Beta tungsten thin films with giant spin Hall effect for use in compositions and structures with perpendicular magnetic anisotropy

Abstract

Methods, devices, and compositions for use with spintronic devices such as magnetic random access memory (MRAM) and spin-logic devices are provided. Methods include manipulating magnetization states in spintronic devices and making a structure using spin transfer torque to induce magnetization reversal. A device described herein manipulates magnetization states in spintronic devices and includes a non-magnetic metal to generate spin current based on the giant spin Hall effect, a ferromagnetic thin film with perpendicular magnetic anisotropy, an oxide thin film, and an integrated magnetic sensor. The device does not require an insertion layer between a non-magnetic metal with giant spin Hall effect and a ferromagnetic thin film to achieve perpendicular magnetic anisotropy.

Claims

1. A multilayer composition for manipulating magnetization states in spintronic devices comprising: a non-magnetic metal layer having a spin Hall angle that converts an electrical current to a traverse spin current, the non-magnetic metal layer having a thickness that is less than the critical thickness (t.sub.c) of the non-magnetic metal layer, and the non-magnetic metal layer having a substantially constant resistivity throughout a temperature range of about 5 K to about 600 K; a free layer having perpendicular magnetic anisotropy and an interface with the non-magnetic metal layer, and the free layer is configured to accept the traverse spin current from the non-magnetic metal layer to produce spin-transfer torque (STT) in the free layer and magnetization switching; and an oxide barrier layer with a crystalline structure having an interface with the free layer.

2. The composition according to claim 1, wherein the non-magnetic layer is characterized by at least one of the following properties: the resistivity is constant to within about 15% throughout the temperature range; has an interface with a thermally oxidized Si wafer; the SST induces switching in the free layer at a current density of less than or equal to about 10.sup.6 A/cm.sup.2 in the non-magnetic layer; the spin Hall angle is selected from the group consisting of: at least about 0.01, at least about 0.05, at least about 0.1, at least about 0.5, at least about 1, at least about 5, and at least about 10; the resistivity selected from the group consisting of: at least about 50 μΩ-cm, at least about 150 μΩ-cm, and at least about 250 μΩ-cm, at least about 300 μΩ-cm, at least about 350 μΩ-cm, at least about 400 μΩ-cm, at least about 450 μΩ-cm, and at least about 500μΩ-cm; has a spin diffusion length selected from the group consisting of: at least about 0.1 nm, at least about 0.5 nm, at least about 1 nm, at least about 5 nm, at least about 10 nm, at least about 20 nm, at least about 30 nm, and at least about 40 nm; and is at least one metal selected from the group consisting of: scandium, titanium, vanadium, yttrium, zirconium, niobium, molybdenum, ruthenium, rhodium, palladium, hafnium, tantalum, tungsten, osmium, iridium, platinum, gold, gallium, germanium, arsenic, selenium, indium, tin, antimony, tellurium, thallium, lead, and bismuth.

3. (canceled)

4. The composition according to claim 1, wherein the free layer is characterized by at least one of: has a coercivity selected from the group consisting of: at least about 0.1 Oe, at least about 1 Oe, at least about 10 Oe, and at least about 100 Oe; is a ferromagnetic thin film; the ferromagnetic thin film is (Co.sub.xFe.sub.100-x).sub.100-yB.sub.y, wherein 0<x<100 and 0<y<100 (atomic percent), referred to as CoFeB; and the percent composition of the CoFeB is Co.sub.40Fe.sub.40B.sub.20.

5-6. (canceled)

7. The composition according to claim 1, further characterized by at least one of the following: has a stacked structure; comprises a capping layer or overlayer to prevent oxidation, the capping layer or overlayer having an interface with the oxide barrier layer; the capping layer or overlayer comprises tantalum; has an insertion layer; the insertion layer comprises hafnium; and the insertion layer is not between the non-magnetic layer and the free layer.

8-11. (canceled)

12. The composition according to claim 1, wherein the non-magnetic metal layer is β-form tungsten, optionally the β-form tungsten being a single phase A.sub.3B solid with A15 crystal structure prior to being annealed.

13. The composition according to claim 12, wherein the tungsten has a thickness selected from a range of about 0.5 nm to about 100 nm.

14-17. (canceled)

18. The composition according to claim 1, wherein the non-magnetic metal layer and the multilayer composition are annealed to obtain at least one of the perpendicularly magnetic anisotropy or a stabilized structure of the non-magnetic metal layer.

19. (canceled)

20. The composition according to claim 1, wherein the multilayer composition is annealed under a magnetic field perpendicular to the plane of the multilayer composition having a strength selected from the group consisting of: at least about 0.01 T, at least about 0.1 T, at least about 1.0 T, at least about 10 T, and at least about 20 T.

21. (canceled)

22. The composition according to claim 18, wherein the β-form tungsten is single phase having an A.sub.3B solid with A15 crystal structure prior to being annealed.

23-24. (canceled)

25. A method of manipulating magnetization states in a spintronic device, the method comprising: applying an electric current to a multilayer composition having: a non-magnetic metal layer having a spin Hall angle of sufficient size to convert an electrical current to a traverse spin current, a free layer with perpendicular magnetic anisotropy having an interface with the non-magnetic metal layer, and an oxide barrier layer having an interface with the ferromagnetic layer; and creating spin-transfer torque (STT) current in the free layer by the traverse spin current from the non-magnetic metal layer yielding magnetization switching.

26. The method according to claim 25, further comprising the step of measuring the resistivity of the non-magnetic metal to identify the phase.

27. The method according to claim 25, wherein creating the STT current for magnetization reversal further comprises selecting from a range of about 10.sup.4 A/cm.sup.2 to about 10.sup.7 A/cm.sup.2.

28-30. (canceled)

31. The method according to claim 25, further comprising detecting with a magnetic sensor, magnetic states representing digital bits, wherein the magnetic sensor is a magnetic tunneling junction or a giant magnetoresistive (GMR) element or an anomalous Hall effect sensor.

32. (canceled)

33. The method according to claim 25, further comprising, prior to applying the electric current, annealing the multilayer composition, the annealing performed at a temperature selected from the group consisting of: at least about 50° C., at least about 100° C., at least about 200° C., at least about 300° C., at least about 400° C., at least about 500° C., at least about 600° C., and at least about 700° C.

34-37. (canceled)

38. A method of making a structure for manipulating magnetization states in a spintronic device, the method comprising: depositing sequentially layers of tungsten, (Co.sub.xFe.sub.100-x).sub.100-yB.sub.y (CoFeB), wherein x and y are each an atomic percent and 0<x<100 and 0<y<100, and MgO on a thermally oxidized Si wafer to form a stacked structure using a high vacuum magnetron sputtering system, and the tungsten layer of the stacked structure has a thickness selected from the range of about 1 nm to about 100 nm.

39. The method according to claim 38, further comprising, after depositing tungsten, CoFeB, and MgO, depositing a capping layer or an overlayer having an interface with MgO that prevents oxidation of the stacked structure.

40. (canceled)

41. The method according to claim 38, wherein the sputtering system is characterized by at least one of the following properties: a base pressure selected from the group consisting of: less than about 1×10.sup.−10 Torr, less than about 1×10.sup.−9 Torr, less than about 1×10.sup.−8 Torr, less than about 1×10.sup.−7 Torr, and less than about 1×10.sup.−6; and a sputtering pressure selected from the group consisting of: at least about 0.1 mTorr, at least about 1 mTorr, at least about 5 mTorr, at least about 10 mTorr, at least about 15 mTorr, and at least about 20 mTorr.

42. (canceled)

43. The method according to claim 38, wherein the tungsten is in β-form.

44. The method according to claim 38, wherein the CoFeB layer of the stack is deposited at a thickness selected from the group consisting of: at least about 0.1 nm, at least about 1.0 nm, at least about 10 nm, at least about 20 nm, at least about 30 nm, and at least about 40 nm.

45. The method according to claim 38, wherein depositing is performed at least one of the following: selecting a DC and RF sputtering power from the group consisting of: at least about 1 W, at least about 5 W, at least about 10 W, at least about 50 W, at least about 100 W, at least about 500 W, and at least about 1000 W; selecting a rate from the group consisting of: less than about 0.005 nm/s, less than about 0.01 nm/s, less than about 0.05 nm/s, less than about 0.1 nm/s, less than about 0.5 nm/s, less than about 1.0 nm/s, less than about 5 Innis, and less than about 10 nm/s; and depositing the CoFeB layer of the stack at a thickness selected from the group consisting of: at least about 0.1 nm, at least about 1.0 nm, at least about 10 nm, at least about 20 nm, at least about 30 nm, and at least about 40 nm.

46. (canceled)

47. The method according to claim 38, further comprising annealing the tungsten layer as a component of the multilayer composition.

48. The method according to claim 38, further comprising, after the depositing step, annealing the stacked structure at a temperature selected from the group consisting of: greater than about 50° C., greater than about 100° C., greater than about 200° C., greater than about 400° C., and greater than about 600° C.

49. A method of fabricating a metastable thin film, the method comprising: cooling a substrate to a temperature less than ambient temperature for deposition; combining argon and oxygen to form a sputtering gas; and depositing onto the substrate the metastable thin film using the sputtering gas at a sputtering power up to 1000 W, and at a rate in a range of about 0.02 nm/s to about 10 nm/s.

50. The method according to claim 49, wherein the metastable thin film is β-form tungsten.

51. The method according to claim 49, wherein cooling the substrate is to the temperature of liquid nitrogen.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0039] FIG. 1 is a line graph of the sheet resistance (R.sub.□) of W(thickness)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm) multilayer stacks as a function of W thickness (t) in the range of 2.5 nm-7.0 nm. The solid line is the best fit to the data based on extracted resistivities of ρ.sub.W of about 209 μΩ-cm and ρ.sub.FeCoB of about 80 μΩ-cm.

[0040] FIG. 2A is a schematic of the W/CoFeB bilayer in the Hall bar configuration for magnetotransport measurement under an external magnetic field (B.sub.ext) and an excitation DC current (I). β is the angle between the vector of B.sub.ext and y-axis. J.sub.c is the charge current density in the W layer, and J.sub.s is the SHE converted spin current density into the CoFeB layer. M is the magnetization vector of the CoFeB FM layer. θ is the angle between the vector M and the y-axis.

[0041] FIG. 2B is a graph of anomalous Hall resistance (R.sub.H) as a function of cycling external magnetic field, B.sub.ext (Oe), applied perpendicular (β=90°) to the stacks of W(t)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm), in which t is in the range of 3 nm-7 nm. The square-shaped loops for every W thickness indicate that CoFeB exhibits PMA for each sample with a different W thickness. PMA in this structure has not previously been observed.

[0042] FIG. 2C is a graph of current-induced magnetic switching curves in the W(7.0 nm)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm) multilayer stack sample, under a negative (β=0°) external field B.sub.ext at each of 0.2 mT, 0.4 mT, 0.7 mT, and 2 mT. From each switching curve, critical current (I.sub.C) was obtained.

[0043] FIG. 2D is a graph of current-induced magnetic switching curves in the W(7.0 nm)/CoFeB(1 nm)/MgO(1.6 nm)/Ta(1 nm) sample, under a positive (β=180°) external field B.sub.ext at each of 0.2 mT, 0.4 mT, 0.7 mT, and 2 mT. From each switching curve, critical current (I.sub.C) was obtained.

[0044] FIG. 3A is a line graph of normalized Hall resistance (i.e., sin θ) as a function of nearly in-plane magnetic field B.sub.ext (β=2°) under a positive or negative current (+/−1 mA) for β-W in W(6.5 nm)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm). θ is the angle between the magnetization vector M and the Y-axis. B.sub.+(θ) and B.sub.−(θ) are the magnetic fields required to rotate the M to θ corresponding to the positive and the negative current.

[0045] FIG. 3B is a line graph of linear relationships between β.sub.+(θ)-β(θ) and 1/sin(θ−β) under different excitation currents in the range of 1 mA-2 mA for W at a thickness of 6.5 nm in W(6.5 nm)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm). The slope for each fitted straight line is the net STT, Δτ.sub.ST.sup.0, between the positive and negative excitation current.

[0046] FIG. 3C is a graph of spin-transfer torques (τ.sub.ST.sup.0) as a function of excitation current for W(t)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0), with tin the range of 3.0 nm to 7.0 nm. Torque values are linear in current and vanish as current approaches zero. Thicker layers of W generate more torque per unit of current.

[0047] FIG. 4A is a line graph of spin Hall angles versus W thickness for W(t)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0), with tin the range of 3.0 nm to 7.0 nm. The line represents theoretical fitting to the data assuming a finite spin diffusion length in the β-W film. For the bulk β-W film, spin Hall angle was observed to be 0.40±0.05 and spin diffusion length was observed to be 3.1±0.4 nm at room temperature.

[0048] FIG. 4B is a line graph of spin Hall angles versus W thickness for W(t)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm), with tin the range of 3.0 nm to 22.1 nm. The lines represents theoretical fitting of the data assuming a finite spin diffusion length in the β-W film. For the bulk β-W film, spin Hall angle was observed to be 0.49±0.02 and spin diffusion length was observed to be 4.3±0.3 nm at room temperature.

[0049] FIG. 5 is a magnetic switching phase diagram of W(7.0 nm)/CoFeB(1.0 nm)/MgO(1.6 nm)/Ta(1.0), in the parameter space of β.sub.ext and critical current (I.sub.c) or J.sub.c. The arrows (↑ or ↓) denote the directions of the M vector in various regions. I.sub.c is the net current into the Hall bar, and J.sub.c is the corresponding current density only in the W layer.

[0050] FIG. 6 contains plots of θ-2θ x-ray diffraction patterns for as-deposited and annealed W thin films having various thicknesses.

[0051] FIG. 7A is a line graph of lattice constant determined from x-ray diffraction as a function of W film thickness for as-deposited and annealed samples.

[0052] FIG. 7B is a line graph of grain size determined by using the Scherrer equation as a function of W film thickness.

[0053] FIG. 8A is a plot of temperature dependence of resistivity for β-W and α-W thin films between 10 K and 380 K.

[0054] FIG. 8B is a line graph of resistivities at 300 K and 10 K as a function of an inverse of film thickness (1/t) for β-W films in the range of 3.0 nm to 26.7 nm. The solid straight line is the theoretical fit using Eq. (6), in the range of 3.0 nm and 22.1 nm based on the finite-size effect of thin film resistivity.

[0055] FIG. 9 is a line graph of temperature dependence of normal Hall coefficient of β-W at a thickness of 26.7 nm, and α-W at thicknesses of 24.1 nm and 26.7 nm. Normal spin Hall effect was measured between −5 T and +5 T. There is a sign change at about 130 K for α-W.

[0056] FIG. 10 is a plot of current-induced magnetic-switching curves in the W(6.0 nm)/Co.sub.40Fe.sub.40B.sub.20 (1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm) sample, under an in-plane magnetic field B.sub.ext of +/−2 mT (+ parallel, and − antiparallel to current direction). Magnetic switching in the PMA of the Co.sub.40Fe.sub.40B.sub.20 layer was sensed by measuring the anomalous Hall voltage in the layer as current was applied to the Hall bar sample.

DETAILED DESCRIPTION

[0057] A magnetoresistive stack which includes a pinned layer, a non-magnetic spacer layer, and a magnetic free layer contacting a non-magnetic base layer with GSHE was developed by Buhrman et al., U.S. patent application publication 2014/0169088. The pinned layer and the free layer of the stack of Buhrman are ferromagnetic. Ibid. The base layer of the stack of Buhrman has a thickness no greater than about five times the spin diffusion length of the base layer, and more preferably has a thickness of about 1.5 times to 3 times the spin diffusion length. Ibid.

[0058] Certain embodiments herein provide a stacked structure containing a non-magnetic layer, a free layer which is ferromagnetic, and an oxide barrier layer. The non-magnetic layer of the structures herein was a metastable β-W thin film. The stacked structure is illustrated in FIG. 1. The as-deposited W thin films were in the β-W phase for thicknesses from 3.0 to 26.7 nm. The β-W remains intact below a critical thickness (t.sub.c) of 22.1 nm, even after magnetic thermal annealing at 280° C., which induces PMA in the stacked structure. Results of x-ray diffraction data herein are evidence of β-W structure after magnetic thermal annealing for the range of thicknesses from about 3.0 nm to about 22.1 nm. See, FIG. 6 and FIG. 7A. β-W with a thickness greater than the spin diffusion length (3.5 nm to 4.3 nm) and less than 22.1 nm was observed to remain in β-form after annealing. The crystalline grain size of the annealed β-W thin films is larger than the spin diffusion length (3.5 nm to 4.3 nm). See, FIG. 7B. The resistivity of annealed β-W was observed to be nearly constant at a range of thicknesses and at temperatures in a range of 10 K and 300 K. See, FIG. 8A. The relatively constant resistivity over such a wide temperature range and thickness range is desirable for spintronics application, since the magnetic switching power from the GSHE would be less sensitive to temperature.

[0059] Methods and data herein provide strategies to obtain giant spin Hall effect and STT induced switching in a layered stack of W/CoFeB/MgO with perpendicular magnetic anisotropy. Further, the data from examples herein show that giant spin Hall effect in β-W resulted in reduced the critical current density for STT induced magnetic switching in the free layer, CoFeB. An MRAM or SLD with PMA results in lower power consumption, higher reliability, higher durability, and reduced or eliminated volatility compared to the characteristics of earlier generations of MRAM.

[0060] Robust PMA was observed in the Examples herein using a structure obtained by a simple fabrication technique, a β-W/CoFeB/MgO layered structure with W having a thickness in a range of 3.0 nm-22.1 nm without an insertion layer between W and CoFeB. As a result of GSHE in β-W, the critical current density for the STT induced magnetic switching in CoFeB was observed to have been reduced in the Examples herein. Therefore, the elemental β-W is a strong candidate for magnetic memory and spin-logic applications.

Example 1—Analysis of the Interface Between a GSHE Solid and a Ferromagnetic (FM) Thin Film

[0061] The layered structures herein contain an interface between a GSHE solid and a FM thin film with PMA, commonly referred to as a free layer. The injected spin current from the GSHE solid yielded STT inside the free layer to effect a magnetization switching. See, Slonczewski et al., Magn. Mater. 159, L1 (1996); Katine et al., Appl. Phys. Lett. 76, 354 (2000). The magnetic states representing the digital bits are sensed by an integrated magnetic sensor, such as an anomalous Hall effect sensor, a magnetic tunneling junction (MTJ), or a giant magnetoresistive (GMR) element. See, Liu et al., Science 336, 555 (2012); Cubukcu et al, Appl. Phys. Lett., 104, 042406 (2014). STT-MRAM with PMA has a number of advantages, including low power consumption, high reliability, durability, and data non-volatility compared to earlier generations of MRAM.

[0062] Among the limited number of GSHE solids previously discovered, metastable β-W was observed to have high resistivity. The large SOC of β-W was observed to be useful in various applications. The spin Hall angle of β-W was observed to be 0.3, which is considered large among transition metals. See, Pai et al., Appl. Phys. Lett., 104, 082407 (2014). The preparation of β-W thin films is compatible with modern semiconductor fabrication processes.

[0063] The range of thicknesses of β-W was extended to 26.7 nm for preparation of the structure herein to observe variation of spin Hall angle over a broad range of β-W thickness. FIG. 4A shows that spin Hall angle as a function of thickness between 3.0 nm and 7.0 nm. FIG. 4B shows the spin Hall angle as a function of thickness between 3 nm and 22.1 nm. From this analysis, the bulk-limit spin Hall angle was observed to be at least 0.49, which is the largest among transition metals, and the spin diffusion length was observed to be in the range of 3.1 nm to 4.3 nm for β-W. Both parameters are important for understanding the advantages of β-W in the context of GSHE, and for development of STT-MRAM and spin-logic devices incorporating β-W and FM thin films with PMA. Results herein show that the amount of STT-induced switching current for magnetization reversal was about 10.sup.6 A/cm.sup.2, one order of magnitude smaller than what was previously observed with similar structures. See, Liu, Science 336, 555 (2012); Pai et al., Appl. Phys. Lett. 104, 082407 (2014).

[0064] GSHE was observed in the β-W/CoFeB/MgO system herein with PMA. The spin Hall angle was observed to be 0.49±0.02 and spin diffusion length was observed to be 4.3±0.4 nm in bulk β-W film at room temperature. This was the largest spin Hall angle that had been observed among elemental solids with a large spin-orbit coupling. See, Azevedo et al., Phys. Rev. B 83, 144402 (2011); Vlietstra et al., Appl. Phys. Lett. 103, 032401 (2013); Liu et al., Phys. Rev. Lett. 106, 036601 (2011); Lee et al., Phys. Rev. B 89, 024418 (2014); Ganguly et al., Appl. Phys. Lett. 104, 072405 (2014); Liu et al., Science 336, 555. (2012); Pai et al., Appl. Phys. Lett. 104, 082407 (2014); Pai et al., Appl. Phys. Lett. 101, 122404 (2012). Furthermore, PMA, which is critical for spintronics applications has not previously been obtained in β-W/CoFeB/MgO.

Example 2—Preparation of Layered Structures

[0065] The multilayered structures or compositions (stacks) herein were prepared on thermally oxidized Si wafers using a high vacuum magnetron sputtering system. During preparation, the base pressure was set at less than 2×10.sup.−8 Torr and the Ar sputtering pressure was set at about 2.2 mTorr. For each sample, the stack was sequentially deposited in the order of W/CoFeB/MgO/Ta. The Ta layer which had a thickness of 1 nm was a capping layer or overlayer to prevent oxidation of the active layers by the atmosphere.

[0066] The DC sputtering power for the CoFeB layer was 10 W. The thickness of the Co.sub.40Fe.sub.40B.sub.20 layer was fixed at 1 nm to allow CoFeB to develop PMA. For formation of β-W, a low DC sputtering power of 3 W was applied intermittently to keep the deposition rate less than 0.02 nm/s. Multiple stacks were prepared, each with W at a thickness in the range of the 2.5 to 26.7 nm. These stacks were patterned using photolithography into standard Hall bars with areas of 20×55 μm.sup.2 for measurements of Hall Effect and resistivity. The stacks were annealed at 280° C. for one minute with two hours of ramping up and six hours of natural cooling in a vacuum at 1×10.sup.−6 Torr, and a magnetic field of 0.45 Tesla was applied perpendicular to the plane of the stacks. Magnetotransport measurements were performed on these stacks using an electromagnet at room temperature. The preparation parameters above can be further optimized. For example, the sputtering power is increased to reduce thin film deposition time. The annealing time including ramping up/down and duration at the fixed temperature can be reduced to increase the throughput of the annealing process. The magnetic field can be selected in a range including 0.45 Tesla, and the annealing temperature can be selected from a range including about 280° C.

[0067] The Quantum Design® Physical Property Measurement System (PPMS) was used to measure the saturated magnetization (M.sub.s) of the stacks at room temperature. M.sub.s for CoFeB at a thickness of 1 nm in the stacks was observed to be about 1100 emu/cm.sup.3.

[0068] To identify the phase of the W layer, sheet resistance (R.sub.□) of the stacks W(thickness, t)/CoFeB(1.0)/MgO(1.6)/Ta(1.0) was measured. FIG. 1 is a graph of the value of R.sub.□ as a function of W thickness in the range of 2.5 nm to 7.0 nm. The solid line is the best fit to the data based on resistivities of ρ.sub.W which was observed to be about 209 μΩ-cm and ρ.sub.FeCoB which was observed to be about 80 μΩ-cm. Typical resistivities for the stable α-W phase and the metastable β-W phase are less than 40 μΩ-cm and greater than 150 μΩ-cm, respectively. These high resistivity values indicate that β-W was present in the samples herein.

[0069] Conditions for the magnetotransport measurement are illustrated in the schematic of FIG. 2A. A DC current was sent along the Y-axis of a Hall-bar sample, and the Hall voltage along the X-axis was measured. An external magnetic field (B.sub.ext) was applied in the YZ plane with an angle β between the field and Y-axis. The resulting magnetization vector (M) was observed in the YZ plane at an angle θ from the Y-axis.

[0070] FIG. 2B is a graph of the anomalous Hall resistance (R.sub.H) as a function of magnetic field applied perpendicularly to the sample plane (β=90°), W(t)/CoFeB(1.0)/MgO(1.6)/Ta(1.0), for a series of samples with W thicknesses in the range of 3 nm to 7 nm. The square hysteresis loops for each sample in FIG. 2B revealed PMA in the W/CoFeB without an insertion layer between W and CoFeB. The switching field, or coercivity H.sub.c, was varied from 5 Oe to 22 Oe. The square-shaped loops for each W thickness indicate that CoFeB exhibited PMA for each sample at each W thickness. PMA in this structure has not previously been observed in the stacked structure.

Example 3—Analysis of Magnetization Switching

[0071] The anomalous Hall Effect (AHE) provides a sensing mechanism to measure the magnetization state of the CoFeB layer. This example illustrates the response of the magnetization state of the CoFeB layer to excitation current (I) in the W layer and B.sub.ext. FIG. 2C and FIG. 2D are graphs of current-induced magnetic switching behavior of a representative sample, W(7.0)/CoFeB(1.0), under a series of positive (β=0°) or negative (β=180°) in-plane fields at values of B.sub.ext equal to 0.2 mT, 0.4 mT, 0.7 mT, and 2 mT. The bi-stable states of M (up and down) are accessible by cycling current in both directions through a critical value (I.sub.c), under either a positive or negative B.sub.ext. See, FIG. 2C and FIG. 2D. Critical current (I.sub.c) was defined as the average of the positive and negative switching current. For W(7.0 nm)/CoFeB(1.0 nm), I.sub.c was observed to be about 3.1 mA, corresponding to a critical current density (J.sub.c) of about 1.5×10.sup.6 A/cm.sup.2 in the W layer under an in-plane field of 2 mT. See, FIG. 2C. This value is about one order of magnitude smaller than those obtained in other PMA structures: Ta/CoFeB, Pt/Co, and W/Hf/CoFeB. See, Liu et al., Science 336, 555 (2012); Pai et al., Appl. Phys. Lett. 104, 082407 (2014); Zhang et al., J. Appl. Phys. 115, 17C714 (2014); Qiu et al., Scientific Reports 4, 4491 (2014); Liu et al., Phys. Rev. Lett. 109, 096602 (2012).

[0072] The current-induced magnetic switching in PMA structures results from the STT mechanism due to the injected spin current density (J.sub.S) from the SOC solid with GSHE. See, Liu et al., Phys. Rev. Lett. 109, 096602 (2012). Under the measurement conditions of FIG. 2A, the equilibrium orientation (θ) of M in the PMA CoFeB layer was calculated from the condition that the net torque on M is zero, using the following equation:

τ.sub.tot≡{circumflex over (x)}.Math.({right arrow over (τ)}.sub.ST+{right arrow over (τ)}.sub.ext+{right arrow over (τ)}.sub.an)=τ.sub.ST.sup.0+B.sub.ext sin(θ−β)−B.sub.an.sup.0 sin θ cos θ=0  (1),

[0073] where

[00001]${\tau }_{\mathrm{ST}}^0=\frac{\hslash }{2.Math.{\mathrm{eM}}_s.Math.t}.Math.J_S$

is the torque per unit moment, and B.sub.an.sup.0 is the perpendicular anisotropy field. See, Liu et al., Phys. Rev. Lett. 109, 096602 (2012). This macrospin model was used to predict the current-induced magnetic switching in FIG. 2C, at sufficient current density greater than J.sub.c, or corresponding STT, τ.sub.ST.sup.0 per unit moment. In FIG. 2C, the lowest J.sub.C of about 1.5×10.sup.6 A/cm.sup.2 was observed at 2 mT, an amount which was about 5 to 10 times less than that reported for other PMA systems. See, Liu et al., Science 336, 555 (2012); Pai et al., Appl. Phys. Lett. 104, 082407 (2014); Zhang, J. Appl. Phys. 115, 17C714 (2014); Qiu et al., Scientific Reports 4, 4491 (2014); Liu et al., Phys. Rev. Lett. 109, 096602 (2012).

[0074] In the coherent spin rotation regime, Eq. (1) was also used as a method to measure τ.sub.ST.sup.0, hence, the converted spin current density J.sub.S from which the spin Hall angle can be derived using θ=J.sub.s/J.sub.c. See, Liu et al., Phys. Rev. Lett. 109, 096602 (2012). In Eq. (1), the spin Hall angle (θ) was obtained from the anomalous Hall resistance, R.sub.H/R.sub.0=sin θ, where R.sub.0 is the maximum Hall resistance when M is perpendicular to the sample plane.

[0075] According to Eq.(1), as B.sub.ext approaches zero or infinity, θ reaches 0 or 90°, respectively. Under an intermediate B.sub.ext, and a positive or a negative current of 1 mA, θ was observed to be dependent on τ.sub.ST.sup.0, B.sub.ext, and B.sup.0.sub.an which illustrates how the relationship between the R.sub.H and sin θ varies as a function of B.sub.ext. See, FIG. 3A. At an arbitrary sin θ, two B.sub.ext values exist, B.sub.+(θ) and B.sub.−(θ), corresponding to the positive and negative current, respectively. See, FIG. 3A. From Eq.(1),

τ.sub.ST.sup.0(+J.sub.S)+B.sub.+(θ)sin(θ−β)−B.sub.an.sup.0 sin θ cos θ=0  (2)

τ.sub.ST.sup.0(−J.sub.S)+B.sub.−(θ)sin(θ−β)−B.sub.an.sup.0 sin θ cos θ=0  (3)

[0076] By solving the simultaneous equations using combination of Eqs. (2)±(3), the following are obtained:

[B.sub.+(θ)−B.sub.−(θ)]=Δτ.sub.ST.sup.0/sin(θ−β)  (4)

[B.sub.+(θ)+B.sub.−(θ)]2B.sub.an.sup.0 sin θ cos θ/sin(θ−β)  (5),

where Δτ.sub.ST.sup.0=τ.sub.ST.sup.0(+J.sub.S)−τ.sub.ST.sup.0(−J.sub.S)=2τ.sub.ST.sup.0(|J.sub.S|. FIG. 3A is a graph of the quantities of B.sub.+(θ), B.sub.−(θ), and θ for W(6.5 nm)/CoFe(1.0 nm)/MgO(1.6 nm). Using Eq.(4) and (5), τ.sub.ST.sup.0(|J.sub.S|) and B.sub.an.sup.0 were calculated.

[0077] FIG. 3B, a graph of [B.sub.+(θ)−B.sub.−(θ)] as a function of 1/sin(θ−β), uses the data obtained in FIG. 3A. As predicted by Eq.(4), linear relations were observed at various currents, and the slope was Δτ.sub.ST.sup.0, for each current. See, FIG. 3B. Using this method, the STT, τ.sub.ST.sup.0(|J.sub.S|) as a function of current was determined for samples with variable W thicknesses. See, FIG. 3C. Using the data in FIG. 3C, the spin Hall angle was calculated according to

[00002]$\theta =J_s/J_c=\left(\frac{2.Math.{\mathrm{eM}}_s.Math.t}{\hslash }\right).Math.\left({\tau }_{\mathrm{ST}}^0/J_c\right).$

[0078] FIG. 4A shows the relationship of the spin Hall angle as a function of W thickness (t) in the range of 3 nm to 7 nm for W(O/CoFeB(1.0)/MgO(1.6)/Ta(1.0) stacked structure with PMA. At 7.0 nm, the spin Hall angle was observed to be θ(7 nm)=0.33±0.03. At the thinner limit, spin Hall angle was observed to decrease, as the W thickness approached the spin diffusion length (λ.sub.sf). See, FIG. 4A and FIG. 4B. Variation of the Hall angle versus W thickness was evidence that θ(∞)=0.49±0.02 and λ.sub.sf=4.3±0.3 nm in the bulk β-W film, according to

[00003]$\frac{J_s\left(t\right)}{\mathrm{Js}\left(\infty \right)}=\frac{\Theta \left(t\right)}{\Theta \left(\infty \right)}=1-\mathrm{sech}\left(\frac{t}{{\lambda }_{\mathrm{sf}}}\right)$

which was used to fit the data in FIG. 4A and FIG. 4B. See also, Liu et al., Phys. Rev. Lett. 106, 036601 (2011). In comparison, θ(5.2 nm)=0.33±0.06 was observed for W/CoFeB/MgO with in-plane magnetic anisotropy, and effective θ(4 nm)=0.34±0.05 was observed in W/Hf/CoFeB/MgO with Hf-induced PMA. In certain embodiments, an insertion layer, Hf, is not needed between W and CoFeB to obtain PMA. See, FIG. 2A.

[0079] Determination of the bulk θ(∞) and λ.sub.sf for β-W is beneficial to further theoretical understanding of this SOC solid and to designing spintronic devices by selecting appropriate β-W thickness. Moreover, PMA was achieved in W/CoFeB/MgO without an insertion layer which is known to reduce spin Hall angle. See, Pai et al., Appl. Phys. Lett. 104, 082407 (2014).

[0080] As shown in FIG. 2C, with sufficient current, for example STT, M of the CoFeB layer undergoes the switching as described by Eq.(1). The largest Hall angle in the sample series herein was observed in the magnetic switching phase diagram for a representative sample W(7.0)/CoFeB(1.0)/MgO(1.6), each sample having an area of 20×55 μm.sup.2. See, FIG. 5. The critical switching current density (J.sub.C) in the W layer was observed to decrease rapidly and linearly with increasing field (B.sub.ext) up to a characteristic field B.sub.0 of about 10 Oe, and at a slower rate when B.sub.ext was greater than B.sub.0. See, FIG. 5. As a comparison, B.sub.0 of about 150 Oe and 3000 Oe were observed for two other PMA systems: Ta(5 nm)/CoFe(0.6 nm)/MgO(1.8 nm) having an area of 1.2×15 μm.sup.2 and Pt(2 nm)/Co(0.6 nm)/AlO.sub.x having an area of 20×200 μm.sup.2, respectively. See, Liu et al., Phys. Rev. Lett. 109, 096602 (2012); Perez et al., Appl. Phys. Lett. 104, 092403 (2014).

[0081] The lower B.sub.0 observed in this PMA system led to reliable switching under a small external field. In STT-MRAM or spin-logic applications, a low biasing field is more easily implemented than a ten-time larger field. B.sub.0 was observed to have the magnitude of the nucleation field in the stacks herein (see coercivity values in FIG. 2B).

[0082] The magnetic switching phase diagram of the PMA CoFeB driven by STT from the β-W was analyzed. Under an in-plane biasing field of only 20 Oe, the switching current density was observed to be about 1.6×10.sup.6 A/cm.sup.2 which is the lowest among other PMA systems with GSHE. Thick β-W films integrated with CoFeB ferromagnetic film without an insertion layer were observed to achieve robust PMA. The large Hall angle and acquired PMA makes β-W a candidate for STT-MRAM and spin-logic applications, with the added advantage of its compatibility with modern semiconductor fabrication.

Example 4—Fabrication of Metastable Thin Films Using Sputtering

[0083] Examples herein provide data which are evidence of structure, electron transport, and GSHE of β-W thin films. Thermal annealing of the β-W films resulted in a lower switch current density compared to the β-W films prior to annealing indicating thermal annealing is useful for materials used in spintronic MRAM and SLD. See, Hao et al., Applied Physics Letters 106, 182403 (2015), which is hereby incorporated by reference in its entirety. Results herein show that films having a thickness up to 22.1 nm remained in the β phase through the annealing process, and films with a thickness greater than 22.1 nm were transformed into α phase by the annealing process. Therefore, the thickness of β-W films that were magnetic thermal annealed for use in spintronic devices was observed to have been greater than the spin diffusion length (λ.sub.sf) of 3.5 nm to 4.3 nm and less than 22.1 nm to have maximized GSHE in β-W films.

[0084] Spintronic MRAM and logic processors have exploited the physics of the spin Hall effect for switching bits, and the materials engineering for prior structures remains challenging. The large spin-orbit coupling in β-W yielded a very low critical current density for magnetization switching after annealing. Methods of fabrication of the technologically promising structures of the various embodiments of the invention herein are provided.

[0085] The W films herein were prepared on thermally oxidized Si wafers under ambient conditions using a high vacuum magnetron sputtering system equipped with a cryopump. The fabrication conditions included a base pressure that was less than 2×10.sup.−8 Torr and the Ar sputtering pressure that was about 2.2 mTorr. For the formation of β-W, a low DC sputtering power of 3 W was applied intermittently to keep the deposition rate constant at 0.02 nm/s. The films were patterned into standard Hall bars using photolithography for spin Hall effect and resistivity measurements, with the longitudinal dimensions of 20×55 μm.sup.2 in area. In addition to these as-deposited samples, a corresponding set of samples was prepared that were annealed at 280° C. for 1 minute with 2 hours of ramping up and 6 hours of natural cooling in vacuum (1×10.sup.−6 Torr) and under a magnetic field (0.45 T) perpendicular to the films.

[0086] In certain embodiments, the metastable thin films were fabricated by a method including depositing a metastable thin film onto a substrate using a sputtering system. For example, a β-W thin film was deposited using a sputtering power in the range of from about 3 W to about 1000 W and at a rate in the range of from about 0.02 nm/s to about 10 nm/s. Improved formation of metastable thin films is obtained by depositing the β-W thin film onto a substrate cooled to a temperature less than the ambient temperature, for example, the temperature of liquid nitrogen.

[0087] Methods are provided herein of making the metastable thin films by sputtering a material onto a substrate. In these methods, the metastable thin film was stabilized prior to inclusion in the multilayered structures. The metastable thin film used was generally β-W thin film. The fabrication process was performed using pure argon as the sputtering gas. Improved stabilization of metastable thin films is expected by using a combination of argon and oxygen as the sputtering gas. For example, oxygen content is 2% to 60% of the total sputtering gas in the sputtering process.

[0088] A magnetron sputtering process which generated giant spin Hall effect was used to fabricate β-W thin films. As-deposited thin films were observed to be in the metastable β-W phase at thicknesses from 3.0 nm to 26.7 nm. The β-W phase was observed to have remained intact below a critical thickness of 22.1 nm even after high vacuum magnetic thermal annealing at 280° C., which induces perpendicular magnetic anisotropy in the layered structures herein of β-W/Ca.sub.40Fe.sub.40B.sub.20/MgO. Intensive annealing was observed to have transformed the films with thicknesses greater than 22.1 nm into the stable α-W phase.

Example 5—Analysis of Structure and Grain Size of Thin Films

[0089] The structure and grain size of β-W and α-W thin films were analyzed by x-ray diffraction analysis with the Bruker D8 Discover X-Ray Diffraction (XRD) System. Electron transport in terms of resistivity and normal Hall effect was observed through a temperature range of as low as about 10 K to at least about 300 K on each sample. Low switching current density was observed in β-W/Co.sub.40Fe.sub.40B.sub.20/MgO, a result that indicates that the films are technologically promising for future generations of spintronic magnetic random access memories (MRAM) and spin-logic devices.

[0090] Highly resistive β-W was observed to have strong SOC leading to GSHE in this metal. See, Pai et al., Appl. Phys. Lett. 101, 122404 (2012); Pai et al., Appl. Phys. Lett. 104, 082407 (2014); Hao et al., Phys. Rev. Appl. 3, 034009 (2015). The spin Hall angle of β-W was observed to approach 0.49, which is the largest among transition elements, and β-W was observed to convert charge current into spin current efficiently. See, Hao et al., Phys. Rev. Appl. 3, 034009 (2015), which is hereby incorporated by reference in its entirety. Robust PMA was observed in a layer structure combining metastable β-W and Ca.sub.40Fe.sub.40B.sub.20 thin film. Ibid. The GSHE yielded, after suitable thermal magnetic annealing, a low critical current density for magnetization switching. Ibid.

[0091] Future generations of spintronic MRAM and spin-logic devices will increasingly rely on GSHE in β-W films. See, Liu et al., Science 336, 555 (2012); Datta et al., Appl. Phys. Lett. 101, 252411 (2012).

[0092] A series of β-W thin films with a broad range of thickness (3.0-26.7 nm) were prepared using a magnetron sputtering process. Based on this process, PMA and a large spin Hall angle (0.49) in the bulk β-W in a structure of β-W/Co.sub.40Fe.sub.40B.sub.20/MgO were achieved. See, Hao et al., Phys. Rev. Appl. 3, 034009 (2015). The effect of thermal annealing on the structure of the W films was examined. See, Hao et al., Appl. Phys. Lett. 106, 182403 (2015).

[0093] FIG. 6 is a recording of the XRD patterns for as-deposited and annealed W films, each with a thickness up to 26.7 nm. The as-deposited films had single phase β-W, which is an A.sub.3B solid with the A15 type crystal structure. The annealed films remained in β-W phase up to a critical thickness (t.sub.c) of 22.1 nm. Above 22.1 nm, films were transformed into α-W phase, which had the bcc crystal structure.

[0094] FIG. 7A is a graph of the lattice constant as a function of W film thickness, which revealed the effect of annealing. Up to a thickness of 26.7 nm, the lattice constant of the as-deposited β-W films were observed to be unchanged. See, FIG. 7A. Post-annealing and beyond t.sub.c, the lattice constant becomes that of the α structure. Ibid. Therefore, the thickness of β-W which has been magnetically thermally annealed and has GSHE intended for spintronics devices, is smaller than t.sub.c. The spin diffusion length (λ.sub.sf) in β-W film is in a range of 3.5 nm to 4.3 nm, which is also smaller than t.sub.c. See, Hao et al., Phys. Rev. Appl. 3, 034009 (2015). Therefore, the full strength of the GSHE of β-W film is exploited at a thickness that is less than t.sub.c, and is larger than λ.sub.sf.

[0095] Based on XRD data, the grain size or size of crystallites was obtained using the Scherrer equation as a function of W film thickness for as-deposited and annealed samples. See, FIG. 7B. According to the Scherrer equation, which relates the grain size to the broadening of a diffraction peak, the average grain size was observed by Kλ/β cos θ, where K is a shape factor (typically about 0.9), λ is the x-ray wavelength, β is the width of a diffraction line at half maximum intensity, and θ is the Bragg angle. The Scherrer equation was observed to have remained valid, to the extent that the peak broadening is primarily due to grain size rather than other inhomogeneities. The grain size was observed to be smaller than, and increased with the film thickness. Annealing was observed to increase the grain size by about 50% to 70%. The grain size in all samples was larger than the λ.sub.sf, which is 3.5 nm in β-W, by a factor of 2 to 4. See, Hao et al., Phys. Rev. Appl. 3, 034009 (2015). Therefore, λ.sub.sf was observed not to be predominantly affected by the grain size, and was more dependent on the local structures within the β-W crystallites, such as atomic and lattice disorders.

Example 6—Analyzing Electron Transport Properties of the Films

[0096] FIG. 8A is a graph of the resistivities of the W films as a function of temperature in the range of 5 K to 380 K. The as-deposited β-W films (3.0 to 24.1 nm) were observed to be characterized by very large resistivities of 185 to 210μΩ-cm at 300 K. See, FIG. 8A. The temperature coefficient of resistivity, (1/ρ)Δρ/ΔT, was observed to be small for the β-W films through the temperature range of 5 K to 380 K. For example, for the 14.5 nm-thick β-W film, resistivity was observed to have remained nearly constant at 188μΩ-cm from 5 K to 380 K. Ibid. This property is advantageous if β-W is used to generate spin current based on the GSHE because the magnetic switching power will not depend on temperature, facilitating a wide temperature range of operations for spintronic devices.

[0097] For a continuous thin film with a thickness (t) larger than the effective electron mean free path (λ.sub.eff), the thin film resistivity is expressed as:

ρ(t)=ρ.sub.B+⅜ρ.sub.Bλ.sub.eff/t,  (6)

where ρ.sub.B is the bulk resistivity.

[0098] FIG. 8B is a graph of the resistivities of the β-W thin films as a function of the inverse thickness (1/t). In the range of 3.0 to 22.1 nm, Eq.(1) fits the observed resistivity data. From the fit, ρ.sub.B(300 K) was observed to be about 195 μΩ-cm, λ.sub.eff(300 K) was observed to be about 0.45 nm, ρ.sub.B(10 K) was observed to be about 192 μΩ-cm, and λ.sub.eff (10 K) was observed to be about 0.95 nm for the bulk β-W film. The thermally induced resistivity, Δρ=ρ.sub.B(300 K)−ρ.sub.B(10 K) was observed to be about 3μΩ-cm, which indicates that electron-phonon inelastic scattering was relatively weak compared to disordered elastic scattering.

[0099] The value of λ.sub.eff was observed to be smaller than the thickness and the grain size of the film by a factor of 5 to 10. β-W phase is stabilized by small amounts of oxygen. See, Petroff et al., J. Appl. Phys. 44, 2545 (1973); Narasimham, AIP Advances 4, 117139 (2014). Therefore, the finite-size effect and the grain boundary scattering are not sufficient to account for the large resistivity of the β-W film. Further, the small amounts of oxygen which are responsible for electron disordered scattering are associated either with charge-dependent impurity scattering or spin-orbit scattering. Based on the large spin Hall angle observed in β-W films, the disordered spin-orbit scattering plays a significant role in the resistivity of the β-W films. See, Pai et al., Appl. Phys. Lett. 101, 122404 (2012); Pai et al., Appl. Phys. Lett. 104, 082407 (2014); Hao et al., Phys. Rev. Appl. 3, 034009 (2015).

[0100] Above 22.1 nm, as shown in FIG. 8B, the resistivities were observed to have decreased significantly as calculated by Eq.(1). These thicker films have a small mixture of α-W phase, which has a smaller resistivity than β-W phase. For the annealed films, resistivity was observed to be in the range of 100-260 μΩ-cm for films that remained β-W phase.

Example 7—Analysis of Temperature Dependence of Resistivity and Hall Effect

[0101] An investigation of the temperature dependence of resistivity and Hall effect provided insight into the β-W solid with a strong SOC. The Quantum Design Physical Property Measurement System was used to measure resistivity and Hall effect as functions of temperature between about 10 K and about 360 K.

[0102] FIG. 8A is a graph of the temperature dependence of the Hall effect in β-W and α-W thin films. After annealing, the 24.1 nm and 26.7 nm films were observed to have been transformed into α-W phase. For the 26.7 nm thick α-W thin film, ρ(300 K) was observed to have been about 40 μQ-cm and ρ(10 K) was observed to be about 30 μΩ-cm. See, FIG. 8A. The resistivity of the α-W phase was observed to be more dependent on temperature compared with β-W. For example, the thermally induced resistivity, Δρ=ρ.sub.B(300 K)−ρ.sub.B(10 K) was observed to be 10μΩ-cm for α-W film with a thickness of 26.7 nm, indicating a significant contribution from the electron-phonon scattering.

[0103] The normal Hall effect was measured which is an integral part of the electron transport of a metal. An investigation of both the resistivity and Hall effect and their temperature dependence provides insight into the β-W solid with a strong SOC. FIG. 9 shows the temperature dependence of the Hall effect of the β-W and α-W thin films. The Hall coefficient, R.sub.H(300 K), for β-W thin film was observed to be 1.62×10.sup.−8 Ω-cm/T which is equal to 1.62×10.sup.−10 m.sup.3/C. R.sub.H(T), carries a negative sign and is linearly dependent on temperature in the range of 10 K to 300 K. At 10 K, the magnitude of R.sub.H(10 K) was observed to be −1.14×10.sup.−8 Ω-cm/T, and which was observed to have been reduced by 30% from the room temperature value. Charge carriers are predominantly electrons in the β-W thin film. R.sub.H was observed to be about 5 times smaller than the bulk value; for example R.sub.H(300 K) was observed to be +8.6×10.sup.−11 m.sup.3/C compared to a bulk value of +11.8×10.sup.−11 m.sup.3/C. See, Physical and Chemical Properties of the Elements, edited by G. V. Samsonov (Naukova Dumka, Kiev, 1965); AIP Handbook, 3rd Ed., edited by Dwight E. Gray (American Institute of Physics, New York, 1972).

[0104] R.sub.H(T) for α-W thin films was observed to be linearly dependent on temperature. For α-W thin films, R.sub.H(300 K) was observed to be −1.91×10.sup.−9 Ω-cm/T which is equal to −1.91×10.sup.−11 m.sup.3/C, consistent with prior studies of α-W thin films. See, Bastl, Thin Solid Films 10, 311-313 (1972). R.sub.H(T) was observed to change signs from negative to positive as temperature was reduced below 130 K, indicating a competition between electron and hole carriers from multi-bands in α-W thin films. The spin Hall effect data presented in FIG. 9 is evidence of the effects of band structure, scattering mechanisms, and surface electronic states of β-W and α-W thin films incorporating the strong SOC. Small amounts of oxygen in β-W or α-W thin films were observed to affect temperature dependence of the spin Hall effect.

[0105] Using sputtering methods herein, a layered structure in the form of W(t)/Co.sub.40Fe.sub.40B.sub.20 (1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm) was made. After annealing, Co.sub.40Fe.sub.40B.sub.20 was observed to develop PMA as the W layer remained in β-phase. See, Hao et al., Phys. Rev. Appl. 3, 034009 (2015).

[0106] FIG. 10 is a graph of the magnetic switching of the Co.sub.40Fe.sub.40B.sub.20 (1.0 nm) layer by the spin current generated in the β-W layer (6.0 nm). Switching was detected by measuring the anomalous Hall effect of the Co.sub.40Fe.sub.40B.sub.20 (1.0 nm) layer. Under an in-plane magnetic field of 2 mT, the sample W(6.0 nm)/Co.sub.40Fe.sub.40B.sub.20 (1.0 nm)/MgO(1.6 nm)/Ta(1.0 nm) was observed in FIG. 10 to be magnetically switched with a critical current (I.sub.c) of 2.2 mA, corresponding to a critical current density (J.sub.C) of 1.2×10.sup.6 A/cm.sup.2 in the W layer, about an order of magnitude smaller than those obtained in other similar structures. See, Pai et al., Appl. Phys. Lett. 104, 082407 (2014); Liu et al., Science 336, 555 (2012); Zhang et al., J. Appl. Phys. 115, 17C714 (2014); Qiu et al., Scientific Reports 4, 4491 (2014); Liu et al., Phys. Rev. Lett. 109, 096602 (2012).

[0107] In general, β-W thin films that are embodiments of the invention provided herein were fabricated using a magnetron sputtering process. The as-deposited films were observed to be in β-W phase from 3.0 nm to 26.7 nm. In order to obtain PMA in layered structures of β-W/Co.sub.40Fe.sub.40B.sub.20/MgO, high vacuum thermal magnetic annealing was performed at a temperature of 280° C. After annealing, β-W phase was observed to have been transformed into α-W phase at a critical thickness of greater than 22.1 nm, which is larger than the spin diffusion length of 3.5 nm for β-W thin films. The stabilization of β-W up to 22.1 nm after annealing was observed to be capable of exploiting the full GSHE in β-W thin films. In the films provided herein, the typical grain size was observed to be about one third to one half of the thin film thickness. The room temperature resistivity of the bulk β-W film was observed to be about 195 μΩ-cm and the electron mean free path was observed to be about 0.45 nm.

[0108] The resistivity of β-W was observed to be nearly insensitive to factors of temperature and thickness. At 14.5 nm, the temperature coefficient of resistivity was observed to be almost zero in the range of 10 K to 360 K. This property is highly desirable for spintronics applications because the magnetic switching power from GSHE does not depend on temperature. The normal Hall coefficient for β-W thin films was measured between 10 K and 300 K. The Hall coefficient for the β-W thin films was observed to have the same magnitude as bulk α-W, and have a negative sign. Using the sputtering and magnetic thermally annealing process herein, PMA and low switching current density was observed in a layered structure of β-W/CO.sub.40Fe.sub.40B.sub.20/MgO.