 | 🚫 | ✔ | 🚫 | |
| ✔ | ✔ | 🚫 | ✔ | ✔ | 🚫 | |
| ✔ | 🚫 | ✔ | ✔ | ✔ | ✔ |
Table S1. TABLE S1. The symmetry constraints of the nonlinear magnetoelectric coefficient under time-reversal symmetry. Here is the inversion symmetry, is the -fold rotation symmetry with rotation axis along direction. is the mirror symmetry with the mirror line along direction. Allowed: ✔; Forbidden: 🚫.
In experiments, the out-of-plane magnetization in WTe bulk is generated by applying a charge current [Fig. 1(d)], effectively generating an electric field denoted as , where represents the longitudinal resistance ( in our device), and is the channel length ( m in our device). The current-induced magnetization is expressed as , which remains unchanged as reversing the direction of current.
Current-induced out-of-plane magnetization in WTe
To directly measure the current-induced out-of-plane magnetization in WTe, we employ a few-layer FeGeTe (FGT) as a ferromagnetic (FM) probe. Results from two devices, namely devices S1 and S2, are presented, with the focus on device S1 in the main text. The optical image of device S1 is depicted in Figure 2(a). A layer of h-BN thin film was interposed between WTe and FGT to prevent magnetic proximity effect from FGT.
Figure 2. (a) Optical image of a device. (b) Schematic of circular disc electrodes. The angle is defined as the angle between the axis of WTe and the direction of applied current. (c) The Hall resistance of FGT, showing hysteresis with magnetic field. (d) Schematic of magnetization detection in WTe via FGT probe. (e) Magnetization-related second-harmonic voltage at and kV/m. (f) Using two Au electrodes for the voltage measurement at and kV/m, no hysteresis loop is observed. All data were acquired at 5 K.
FGT. Disc electrodes were fabricated to apply current along various directions, as illustrated in Figure 2(b), where the angle represents the relative angle between the applied current and a pair of electrodes (along the crystalline axis with a small misalignment of ). The perpendicular magnetism anisotropy of FGT is confirmed by a square-shaped hysteresis loop in the anomalous Hall resistance, as shown in Figure 2(c) (also refer to Figure S5 in the Supplementary Material [47]).
Figure S5. The hysteresis curve of the Hall resistance of FGT as sweeping in-plane (black) and out-of-plane (red) magnetic field.
The experimental setup for measuring the current-induced magnetization in WTe is illustrated in Figure 2(d). In these measurements, a fixed ac with a frequency is applied to generate magnetization through the nonlinear Edelstein effect. Given that , a second-harmonic out-of-plane magnetization with a frequency of is anticipated. The magnetization in WTe is then probed using the FGT probe. When is antiparallel or parallel to the magnetization of FGT (), a high- or low-voltage state will be detected, respectively (see Supplementary Note 3 for detailed mechanisms [47]). By varying through sweeping an external magnetic field and simultaneously measuring the voltage using the FGT probe with respect to a reference Au electrode via the lock-in method, the current-induced in WTe can be directly obtained (see Figure S6 [47] for the detailed measurement configuration). The magnetization-dependent voltage is determined by subtracting the background from , denoted as (refer to Figure S7 [47]).
Figure S6. (a) Measurement scheme for the nonlinear orbital Edelstein effect. (b) The illustration of measurement configuration corresponding to (a). The red numbers denote the electrodes depicted in (a).
Figure S7. The measured second-harmonic voltage () via the FGT probe under varying excitation current at fixed direction ().
The measured with a current applied at an angle deviated from the axis () is presented in Figure 2(e). Hysteresis behavior is evident in during both forward and backward sweeping magnetic fields. The high- and low states of correspond to the current-induced in WTe being antiparallel and parallel to , respectively. A difference in between up- and down sweeping magnetic fields above T is noted, attributed to the cumulative nature of thermal effects over time under the applied ac field kV/m (-mA current). No hysteresis loop is observed for the voltage probe employing two nonmagnetic electrodes with the FGT electrode between them [Fig. 2(f)]. This observation unequivocally dismisses the possibility of the originating from the proximity-induced magnetism in WTe by the FGT electrode.
The at is further investigated by varying the intensity of excitation current, as shown in Figure 3(a). It is found that the loop height increases with excitation current, and its polarity remains unchanged upon reversing the excitation current (by exchanging the source-drain leads). The loop height , defined as the voltage difference at zero magnetic field when forward- and backward sweeping magnetic field, exhibits a quadratic dependence on the excitation current [Fig. 3(b)]. Because , the observations suggest a quadratic relationship between the induced out-of-plane magnetization in WTe and the excitation current, aligning with the anticipated behavior of the nonlinear Edelstein effect.
Figure 3. (a) under various at . (b) The height of hysteresis curve, , as a function of at . The red line is a fit with the formula of . (c) at kV/m with (approximately along the axis). All data were acquired at 5 K.
The is further investigated when applying current at (approximately along the axis). As depicted in Figure 3(c), suppressed hysteresis is observed as sweeping magnetic field. Due to the mirror plane imposing in WTe, inducing out-of-plane magnetization with along the axis is not feasible. The slight, nonvanishing hysteresis in Figure 3(c) is attributed to a misalignment of between and the axis. Consistent results were obtained with excitation currents along different directions (refer to Figure S8 [47]) and under various measurement configurations (see Figure S9 [47]). Furthermore, reproducible results were confirmed in device S2 (refer to Figure S10 and Figure S11 in Supplemental Material [47]).

(a)

(b)

(c)

(d)

(e)

(f)
Figure S8. Current induced second-harmonic out-of-plane magnetization of at 5 K.
Figure S9. Second-harmonic magnetization-dependent voltage under different reference Au electrodes.
a-c, The voltage measurement configurations with the fixed FGT probe and different reference Au electrodes.
d-f, The measured under corresponding measurement configurations in a-c, respectively.
Figure S10. The measured second-harmonic voltage () via the FGT probe under at (a) and (b) in device S2 at 5 K.
Figure S11. Anisotropic transport in device S2. a-d, (a) Longitudinal resistance , (b) third-order nonlinear Hall effect , (c) second-harmonic magnetization signal , and (d) first-harmonic magnetization signal as a function of angle at 5 K.
The temperature dependence of current-induced magnetization is further investigated. As depicted in Figure 4, the diminishes with increasing temperature, and the hysteresis loop becomes barely discernible above 50 K. Concurrently, the third-order nonlinear Hall effect of WTe was measured at various temperatures. Remarkably, when plotting against , a clear linear dependence emerges, as illustrated in Figure 4(c). Here, , , is the third-harmonic Hall voltage, and and are the channel width and length, respectively. Note that the quantity is directly related to the BCP tensors in WTe [44]. The BCP tensors contribute to the by , where is the BCP tensors and is the (intrinsically) -independent part of [42]. Therefore, the observed linear correlation between and indicates that the current-induced nonlinear magnetization in WTe also possesses an intrinsic origin from the BCP tensors.
Figure 4. (a) Second-harmonic under kV/m along at different temperatures as indicated. (b) Temperature dependence of the hysteresis height . (c) The at fixed kV/m as a function of the third-order nonlinear Hall signal along . The data were collected across a temperature range of 8 to 50 K.
Discussion and conclusions
The magnetization-dependent second-harmonic voltage signals distinctly indicate the nonlinear out-of-plane magnetization induced by in-plane current, that is, the nonlinear Edelstein effect. We also examine other mechanisms contributing to this nonlinear magnetization (details in Supplemental Note 7), and it has been determined that the nonlinear spin- (orbital) Hall effect and spin Seebeck effect are not responsible for the observed phenomena.
Recent reports indicate that in-plane and out-of-plane magnetization with and 30 V, respectively, can be induced by the linear Edelstein effect with an applied current of 1 mA along the axis in WTe, where the outof-plane magnetization is allowed by the surface symmetry breaking [51, 52, 53, 54, 57]. Despite its relatively weaker magnitude observed in this work, the current-induced magnetization from the nonlinear Edelstein effect demonstrates a quadratic relationship with the current, ensuring its direction remains unchanged upon the reversal of the current. Furthermore, the lower-symmetry requirements of the nonlinear Edelstein effect render it more applicable across a broader range of crystalline materials. These characteristics endow the nonlinear Edelstein effect with unique potential applications in charge-to-magnetization conversion.
In summary, our results reveal the emergence of nonlinear Edelstein effect in few-layer . The application of a current induces an exotic orbital texture through the positional shift of Bloch electrons, further leading to the generation of magnetization. The induced magnetization
demonstrates a quadratic dependence on the applied current, directly observed using a ferromagnetic FGT probe with second-harmonic voltage measurements. This finding represents a significant advancement in understanding the nonlinear charge-to-spin conversion mechanisms, particularly highlighting the role of current-induced orbital magnetization facilitated by BCP tensors. Considering the widespread existence of BCP in diverse materials, regardless of SOC [34, 42, 43, 44], the nonlinear Edelstein effect holds significant promise in a broad range of materials with weak SOC for the electric control of magnetization.
Acknowledgments. This work was supported by the National Natural Science Foundation of China (Grants No. 62425401 and No. 62321004), and Innovation Program for Quantum Science and Technology (Grant No. 2021ZD0302403).
SM
Method
Device fabrications. The high-quality bulk WTe crystals were obtained from HQ Graphene. Few-layer WTe flakes were obtained by a mechanical exfoliation method, where the WTe crystal was repeatedly exfoliated with scotch tape, transferred onto polydimethylsiloxane (PDMS), and then covered onto a precleaned Si substrate with 285 nm-thick SiO. The PDMS was heated for about 1 minute at 90°C to transfer the few-layer WTe onto the Si substrate. Ti/Au electrodes (around 10 nm thick) were patterned onto the SiO/Si substrate using e-beam lithography, metal deposition, and lift-off. The exfoliated h-BN, (FGT), and few-layer WTe were sequentially picked up and transferred onto the Ti/Au electrodes using a polymer-based dry transfer technique.
In the fabrication of the device, ~7-nm-thick WTe flake was chosen and was stacked with monolayer or bilayer h-BN and 10-nm-thick FGT strip, where the h-BN served as the spacer to prevent the magnetic proximity effect between the FGT and WTe. A 20-30 nm thick h-BN layer was placed on top for coverage. The entire exfoliation and transfer process were performed in an argon-filled glove box with an O and HO content below 0.01 parts per million to prevent sample degradation.
Berry connection polarizability in Td monolayer WTe
The Berry connection polarizability (BCP) tensor is a band geometric quantity that is related to the field-induced positional shift of Bloch electrons, given by , where is the interband Berry connection and is the electron charge [42]. The BCP tensor can generate Berry connection through an external electric field following , where is the BCP tensor, is the external electric field, and the superscript “(1)” represents that the physical quantity is the first order term of electric field.
The BCP tensors are calculated in monolayer Td WTe using the effective Hamiltonian
where , , , and (). All the parameters used are the same as in Ref. [32]. The calculated BCP tensors are shown in Figure S1. Giant BCP tensors are found near band edge. Importantly, even when the spin-orbit coupling is absent, the BCP tensors are still significant in WTe.
Figure S1. (a) Energy band of the monolayer WTe, where the spin-orbit coupling is turned off, leading to spin-degenerate bands. (b)–(d) Distribution of the BCP component (b) , (c) , and (d) in momentum space.
It is worth noting that the monolayer model is a simplified approach to understanding the electronic properties of the few-layer WTe. Nevertheless, this model captures the main features of the BCP tensors when compared with the results from first-principles calculations of the bulk system (ref. 44). These main features are crucial for understanding the nonlinear magnetization signals observed in our experiments.
However, the monolayer model cannot capture the detailed stacking interactions and variations in electronic properties that arise in the few-layer and bulk systems. First-principles calculations of the bulk system, as shown in ref. 44, reveal more detailed features that the monolayer model cannot account for.
The field-induced orbital magnetic moment in monolayer Td WTe is calculated by a Kubo-like formula,
where the -th band BCP tensors are calculated as
as shown in Fig. 1 of main text.
To demonstrate this field-induced orbital magnetic moment, calculations are also implemented in a two-dimensional (2D) gapped Dirac model with the Hamiltonian
where is the 2 × 2 identity matrix, is Pauli matrices, describes the tilting of Dirac cone, and describe the anisotropy of Fermi velocity, is the energy gap [42]. The energy band of the 2D gapped Dirac model is shown in Figure S2. The BCP tensors are evaluated, as shown in Figure S2-(Figure S2). It is found that the BCP tensors show maximum value near the band edge. Under the application of an external electric field, orbital magnetic moment can be generated by the BCP tensors. The field-induced orbital magnetic moment shows a dipole-like distribution in momentum space, leading to nonzero kinetic magnetoelectric coefficient , as shown in Figure S2-(Figure S2).

Figure S2. The 2D gapped Dirac model. a, Energy band of the gapped Dirac model. b-d, Distribution of the BCP component (b) , (c) , and (d) in momentum space. e,f, Field induced orbital magnetic moment under an external electric field applying along the (e) axis and (f) axis.
Nonlinear orbital Edelstein effect
The orbital magnetic moment arises from the self-rotation of Bloch electrons around their center of mass [22, 23]. An unusual orbital texture characterized by orbit magnetic moment-momentum locking can result in a nonzero magnetoelectric coefficient , and the out-of-plane orbital magnetization is given by [34], known as the out-of-plane orbital Edelstein effect. However, in materials with inversion symmetry, the is forced to be zero, leading to this out-of-plane orbital Edelstein effect vanishing [27]. Alternatively, the can be induced by applying an electric field through the mechanism of BCP tensors, where the restriction of inversion symmetry is removed.
As depicted in the upper panel of Figure S3, the applied current generates an antisymmetric orbital texture with a nonzero due to the field-correction of Berry connection induced by BCP tensors. The applied current further induces orbital magnetization, as illustrated in the lower panel of Figure S3, through the Fermi contour shift. Therefore, the current induces orbital magnetization with , representing the nonlinear orbital Edelstein effect.
Figure S3. Illustration of BCP-induced nonlinear magnetoelectric effect. Upper panel: The current-generated orbital texture in with . Lower panel: Illustration of the current-induced orbital magnetization with nonequilibrium electron distribution.
The out-of-plane nonlinear orbital Edelstein effect relies on the mechanism of BCP tensors. Given the widespread presence of BCP tensors across various materials, the nonlinear orbital Edelstein effect is anticipated to manifest in a diverse range of substances. Unlike the linear orbital Edelstein effect exclusive to gyrotropic crystals [27], the nonlinear counterpart exhibits a more lenient symmetry constraint [48]. To provide a generalized framework for the nonlinear orbital Edelstein effect in diverse materials, we summarize the symmetry constraints. Specifically, we consider the time-reversal-even nonlinear magnetoelectric coefficient in nonmagnetic materials, focusing on in-plane electric field-induced out-of-plane magnetization. Assuming the is applied in the plane () and the induced magnetization is along the out-of-plane axis (), the obtained symmetry constraints are presented in Table S1.
Notably, Table S1 reveals that the nonlinear magnetoelectric coefficient is permissible in bulk, precisely aligning with the results obtained in this work.
In this work, we choose few-layer to demonstrate the nonlinear orbital Edelstein effect based on the following considerations:
exhibits metallic behavior, which facilitates an efficient Fermi contour shift driven by electric current. This shift is crucial for the nonlinear orbital Edelstein effect, representing a kinetic magnetoelectric effect.
The bulk of has significant BCP tensors, leading to the generation of a Berry curvature dipole under an electric field. This effect has already been considered as the origination of the third-order nonlinear Hall effect observed in bulk. Consequently, the nonlinear orbital Edelstein effect can be highlighted as the dominant term.
Magnetization detection using a ferromagnetic probe
The current induced out-of-plane magnetization in is directly measured through the detection of a magnetization-dependent voltage using a ferromagnetic (FM) probe [6, 7, 8, 9, 10]. The depends on the relative orientation between the and the magnetization of the FM probe (). When , electrons with magnetic moment of will diffuse from to the FM probe, causing the FM probe to become negatively charged and yielding . Conversely, when , electrons with magnetic moment of will diffuse from the FM probe to , causing the FM probe to become positively charged and yielding . These processes are illustrated in Figure S4-(Figure S4). One can determine that when , , and when , . Thus, by adjusting the external magnetic field to vary , a high voltage when and a low voltage when can be measured by the FM probe.
Figure S4. Illustration of the principle of magnetization detection using a ferromagnetic probe.
a,b, Left part: Application of an electric field induces polarization of the magnetic moment () in WTe. Right part: The magnetization of the FeGeTe (FGT) probe () that can be altered by an external magnetic field. The electrochemical potential balance between states with the same in WTe and FGT results in charge transfer, and thus a voltage () can be measured by the FGT probe.
c, The conversion between orbital (red arrows) and spin (blue arrows) magnetic moments in WTe/ferromagnet bilayers. There exist two mechanisms for spin-orbital conversion. For the first channel, the orbital magnetic moment is converted into spin through the spin-orbit coupling (SOC) in WTe and then the polarized spin is injected into the FM probe. For the second channel, the orbital angular momentum is injected into FM probe and then is converted into spin through the SOC of the FM probe. For both ways, the orbital magnetization in WTe can induce spin chemical potential imbalance in the FM probe.
This approach can be used to detect not only the spin magnetization, but also the orbital magnetization in . As illustrated in Figure S4, the orbital magnetic moment can be converted into spins through the spin-orbit coupling [49]. Therefore, the orbital polarization in can induce spin chemical potential imbalance in the FM probe, leading to a measurable related to the orbital magnetization in .
The choice of ferromagnetic probe electrode:
In this work, we opt for a few-layer (FGT) flake, a 2D layered van der Waals ferromagnet, as the magnetic probe for measuring the current-induced magnetization in . The orbital magnetization in aligns out-of-plane due to the confinement of charge carriers within a 2D plane. Given the perpendicular magnetic anisotropy of FGT, it is particularly suited for detecting the current induced out-of-plane magnetization. To provide further clarity, Figure S5 shows the hysteresis curves of the Hall resistance of the FGT as sweeping the in-plane and out-of-plane magnetic fields, respectively. As alterations in the magnetization state of FGT can impact the Hall resistance due to the anomalous Hall effect, the hysteresis curves in Figure S5 indeed reflects the magnetization variation. The distinct transitions observed in the hysteresis curves only when a magnetic field is applied along the out-of-plane direction confirm the perpendicular magnetic anisotropy of FGT in our device.
Additionally, the Curie temperature of few-layer FGT used in this work is K [50]. In our study, the temperature for probing current-induced magnetization is below 50 K, ensuring that few-layer FGT meets the requirements for this investigation. During the measurement of current-induced magnetization in , we introduced a perpendicular magnetic field to manipulate the magnetization state of FGT. This enabled us to detect the magnetization-dependent voltage, representing FGT magnetization in both parallel and antiparallel orientations with respect to the current-induced magnetization. The magnitude of the applied magnetic field exceeded the coercive force of FGT, ensuring that the magnetic domain distribution did not influence the measurement results.
In our experiments, we strategically incorporate a thin -BN layer between the FGT and WTe layers. Its role is to act as an insulating barrier, preventing the magnetic proximity effect from FGT on the neighboring WTe layer. This ensures that any observed magnetic signals originate solely from current-induced magnetization in WTe, without influence from FGT's magnetism.
Supplemental Note 4: Measurement Scheme for the nonlinear orbital Edelstein effect
The measurement scheme for the magnetization detection is shown in the Figure S6.
The disc-shaped electrodes 1-12 are used to apply current along different directions (i.e., along different angle ). Electrodes 13 and 14 are voltage contacts. Electrode 13 connects to the WTe, which is the reference non-magnetic electrode, while the electrode 14 connects to the ferromagnetic probe FeGeTe, which is regarded as the magnetic probe (see also Figure S6). Figure S6 shows the measurement configuration for current along . To rotate the current direction, we just rotate the current electrodes, for example, from electrodes 1/7 (I+/I-) to electrodes 2/8 (I+/I-), while keep the voltage contacts 13 and 14 unchanged.
Additional results of device S1
To investigate current-induced out-of-plane magnetization in , a magnetization-dependent voltage is collected via a FGT probe. An AC current is injected using the standard lock-in method, and the measured voltage is shown in Figure S7. Figure S7 shows the current-dependence of the measured second-harmonic voltage . To obtain the magnetization-related voltage signals , a linear background is removed from .
a,b, Second-harmonic magnetization-dependent voltage at under (a) kV/m and (b) kV/m, respectively.
c, The loop height of the - hysteresis curve, , as a function of at .
d,e, at under (d) kV/m and (e) kV/m, respectively.
f, as a function of at .
The second-harmonic magnetization-dependent voltage () is measured at different angles, all showing consistent results, as shown in Figure S8. With a fixed source-drain current, we measured the voltage drop between the FGT electrode and various reference Au electrodes. The results are shown in Figure S9. It is found that the shows negligible dependence on the reference voltage probe configurations.
Reproducible results in device S2
We fabricated Device S2 with a similar configuration to Device S1. As presented in Figure S10, Device S2 exhibits consistent results with Device S1 in terms of its second-harmonic voltage, .
The angle dependent measurement was implemented in device S2. The results are presented in Figure S11. The angle-dependence of resistance and third-order nonlinear Hall effect is well consistent with previous works [44]. The angle-dependence of is fitted by
where and are resistance along crystalline and axis, respectively. The third-order nonlinear Hall effect was measured via standard lock-in method. The third-order nonlinear Hall effect shows angle-dependence following the formula
where , , is the third-harmonic Hall voltage, is the applied current, and are channel width and length, respectively, is the resistance anisotropy, are elements of the third-order susceptibility tensor, is the angle misalignment between and the crystalline axis. The fitting curves yield a misalignment between the electrode base line and the crystalline axis, where the -axis is at around in device S1 and in device S2. Importantly, the as a function of angle also shows the same symmetry as the third-order nonlinear Hall effect (Figure S11), which is consistent with the BCP mechanism in WTe. In addition to the second-harmonic magnetic voltage , the first-harmonic magnetic voltage is also observed (Figure S11), consistent with previous works [54] about the linear Edelstein effect in . The angle-dependence of the first-harmonic signal shows a sine-shaped curve with maximum value along axis, and zero value along axis. This is distinct from the second-harmonic signal where zero values are observed along both and axis. The different symmetries indicate the different physical origin of the first- and second-harmonic signals.
Discussions about the origin of the current-induced nonlinear magnetization
Contact-induced nonlinearity
The nonlinear voltage can be induced by contact-induced diode effect. However, as shown in Figure S12, the linear - relationship is observed, indicating the ohmic nature of the contacts. Therefore, the scenario of contact-induced nonlinearity can be ruled out.
Figure S12. (a) The characteristic IV curves of device S2 at 5 K. (b) Third-order nonlinear Hall effect at 5 K in device S2.
Nonlinear spin/orbital Hall effect
For the spin/orbital Hall effect, under the application of a current, the spin/orbit, characterized with -component magnetic moments, deflects towards opposing directions and accumulates respectively on two distinct edges of the sample. If there is a nonlinear version of the spin/orbital Hall effect, the measured should be sensitive to the configurations of the probe electrodes. This sensitivity arises from the influence of spin accumulation at the reference electrode, impacting the measured signal magnitude. In our experiments, as demonstrated in Figure S9, the exhibits negligible dependence on the position of reference electrode. In this setup, the FGT electrode is fixed, while the reference electrodes are positioned on the left and right sides of the source-drain current. Hence, the nonlinear spin/orbital Hall effect is not the origin of the observed phenomenon in our experiment.
Spin Seebeck effect
Applying a current in , a temperature gradient may be induced. Under this temperature gradient, a spin-polarized current may emerge if there are differences in the Seebeck coefficients between the spin-up and spin-down states.
Previous reports indicate that a current along the -axis can linearly generate the spin polarization along -direction in [51, 52, 53, 54], implying the presence of a specialized spin texture with some form of spin-momentum locking. Consequently, if the temperature gradient has a component along the -axis, out-of-plane spin polarization may be induced due to the spin Seebeck effect, proportional to .
In our experiments, as illustrated in Figure S11 for the results of device S2, the exhibits the same symmetry as the third-order nonlinear Hall signals , aligning with the the BCP mechanism in . Crucially, the third-order nonlinear Hall signal is zero along the or axis due to the symmetry of [44]. Similarly, the also approaches a minimum along the or axis. This observation contradicts the expectations of the spin Seebeck effect, where the maximum is anticipated along the -axis due to a specific spin texture. Therefore, the spin Seebeck effect can be ruled out as the origin of the observed second-harmonic magnetization signal.
Nonlinear Orbital and Spin Edelstein effect
In this work, the observed second-harmonic magnetization signals are attributed to the nonlinear orbital and spin Edelstein effect through the BCP tensors. Upon applying a current, the BCP tensor can result in Berry connection linearly proportional to the applied current, that is, . Here is the BCP tensors, is the applied electric field, and the superscript “(1)” represents that the physical quantity is the first order term of electric field. Moreover, the field-induced Berry connection would contribution to a correction term of orbital magnetic moment . The field-induced orbital magnetic moment shows antisymmetric distributions in the momentum space with orbit-momentum locking, thus inducing nonzero magnetoelectric coefficient . Therefore, orbital magnetization is induced by the electric field with , that is, nonlinear orbital Edelstein effect.
In the results of device S2 illustrated in Figure S11, the symmetry of is consistent with that of the third-order nonlinear Hall signals (Figure S12). Furthermore, a linear relationship exists between and the third-order nonlinear Hall effect, as depicted in Figure 4(c). These findings suggest a common origin for both the current-induced second-harmonic magnetization and the third-order nonlinear Hall effect, namely, the BCP mechanism. Through this BCP mechanism, the current generates the polarization of orbital magnetic moments by modifying Berry connection.
It is noteworthy that, owing to the presence of spin-orbit coupling in , the orbital magnetization gives rise to a spin polarization, leading to a corresponding nonlinear spin Edelstein effect. While it can be challenging to disentangle the contributions of orbit and spin in systems with strong spin-orbit coupling, the fundamental essence of this observation is rooted in the orbital effect. The spin contribution arises as a consequence of spin-orbit coupling.
References
- [1]J.-C. Rojás-Sanchez, S. Oyarzún, Y. Fu, A. Marty, C. Vergnaud, S. Gambarelli, L. Vila, M. Jamet, Y. Ohtsubo, A. Taleb-Ibrahimi, P. Le Fèvre, F. Bertran, N. Reyren, J.-M. George, and A. Fert, Spin to charge conversion at room temperature by spin pumping into a new type of topological insulator: α−sn films, Phys. Rev. Lett. 116, 096602 (2016).DOI
- [2]B. Dieny, I. L. Prejbeanu, K. Garello, P. Gambardella, P. Freitas, R. Lehndorff, W. Raberg, U. Ebels, S. O. Demokritov, J. Akerman, A. Dec et al., Opportunities and challenges for spintronics in the microelectronics industry, Nat. Electron. 3, 446 (2020).arxiv.org/abs/1908.10584
- [3]J. Sinova, D. Culcer, Q. Niu, N. A. Sinitsyn, T. Jungwirth, and A. H. MacDonald, Universal intrinsic spin Hall effect, Phys. Rev. Lett. 92, 126603 (2004).DOI
- [4]J. Sinova, S. O. Valenzuela, J. Wunderlich, C. H. Back, and T. Jungwirth, Spin Hall effects, Rev. Mod. Phys. 87, 1213 (2015).
- [5]A. Manchon, H. C. Koo, J. Nitta, S. M. Frolov, and R. A. Duine, New perspectives for Rashba spin–orbit coupling, Nat. Mater. 14, 871 (2015).DOI
- [6]C. H. Li, O. M. J. van ’t Erve, J. T. Robinson, Y. Liu, L. Li, and B. T. Jonker, Electrical detection of charge-current-induced spin polarization due to spin-momentum locking in Bi₂Se₃, Nat. Nanotechnol. 9, 218 (2014).DOI
- [7]F. Yang, S. Ghatak, A. A. Taskin, K. Segawa, Y. Ando, M. Shiraishi, Y. Kanai, K. Matsumoto, A. Rosch, and Y. Ando, Switching of charge-current-induced spin polarization in the topological insulator BiSbTeSe₂, Phys. Rev. B 94, 075304 (2016).DOI
- [8]B.-C. Lin, S. Wang, A.-Q. Wang, Y. Li, R.-R. Li, K. Xia, D. Yu, and Z.-M. Liao, Electric control of Fermi arc spin transport in individual topological semimetal nanowires, Phys. Rev. Lett. 124, 116802 (2020).DOI
- [9]J.-i. Inoue, G. E. W. Bauer, and L. W. Molenkamp, Diffuse transport and spin accumulation in a Rashba two-dimensional electron gas, Phys. Rev. B 67, 033104 (2003).DOI
- [10]A.-Q. Wang, P.-Z. Xiang, X.-G. Ye, W.-Z. Zheng, D. Yu, and Z.-M. Liao, Room-temperature manipulation of spin texture in a Dirac semimetal, Phys. Rev. Appl. 14, 054044 (2020).DOI
- [11]A. R. Mellnik, J. S. Lee, A. Richardella, J. L. Grab, P. J. Mintun, M. H. Fischer, A. Vaezi, A. Manchon, E.-A. Kim, N. Samarth, and D. C. Ralph, Spin-transfer torque generated by a topological insulator, Nature (London) 511, 449 (2014).DOI
- [12]L. Liu, C.-F. Pai, Y. Li, H. W. Tseng, D. C. Ralph, and R. A. Buhrman, Spin-torque switching with the giant spin Hall effect of tantalum, Science 336, 555 (2012).DOI
- [13]E. Santos, J. E. Abrão, D. Go, L. K. de Assis, Y. Mokrousov, J. B. S. Mendes, and A. Azevedo, Inverse orbital torque via spin-orbital entangled states, Phys. Rev. Appl. 19, 014069 (2023).arxiv.org/abs/2204.01825
- [14]H. An, Y. Kageyama, Y. Kanno, N. Enishi, and K. Ando, Spin–torque generator engineered by natural oxidation of Cu, Nat. Commun. 7, 13069 (2016).DOI
- [15]K. Hasegawa, Y. Hibino, M. Suzuki, T. Koyama, and D. Chiba, Enhancement of spin-orbit torque by inserting CoOₓ layer into Co/Pt interface, Phys. Rev. B 98, 020405(R) (2018).DOI
- [16]K. Yuito, T. Yuya, A. Hongyu, H. Takashi, G. Tenghua, J. Shi, and A. Kazuya, Spin-orbit torque manipulated by fine-tuning of oxygen-induced orbital hybridization, Sci. Adv. 5, eaax4278 (2019).
- [17]J. Kim, D. Go, H. Tsai, D. Jo, K. Kondou, H.-W. Lee, and Y. Otani, Nontrivial torque generation by orbital angular momentum injection in ferromagnetic-metal/Cu/Al₂O₃ trilayers, Phys. Rev. B 103, L020407 (2021).DOI
- [18]D. Go, D. Jo, C. Kim, and H.-W. Lee, Intrinsic spin and orbital Hall effects from orbital texture, Phys. Rev. Lett. 121, 086602 (2018).DOI
- [19]D. Jo, D. Go, and H.-W. Lee, Gigantic intrinsic orbital Hall effects in weakly spin-orbit coupled metals, Phys. Rev. B 98, 214405 (2018).DOI
- [20]D. Go, D. Jo, T. Gao, K. Ando, S. Blügel, H.-W. Lee, and Y. Mokrousov, Orbital Rashba effect in a surface-oxidized Cu film, Phys. Rev. B 103, L121113 (2021).DOI
- [21]Y.-G. Choi, D. Jo, K.-H. Ko, D. Go, K.-H. Kim, H. G. Park, C. Kim, B.-C. Min, G.-M. Choi, and H.-W. Lee, Observation of the orbital Hall effect in a light metal Ti, Nature (London) 619, 52 (2023).DOI
- [22]M.-C. Chang and Q. Niu, Berry phase, hyperorbits, and the Hofstadter spectrum: Semiclassical dynamics in magnetic bloch bands, Phys. Rev. B 53, 7010 (1996).pubmed.ncbi.nlm.nih.gov/9982146
- [23]D. Xiao, W. Yao, and Q. Niu, Valley-contrasting physics in graphene: Magnetic moment and topological transport, Phys. Rev. Lett. 99, 236809 (2007).DOI
- [24]S. Bhowal and S. Satpathy, Intrinsic orbital moment and prediction of a large orbital Hall effect in two-dimensional transition metal dichalcogenides, Phys. Rev. B 101, 121112(R) (2020).DOI
- [25]S. Bhowal and G. Vignale, Orbital Hall effect as an alternative to valley Hall effect in gapped graphene, Phys. Rev. B 103, 195309 (2021).DOI
- [26]T. Yoda, T. Yokoyama, and S. Murakami, Current-induced orbital and spin magnetizations in crystals with helical structure, Sci. Rep. 5, 12024 (2015).DOI
- [27]T. Yoda, T. Yokoyama, and S. Murakami, Orbital Edelstein effect as a condensed-matter analog of solenoids, Nano Lett. 18, 916 (2018).DOI
- [28]A. El Hamdi, J.-Y. Chauleau, M. Boselli, C. Thibault, C. Gorini, A. Smogunov, C. Barreteau, S. Gargiulo, J.-M. Triscone, and M. Viret, Observation of the orbital inverse Rashba–Edelstein effect, Nat. Phys. 19, 1855 (2023).DOI
- [29]J. Ma and D. A. Pesin, Chiral magnetic effect and natural optical activity in metals with or without Weyl points, Phys. Rev. B 92, 235205 (2015).DOI
- [30]S. Zhong, J. E. Moore, and I. Souza, Gyrotropic magnetic effect and the magnetic moment on the Fermi surface, Phys. Rev. Lett. 116, 077201 (2016).DOI
- [31]J. Rou, C. Şahin, J. Ma, and D. A. Pesin, Kinetic orbital moments and nonlocal transport in disordered metals with nontrivial band geometry, Phys. Rev. B 96, 035120 (2017).DOI
- [32]L.-k. Shi and J. C. W. Song, Symmetry, spin-texture, and tunable quantum geometry in a WTe₂ monolayer, Phys. Rev. B 99, 035403 (2019).DOI
- [33]K. Osumi, T. Zhang, and S. Murakami, Kinetic magnetoelectric effect in topological insulators, Commun. Phys. 4, 211 (2021).DOI
- [34]D. G. Ovalle, A. Pezo, and A. Manchon, Orbital Kerr effect and terahertz detection via the nonlinear Hall effect, Phys. Rev. B 110, 094439 (2024).arxiv.org/abs/2311.11889
- [35]C. Xiao, H. Liu, W. Wu, H. Wang, Q. Niu, and S. A. Yang, Intrinsic nonlinear electric spin generation in centrosymmetric magnets, Phys. Rev. Lett. 129, 086602 (2022).DOI
- [36]M. N. Ali, J. Xiong, S. Flynn, J. Tao, Q. D. Gibson, L. M. Schoop, T. Liang, N. Haldolaarachchi, M. Hirschberger, N. P. Ong, and R. J. Cava, Large, non-saturating magnetoresistance in WTe₂, Nature (London) 514, 205 (2014).DOI
- [37]S. Wu, V. Fatemi, Q. D. Gibson, K. Watanabe, T. Taniguchi, R. J. Cava, and P. Jarillo-Herrero, Observation of the quantum spin Hall effect up to 100 kelvin in a monolayer crystal, Science 359, 76 (2018).DOI
- [38]E. Sajadi, T. Palomaki, Z. Fei, W. Zhao, P. BemPnt, C. Olsen, S. Luescher, X. Xu, J. A. Folk H, and D. H. Cobden, Gate-induced superconductivity in a monolayer topological insulator, Science 362, 922 (2018).DOI
- [39]V. Fatemi, S. Wu, Y. Cao, L. BetheaU, Q. D. Gibson, K. Watanabe, T. Taniguchi, R. J. Cava, and P. Jarillo-Herrero, Electrically tunable low-density superconductivity in a monolayer topological insulator, Science 362, 926 (2018).DOI
- [40]M. Kim, S. Han, J. H. Kim, J.-U. Lee, Z. Lee, and H. Cheong, Determination of the thickness and orientation of few-layer tungsten ditelluride using polarized Raman spectroscopy, 2D Mater. 3, 034004 (2016).DOI
- [41]H. Wang, Y.-X. Huang, H. Liu, X. Feng, J. Zhu, W. Wu, C. Xiao, and S. A. Yang, Orbital origin of the intrinsic planar Hall effect, Phys. Rev. Lett. 132, 056301 (2024).DOI
- [42]H. Liu, J. Zhao, Y.-X. Huang, X. Feng, C. Xiao, W. Wu, S. Lai, W.-b. Gao, and S. A. Yang, Berry connection polarizability tensor and third-order Hall effect, Phys. Rev. B 105, 045118 (2022).DOI
- [43]Y. Gao, S. A. Yang, and Q. Niu, Field induced positional shift of Bloch electrons and its dynamical implications, Phys. Rev. Lett. 112, 166601 (2014).DOI
- [44]S. Lai, H. Liu, Z. Zhang, J. Zhao, X. Feng, N. Wang, C. Tang, Y. Liu, K. S. Novoselov, S. A. Yang, and W.-B. Gao, Third-order nonlinear Hall effect induced by the Berry-connection polarizability tensor, Nat. Nanotechnol. 16, 869 (2021).DOI
- [45]T. Thonhauser, D. Ceresoli, D. Vanderbilt, and R. Resta, Orbital magnetization in periodic insulators, Phys. Rev. Lett. 95, 137205 (2005).pubmed.ncbi.nlm.nih.gov/16197172
- [46]J. C. W. Song and G. Vignale, Low-dissipation edge currents without edge states, Phys. Rev. B 99, 235405 (2019).DOI
- [47]See Supplemental Material about method of device fabrication and transport measurements, theoretical calculations and symmetry discussions about nonlinear Edelstein effect, the schematics about direct magnetic detection, additional experimental results, and the discussions about the origin of the current-induced nonlinear magnetization, which includes Refs. [48–54].
- [48]C. Xiao, W. Wu, H. Wang, Y.-X. Huang, X. Feng, H. Liu, G.-Y. Guo, Q. Niu, and S. A. Yang, Time-reversal-even nonlinear current induced spin polarization, Phys. Rev. Lett. 130, 166302 (2023).DOI
- [49]D. Lee, D. Go, H.-J. Park, W. Jeong, H.-W. Ko, D. Yun, D. Jo, S. Lee, G. J. H. O. Choi, H. K.-J. Kim, B.-G. Park, B.-C. Min, H. C. Koo, H.-W. Lee, O. Lee, and K.-J. Lee, Orbital torque in magnetic bilayers, Nat. Commun. 12, 6710 (2021).DOI
- [50]X.-G. Ye, P.-F. Zhu, W.-Z. Xu, N. Shang, K. Liu, and Z.-M. Liao, Orbit-transfer torque driven field-free switching of perpendicular magnetization, Chin. Phys. Lett. 39, 037303 (2022).DOI
- [51]P. Li, W. Wu, Y. Wen, C. Zhang, J. Zhang, S. Zhang, Z. Yu, S. A. Yang, A. Manchon, and X.-X. Zhang, Spin-momentum locking and spin-orbit torques in magnetic nano-heterojunctions composed of Weyl semimetal WTe₂, Nat. Commun. 9, 3990 (2018).DOI
- [52]B. Zhao, K. Parpiak, D. Khokhlov, A. Johansson, A. Md. Hoque, X. Xu, Y. Jiang, I. Mertig, and S. P. Dash, Unconventional charge-spin conversion in Weyl-semimetal WTe₂, Adv. Mater. 32, 2000818 (2020).DOI
- [53]S. Shi, J. Li, C.-H. Hsu, K. Lee, Y. Wang, J. Yang, J. Wang, Q. Wang, H. Wu, W. Zhang, G. Eda, G. Liang, H. Chang, and H. Yang, Observation of the out-of-plane polarized spin current from CVD grown WTe₂, Adv. Quantum Technol. 4, 2100038 (2021).arxiv.org/abs/2106.04154
- [54]K. Ohnishi, M. Aoki, R. Ohshima, E. Shigematsu, Y. Ando, T. Takenobu, and M. Shiraishi, All-electric spin device operation using the Weyl semimetal, WTe₂, at room temperature, Adv. Electron. Mater. 9, 2200647 (2023).DOI
- [55]I. Baek, S. Han, S. Cheon, and H.-W. Lee, Nonlinear orbital and spin Edelstein effect in centrosymmetric metals, npj Spintronics 2, 33 (2024).arxiv.org/abs/2310.05113
- [56]H. Rostami and M. Polini, Nonlinear anomalous photocurrents in Weyl semimetals, Phys. Rev. B 97, 195151 (2018).DOI
- [57]Z.-C. Pan, D. Li, X.-G. Ye, Z. Chen, Z.-H. Chen, A.-Q. Wang, M. Tian, G. Yao, K. Liu, and Z.-M. Liao, Room-temperature orbit-transfer torque enabling van der Waals magnetoresistive memories, Sci. Bull. 68, 2743 (2023).DOI









