SPIN-ORBIT TORQUE DEVICE

Abstract

A spin-orbit torque device is disclosed, which includes: a magnetic layer; and a non-magnetic layer adjacent to the magnetic layer and comprising a spin-Hall material, wherein the spin-Hall material comprises Ni.sub.xCu.sub.1-x alloy, and x is in a range from 0.4 to 0.8.

Claims

1. A spin-orbit torque device, comprising: a non-magnetic layer comprising a spin-Hall material, wherein the spin-Hall material comprises Ni.sub.xCu.sub.1-x alloy, and x is in a range from 0.4 to 0.8.

2. The spin-orbit torque device of claim 1, wherein x is in a range from 0.7 to 0.8.

3. The spin-orbit torque device of claim 2, wherein x is in a range from 0.75 to 0.8.

4. The spin-orbit torque device of claim 1, wherein the non-magnetic layer has a thickness ranging from 0.1 nm to 8 nm.

5. The spin-orbit torque device of claim 4, wherein the non-magnetic layer has a thickness ranging from 2 nm to 8 nm.

6. The spin-orbit torque device of claim 5, wherein the non-magnetic layer has the thickness ranging from 3 nm to 8 nm.

7. The spin-orbit torque device of claim 1, wherein the spin-Hall material is Ni.sub.80Cu.sub.20 alloy, and the non-magnetic layer has a thickness ranging from 4.5 nm to 5.5 nm.

8. The spin-orbit torque device of claim 7, wherein a Curie temperature of the Ni.sub.80Cu.sub.20 alloy is about room temperature.

9. The spin-orbit torque device of claim 1, wherein the spin-Hall material has the spin Hall angle of 42% to 50% when the spin-Hall material is in the paramagnetic state.

10. The spin-orbit torque device of claim 1, wherein the spin-Hall material has the spin Hall angle of 8% to 15% when the spin-Hall material is in the ferromagnetic state.

11. The spin-orbit torque device of claim 1, wherein the spin-Hall material has the spin diffusion length of 0.2 nm to 0.3 nm when the spin-Hall material is in the paramagnetic state.

12. The spin-orbit torque device of claim 1, wherein the spin-Hall material has the spin diffusion length of 0.4 nm to 0.5 nm when the spin-Hall material is in the ferromagnetic state.

13. The spin-orbit torque device of claim 1, further comprising a magnetic layer adjacent to the non-magnetic layer.

14. The spin-orbit torque device of claim 1, which is a magnetic random access memory.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] FIG. 1A is a diagram showing X-ray diffraction pattern of the 200-nm-thick Ni.sub.xCu.sub.1-x (0<x<1.0) with different composition.

[0023] FIG. 1B is a diagram showing the composition dependence of lattice constant a for Ni.sub.xCu.sub.1-x.

[0024] FIG. 1C is a schematic view of anomalous Nernst effect (ANE) in a ferromagnetic metal (FM) and inverse spin Hall effect (ISHE) and ANE in FM/YIG under a temperature gradient.

[0025] FIG. 2A is a diagram showing Temperature-dependent magnetization of Ni.sub.75Cu.sub.25 (5 nm)/Si measured by the SQUID magnetometer.

[0026] FIG. 2B is a diagram showing Temperature-dependence of ANE of Ni.sub.75Cu.sub.25 (5 nm)/Si.

[0027] FIG. 2C is a diagram showing the spin dependent thermal voltage as a function of magnetic field (H) measured in Ni.sub.75Cu.sub.25 (5 nm)/Si and Ni.sub.75Cu.sub.25 (5 nm)/YIG at 300 K above its T.sub.c.

[0028] FIG. 2D is a diagram showing the spin dependent thermal voltage as a function of magnetic field (H) measured in Ni.sub.75Cu.sub.25 (5 nm)/Si and Ni.sub.75Cu.sub.25 (5 nm)/YIG at 200 K below its T.sub.c.

[0029] FIG. 3A is a diagram showing temperature-dependent ANE (solid circles) and ISHE (hollow circles) voltage for Ni.sub.xCu.sub.1-x with x being 0.5, 0.6, 0.7, 0.75, and 0.8.

[0030] FIG. 3B is a diagram showing values of T.sub.P, T.sub.A, and T.sub.c as a function of Ni.sub.xCu.sub.1-x compositions.

[0031] FIG. 3C is a diagram showing temperature-dependent ANE (solid circles) and ISHE (hollow circles) voltage of Ni.sub.40Cu.sub.60.

[0032] FIG. 4A is a diagram showing composition-dependence of ΔV.sub.ANE(hollow circles) for Ni.sub.xCu.sub.1-x (5 nm)/Si and that of ΔV.sub.ISHE (solid circles) for Ni.sub.xCu.sub.1-x (5 nm)/YIG with 0<x<1.0 at room temperature.

[0033] FIG. 4B is a diagram showing thickness-dependence of ΔV.sub.ISHE/ρ for Ni.sub.80Cu.sub.20, wherein the solid circles and the hollow circles are the fitted results using Eq. (2) (described hereinafter) for Ni.sub.80Cu.sub.20 in PM and FM.

[0034] FIG. 4C is a plot of θ.sub.SH vs λ.sub.sd at room temperature, wherein the hollow circles are the result for Ni.sub.80Cu.sub.20 in the PM and FM states, other solid symbols are results for Pt in literature, and the curve represents θ.sub.SH* λ.sub.sd=0.13 nm.

[0035] FIG. 5 is a cross-sectional view showing a spin-orbit torque device according to one embodiment of the present disclosure.

DETAILED DESCRIPTION OF EMBODIMENT

[0036] Different embodiments of the present disclosure are provided in the following description. These embodiments are meant to explain the technical content of the present disclosure, but not meant to limit the scope of the present disclosure. A feature described in an embodiment may be applied to other embodiments by suitable modification, substitution, combination, or separation.

[0037] It should be noted that, in the present specification, when a component is described to have an element, it means that the component may have one or more of the elements, and it does not mean that the component has only one of the element, except otherwise specified.

[0038] Moreover, in the present specification, when an element is described to be “suitable for” or “adapted to” another element, the other element is an example or a reference helpful in imagination of properties or applications of the element, and the other element is not to be considered to form a part of a claimed subject matter; similarly, except otherwise specified; similarly, in the present specification, when an element is described to be “suitable for” or “adapted to” a configuration or an action, the description is made to focus on properties or applications of the element, and it does not essentially mean that the configuration has been set or the action has been performed, except otherwise specified.

[0039] Moreover, in the present specification, a value may be interpreted to cover a range within ±10% of the value, and in particular, a range within ±5% of the value, except otherwise specified; a range may be interpreted to be composed of a plurality of subranges defined by a smaller endpoint, a smaller quartile, a median, a greater quartile, and a greater endpoint, except otherwise specified.

Experimental Method

[0040] The Si substrate and the YIG substrate were provided, and cleaned by acetone and then isopropyl alcohol about 30 minutes. Then, ethanol was used to remove the possible residual contamination (such as dust or particles) for about 10 minutes. Finally, DI-water was used to remove the residual organic solvent for about 10 minutes.

[0041] The NiXCu.sub.1-x thin films were deposited on the cleaned Si substrate and YIG substrate respectively. The Ni.sub.xCu.sub.1-x thin films can be prepared by any method know in the art. Herein, the sputtering technique was used to prepare the Ni.sub.xCu.sub.1-x thin films on the Si substrate and the YIG substrate. The sputtering system was operated under high vacuum environment with the order of 10.sup.−7˜10.sup.−8 torr, and the sputtering process was performed by using argon plasma. However, the present disclosure is not limited thereto. In another embodiment of the present disclosure, the magnetron sputtering may be used to improve the efficiency of sputtering, especially for magnetic materials.

[0042] By controlling the deposition rates or the sputtering target, the composition of the Ni.sub.xCu.sub.1-x thin films can be adjusted. The X-ray reflectometry and atomic force microscope were used to measure film thickness and surface roughness, the X-ray diffraction (XRD) was used to measure crystal structures and film orientations, and the magnetometers was used to measure the magnetic properties of the Ni.sub.xCu.sub.1-x thin films. For spin-dependent transport measurements, the Ni.sub.xCu.sub.1-x films protected by a 2-nm Al film were patterned into Hall bar structure with widths of 200 μm by photolithography.

Results

[0043] The X-ray diffraction (XRD) patterns of the 200-nm-thick Ni.sub.xCu.sub.1-x alloys show they all are mainly fcc (111)-textured as shown in FIG. 1A. With increasing Ni content, the (111) peak progressively shifts to higher diffraction angles because the lattice constant of 0.351 nm of Ni is smaller than that of 0.361 nm of Cu. The fcc lattice parameter (a) depends linearly on the Ni content, a manifestation of the Vegard's law, as shown in FIG. 1B.

[0044] There are various ways to inject spin-polarized current and pure spin current via electrical (e.g., anomalous Hall, spin Hall), thermal (anomalous Nernst, spin Seebeck), and FMR excitations (e.g., spin pumping). Spin pumping and electrical injection may inadvertently include other contributions especially thermal due to the high current density and FMR heating. Longitudinal thermal injection in the out-of-plane direction via anomalous Nernst effect (ANE) and spin Seebeck effect (SSE) are the simplest injection schemes with little parasitic effects.

[0045] FIG. 1C is a schematic view of anomalous Nernst effect (ANE) in a ferromagnetic metal (FM) and inverse spin Hall effect (ISHE) and ANE in FM/YIG under a temperature gradient.

[0046] The left device shown in FIG. 1C is the device for measuring the voltages under a temperature gradient, wherein the FM film 12 (for example, the Ni.sub.xCu.sub.1-x film) is formed on the Si substrate 11, and a voltage meter (not shown in the figure) is electrically connected to the pads (not shown in the figure) on the FM film 12. In addition, a heater (not shown in the figure) is also provided in this device to provide temperature gradient to the FM film 12. By using the left device shown in FIG. 1C, the ANE voltage (V.sub.ANE) in the FM film 12 can be measured.

[0047] The right device shown in FIG. 1C is similar to the left device, except that the Si substrate 11 shown in the left device is replaced by the YIG substrate 21 shown in the right device. By using the right device shown in FIG. 1C, the ANE voltage (V.sub.ANE) plus ISHE voltage (V.sub.ISHE) in the FM film 12 can be measured by a voltage meter (not shown in the figure). It should be noted that, in the right device shown in FIG. 1C, the ANE voltage and the ISHE voltage are shown separately, but the ANE voltage and the ISHE voltage may be measured together by one voltage meter.

[0048] As shown in FIG. 1C, for a FM with in-plane magnetization along the x direction, a temperature gradient (∇T) in the out-of-plane (z) direction injects a charge current in the z-direction. The spin-orbit coupling (SOC) in the FM causes unequal amount of spin-up and spin-down electrons to deflect laterally in opposite directions, resulting in a spin-polarized current due to the ANE electric field in the y-direction of the following Eq. (1):

E.sub.ANE−Q.sub.S4πM×∇T,  (1)

and detected as an ANE voltage in the y-direction.

[0049] In the longitudinal spin Seebeck effect (SSE) scheme, one places a thin metal film on ferromagnetic YIG also in a vertical temperature gradient, which injects a pure spin current j.sub.S in the z-direction with spin index in the x-direction into the metal, in which the SOC causes both spin-up and spin-down electrons to deflect laterally to the same side in the y-direction via the ISHE electric field proportional to σ×∇T or σ×jd.sub.S, detected as an ISHE voltage. In the case of a FM metal, with magnetization in the x-direction aligned by an external field, the electrical fields due to ANE and ISHE are both in the y-directions, thus their voltages are additive.

[0050] The SQUID magnetometer with high sensitivity is used to measure the small magnetization (˜10.sup.−5 emu) and the magnetic ordering temperature of thin Ni.sub.xCu.sub.1-x alloys, typically 5 nm thick. An example of Ni.sub.75Cu.sub.25 (5)/Si (number in parentheses is the thickness in nanometer) for revealing Curie temperature (T.sub.c) of 260K is shown in FIG. 2A. From Eq. (1) shown above, under a constant |∇T|, the ANE voltage is proportional to the magnetization M. Thus, ANE can electrically measure M. As shown in FIG. 2C and FIG. 2D, the ANE voltage in Ni.sub.75Cu.sub.25 (5)/Si while sizable at 200 K, vanishes at 300 K. The ANE voltage in fact vanishes abruptly at T.sub.c of 260K, as shown in FIG. 2B. The ANE readily measures the hysteresis loop and T.sub.c in thin FM films only a few nm thick, can function as a sensitive magnetometer for measuring FM with in-plane magnetization, in a manner similar to that of anomalous Hall effect (AHE) as a sensitive magnetometer for measuring FM with perpendicular magnetization.

[0051] When Ni.sub.75Cu.sub.25 (5)/YIG is subjected to a similar out-of-plane temperature gradient of 20 K/mm, in addition to the ANE within Ni.sub.75Cu.sub.25 (5), there is also pure spin current injection from YIG via the SSE with the resultant ISHE voltage. As shown in FIG. 2C and FIG. 2D, voltage in Ni.sub.75Cu.sub.25 (5)/YIG at 200 K and also at 300 K even after Ni.sub.75Cu.sub.25 (5) has become paramagnetic at T>260 K, where there is no spin-polarized current and only pure spin current. The lateral voltages measured due to ANE and ISHE shown in FIG. 2C and FIG. 2D saturates at large ±H fields. The value of ΔV is defined by these voltages. In Ni.sub.75Cu.sub.25 (5)/Si, one observes only ΔV=ΔV.sub.ANE of about 2.6 μV at 200 K and 0 μV at 300 K. However, in Ni.sub.75Cu.sub.25 (5)/YIG under similar temperature gradient, one observes ΔV=ΔV.sub.ISHE=5 μV at 300 K when Ni.sub.75Cu.sub.25 (5) is paramagnetic with ΔV.sub.ANE=0 V. Importantly, one observes an even larger ΔV=ΔV.sub.ISHE+ΔV.sub.ANE=9.5 μV at 200 K that contains both the ANE and the ISHE contributions, where ΔV.sub.ANE can be measured in Ni.sub.75Cu.sub.25 (5)/Si. These results provide clear evidences that ferromagnetic alloys exhibit substantial spin-to-charge conversion in the ferromagnetic state as well as in the paramagnetic state. In FIG. 2C and FIG. 2D, the presence of a plateau behavior in the low field region in the ISHE voltage is due to the effect of demagnetizing factor from surface magnetization of YIG The ANE loop at 200 K as shown in FIG. 2D reveals the coercivity of Ni.sub.75Cu.sub.25 (5).

[0052] The present example also displays results in S(μV/K)=ΔV/AT, where ΔT is temperature difference. FIG. 3A shows S(μV/K) of ΔV.sub.ANE and ΔV.sub.ISHE, the latter with ΔV.sub.ANE subtracted from ΔV, as a function of temperature across the phase transition of Ni.sub.xCu.sub.1-x with a range of compositions (0.4≤x0.8). The S(μV/K) of Ni.sub.xCu.sub.1-x/Si (solid circles), consisting of only ANE, reveals a sharp phase transition at T.sub.A, which is just T.sub.C, above which Ni.sub.xCu.sub.1-x is in the paramagnetic state with no ANE. On the other hand, S(μV/K) of Ni.sub.xCu.sub.1-x/YIG (hollow circles), containing ANE and the pure spin current contribution, is always substantial below and even above T.sub.C.

[0053] Also prominently displayed is the pure spin current enhancement due to spin fluctuations, most intensely near T.sub.C, at which S(μV/K) is maximal. These results show clearly that spin fluctuations in Ni—Cu alloys can greatly enhance the already substantial spin-to-charge conversion. Above T.sub.C, spin fluctuation decreases with increasing temperature, so are its effect on pure spin current enhancement. The magnetic ordering temperature of Ni—Cu alloys are shown in FIG. 3B, where the values of T.sub.C determined by SQUID magnetometry, T.sub.A where ANE in Ni—Cu/Si vanishes, and T.sub.P where S(μV/K) in Ni—Cu/YIG is maximal, are in very good agreement. All three methods can be used to determine the ordering temperatures of FM materials. But the ANE method enjoys the clear advantages of higher sensitivity especially beneficial for thin films. In Ni—Cu alloys, the ordering temperature decreases linearly with reducing Ni content, and becomes non-magnetic at about x=0.45. However, as shown in FIG. 3C while there is no magnetic ordering and no ANE in Ni.sub.40Cu.sub.60/Si down to about 20 K, there is substantial S(μV/K) at all temperatures in Ni.sub.40Cu.sub.60/YIG, increasing in values for decreasing temperature reflecting the incipient magnetic ordering and the presence of spin fluctuation. At a low temperature of about 60 K, S(μV/K) decreases sharply towards zero as it should when T=0 K approaches. Noted, the competition between the propagation length of magnon and the concentration of magnon in YIG can lead to nonmonotonic temperature-dependent behavior similar to the enhancement of the spin current (e.g., Pt/YIG). However, two contributions can be clearly distinguished in Ni.sub.80Cu.sub.20, since spin current enhancement occurs at around room temperature, but the contribution of the magnon population from YIG remains at low temperatures.

[0054] Furthermore, the interplay of the pure spin current and the spin-polarized current can also be apparent at room temperature when one compare the spin-dependent thermal voltages of ANE and ISHE for 5-nm Ni.sub.xCu.sub.1-x in a wide range of compositions (0≤x≤1.0) measured. As shown in FIG. 4A, the ΔV.sub.ISHE (solid circles) measured at 300 K increases with the Ni content until Ni.sub.80Cu.sub.20, beyond which ΔV.sub.ISHE decreases with the simultaneous appearance of ΔV.sub.ANE (hollow circles), where 5-nm Ni.sub.xCu.sub.1-x with x>0.8 is ferromagnetic. Therefore, in the specific case of Ni.sub.80Cu.sub.20, the spin current can be substantially enhanced at room temperature through spin fluctuation near the phase transition. Thus, one can exploit spin fluctuations to greatly enhance pure spin current in Ni.sub.80Cu.sub.20 for room temperature operations.

[0055] To quantitatively determine the enhanced spin-to-charge efficiency of Ni.sub.80Cu.sub.20 at room temperature, SSE measurements in Ni.sub.80Cu.sub.20 of a series of thicknesses were performed to evaluate spin Hall angle θ.sub.SH and spin diffusion length λ.sub.sd. The ISHE voltage depends on the thickness t as the following Eq. (2):

[00001]$\begin{array}{cc}\Delta V_{\mathrm{ISHE}}\left(t\right)=2\mathrm{CL}\nabla T\rho \left(t\right){\theta }_{\mathrm{SH}}\frac{{\lambda }_{\mathrm{sd}}}{t}\tanh \left(\frac{t}{2{\lambda }_{\mathrm{sd}}}\right)& \left(2\right)\end{array}$

where L=6 mm is the distance between the voltage terminals, ΔT|=26 K/mm is the temperature gradient, and C is the spin current injection coefficient. The resistivity (t) is also determined through experiments. From the linear interpolation with C(Ni)=1.55 Am.sup.−1K.sup.−1 for Ni and C(Cu)=1.24 Am.sup.−1K.sup.−1 for Cu, C(Ni.sub.80Cu.sub.20)=1.5 Am.sup.−1K.sup.−1 for Ni.sub.80Cu.sub.20 can be obtained.

[0056] For non-magnetic metals (e.g., Pt), ΔV.sub.ISHE(t)/ρ(t) decreases with increasing t in a quasi-hyperbolic manner. However, the results of ΔV.sub.ISHE(t)/ρ(t) shown in FIG. 4B for Ni.sub.80Cu.sub.20 exhibits a clear break at t=7 nm because its ordering temperature is close to 300 K. Samples of Ni.sub.80Cu.sub.20 with t greater and less than 7 nm has T.sub.C above and below 300 K respectively, hence a discontinuity at 7 nm in FIG. 4B. From the fitting in FIG. 4B (solid lines) by Eq. (2), when Ni.sub.80Cu.sub.20 is in the FM state, θ.sub.SH=110 and λ.sub.sd=0.42 nm can be obtained, which are comparable to those of heavy metals. On the other hand, when Ni.sub.80Cu.sub.20 is in the PM state, a large enhancement to value of θ.sub.SH=460 and λ.sub.sd=0.22 nm can be obtained. Under the definition using number of carriers of θ.sub.SH≤1, θ.sub.SH=46% is the largest reported to date.

[0057] Although the θ.sub.SH value of metals may vary greatly (e.g., Pt), depending on the experimental technique or the analyses, empirically the relation of θ.sub.SH.Math.λ.sub.sd≈constant has been suggested as shown in FIG. 4C containing the results of various reports. Our results of θ.sub.SH=110 and λ.sub.sd=0.42 nm in the FM state, and θ.sub.SH=46% and λ.sub.sd=0.22 nm in the PM state of Ni.sub.80Cu.sub.20 appear to be also consistent with this correlation of θ.sub.SH.Math.λ.sub.sd˜0.13 nm denoted as the line. The enhanced spin-to-charge conversion and the larger θ.sub.SH in Ni.sub.80Cu.sub.20 to short-range spin fluctuation leads to the shorter λ.sub.sd. Ni—Cu alloys in general, and Ni.sub.80Cu.sub.20 in particular, not only exhibit much larger spin-to-charge efficiency than those of Pt, they are about three orders of magnitude less costly than Pt.

[0058] In summary, the present disclosure show the strong interplay of the anomalous Nernst effect of the spin-polarized current, the inverse spin Hall effect of the pure spin current, and spin fluctuation in magnetic alloys that has been revealed in Ni.sub.xCu.sub.1-x in a wide range of compositions (0≤x≤1.0) with tailored magnetic ordering temperature. We demonstrate the strong interaction of pure spin current and spin fluctuation can greatly enhance spin-to-charge conversion, yielding remarkably high spin Hall angle of 46% in Ni.sub.80Cu.sub.20 at room temperature, that can be exploited in various spin-based applications and devices. We also show that the spin-dependent thermal transport via the ANE can serve as a sensitive magnetometer to electrically detect magnetic phase transitions.

[0059] FIG. 5 is a cross-sectional view showing a spin-orbit torque device according to one embodiment of the present disclosure.

[0060] The spin-orbit torque device of the present embodiment comprises: a non-magnetic layer 42 comprising a spin-Hall material; and a magnetic layer 41 adjacent to the non-magnetic layer 42. Herein, the spin-Hall material may be the Ni.sub.xCu.sub.1-x alloy described above. The magnetic layer 41 may comprise a ferromagnetic material, and the ferromagnetic material may comprise Fe, Ni, Co or alloy thereof; but the present disclosure is not limited thereto. In addition, the spin-orbit torque device of the present embodiment may be used as a magnetic random access memory.

[0061] When the non-magnetic layer 42 is a Ni.sub.80Cu.sub.20 layer having a thickness of about 5 nm, the spin-orbit torque device can be operated at room temperature. However, the present disclosure is not limited thereto. In another embodiment of the present disclosure, x in the Ni.sub.xCu.sub.1-x alloy and the thickness of the non-magnetic layer 42 can be adjusted to achieve the spin-orbit torque device capable of operating at room temperature or at other temperature under or above the room temperature, according to the need.

[0062] Although the present disclosure has been explained in relation to its embodiment, it is to be understood that many other possible modifications and variations can be made without departing from the spirit and scope of the disclosure as hereinafter claimed.