We demonstrate the spin–orbit torque (SOT) induced perpendicular magnetization switching in an annealed W/CoFeB/Zr/MgO multilayer with high thermal stability. It is found that the thermal stability factor can reach 79 after annealing at 540 °C. With an increase in the annealing temperature, the absolute damping-like efficiency almost keeps a high constant value (about 0.3). The tungsten in the W/CoFeB/Zr/MgO multilayer could convert from the high resistive β-W to a mediate resistive amorphous-like structure. Therefore, the absolute spin Hall conductance increases from 765 of β-W to 1420 (ℏ/e)(Ω cm)−1 of the amorphous-like tungsten. These results pave a realistic way for the practical application of tungsten in the SOT-based spintronics devices with high thermal stability and SOT efficiency.

Spin–orbit torque (SOT) induced magnetization switching in an energy-efficient and fast way has attracted considerable theoretical and experimental interest in the past decade.1–12 In a heavy metal (HM) buffered CoFeB/MgO multilayer with perpendicular magnetic anisotropy (PMA),13 the spin angular momentum transfer from the HM to the ferromagnetic metal (FM) by the large spin Hall effect (SHE) then results in the current-induced magnetization switching (CIMS).1,2 The effective SOT fields are usually proportional to the spin Hall angle (θSH) of the HM, which is the ratio between spin current density and charge current density. Among all the HMs, β-phase tungsten (β-W) is reported to possess the largest spin Hall angle in the HMs.4,14 However, the θSH of W strongly depends on its crystalline structure4,14,15 and the condition during the growth of W layer.15–19 These results open an intriguing question on the crystalline structure dependence of intrinsic spin Hall conductance (SHC) for W and other HMs.

Furthermore, the PMA related SOT driven devices also need to satisfy the following requirements to fulfill the spin logic device applications:20–23 (i) the critical switching current density Jc should be low to ensure low power consumption; (ii) the thermal stability should be high to guarantee data retention; and (iii) the stack should sustain significant processing thermal budgets (∼400 °C) to be compatible with the back-end-of-line processing in a complementary metal–oxide–semiconductor (CMOS) integration technique.21,22 Generally, the thermal stability of a magnetic device is determined by the energy barrier (Eb) separating the two stable orientations of its magnetization. The thermal stability factor Δ (=Eb/kBT, where kB is the Boltzmann constant and T = 300 K) larger than 55 is needed to guarantee data retention for 10 years.22 Recently, we have reported that the PMA in the W/CoFeB/Zr/MgO multilayer by inserting a very thin Zr layer (about 0.43 nm) is robust upon annealing at 600 °C,23 which can well meet the requirement for integrating SOT devices with CMOS.

In this work, we demonstrate the CIMS in the annealed W/CoFeB/Zr/MgO multilayer with a thin Zr insertion layer. The thermal stability factor can reach 79 after the sample is annealed at 540 °C. Accompanied with the annealing temperature (Tann) increasing from 270 to 540 °C, the structure of the W layer changes from the high resistive β-W to a mediate resistive amorphous-like structure, while the absolute SHC increases from 765 of β-W to 1420 (ℏ/e)(Ω cm)−1 of the mediate resistive amorphous-like tungsten. This work provides a scheme to sustain the large SOT efficiency and high thermal stability, which will be beneficial for SOT devices to combine with the mature CMOS integrated techniques.

Multilayer films with a structure of W(tW)/CoFeB(1.2)/Zr(0.43)/MgO(3) (nominal thickness in nanometers) were deposited on thermally oxidized Si/SiO2 substrates at room temperature using a magnetron sputtering system. The thickness of W layer tW was systematically varied from 2 to 15 nm. The base pressure of the sputtering system was less than 8 × 10−6 Pa. The Zr and CoFeB layers were deposited by using a direct current (DC) power source, and the MgO layer was deposited by using a radio frequency (RF) power source. The deposition rates for W, Zr, CoFeB, and MgO layers were 0.47, 0.54, 0.68, and 0.1 Å/s, respectively. The films were annealed in vacuum, which is less than 1 × 10−5 Pa, at temperature up to 600 °C for 1 h to obtain PMA. More details about the magnetic properties can be found in our previous work.23 

The W/CoFeB/Zr/MgO films were patterned into Hall bars of a width of 10 µm by standard photolithography combined with an argon ion etching technique. To protect the devices, a 2-nm-thick Ta was used as a capping layer. The anomalous Hall resistance and current-induced magnetization switching were measured by using a room temperature multipurpose transport measurement system with a 1 T electromagnet; Keithley 2182A and Keithley 6221 were employed as the source for DC and pulse measurements, respectively. For the harmonic Hall voltage measurement, a low frequency alternating current (133.33 Hz) from Keithley 6221 was passed into the bilayer, and the first and second harmonic Hall voltages were measured using two SR830 lock-in amplifiers while sweeping the magnetic field HL(T) along (perpendicular to) the current direction. For the current-induced magnetization switching measurement, a combination of Keithley 6221 current source and 2182A nanovoltmeter were used. In the measurements of switching phase diagrams, a constant pulse width of 200 µs was used and the RH was measured after a 150 µs delay. In the measurements of switching current vs pulse duration, the pulse widths were from 50 µs to 1 s.

The magnetic properties were measured using a vibration sample magnetometer. High-resolution transmission-electron microscopy (HRTEM) images were collected using a FEI Tecnai microscope operating at an accelerating voltage of 300 kV. The Gatan DigitalMicrograph software was used to measure the lattice spacings from the fast Fourier transformation (FFT) pattern of the HRTEM images. These values, and the corresponding FFT diffraction patterns, were simulated using the CrystalMaker/SingleCrystal software package. The polarized neutron reflectometry (PNR) measurement was carried out on the Multipurpose Reflectometer (MR) at the China Spallation Neutron Source (CSNS). The neutron reflectivity curves are recorded at room temperature as the function of momentum transfer Q = 4π sin θ/λ, where λ is the neutron wavelength and θ is the incident angle between the neutron beam and the film plane. To access a broad momentum transfer range, the reflected neutrons were collected at different incident angles.

Figure 1 shows the inverse of the sheet resistance, l/(Rxx · w), against the W thickness under different annealing temperatures, where l and w are the length and width of the Hall bar, respectively. Since the sheet resistance inverse shows a jump at tW ∼ 8 nm, the tW dependence of l/(Rxx · w) can be roughly fitted by two straight lines for tW < 8 nm and tW > 8 nm, respectively. The resistivity of thick W (ρW) is determined from the slope of the fitted line, which is about 17 µΩ·cm (the value of α-W) when tW ≥ 8 nm. In general, the longitudinal resistivity of an individual layer in a multilayer can be determined by the channel resistance of the Hall bar versus thin film thickness measurements, which have been well used in the previous studies.24–26 As shown in Table I, an interesting feature is that the ρW of 4.5-nm-thick W layer after annealing at 540 °C is about 54% smaller than that of the as-deposited film. Meanwhile, the resistivity of the thick W layer changes a little. The x-ray diffraction measurements indicate that the annealed thick W film is α-phase (see the supplementary material). Apparently, the crystalline structure of W and its resistivity have strong thickness dependence, which are in agreement with previous studies.4,14,15 In the following, we focus on the samples with tW = 4.5 nm, hereafter called W/CoFeB/Zr/MgO, to systematically investigate the thermal stability of SOT performance.

FIG. 1.

The W thickness dependence of l/(Rxx · w) in W(tW)/CoFeB/Zr/MgO. (a) As-prepared, (b) annealed at 270 °C, and (c) annealed at 540 °C. The solid lines represent the linear fitting to the data for appropriate thickness ranges.

FIG. 1.

The W thickness dependence of l/(Rxx · w) in W(tW)/CoFeB/Zr/MgO. (a) As-prepared, (b) annealed at 270 °C, and (c) annealed at 540 °C. The solid lines represent the linear fitting to the data for appropriate thickness ranges.

Close modal
TABLE I.

Resistivities of W (μΩ·cm) in the W(tW)/CoFeB/Zr/MgO multilayers.

As-deposited270 °C annealed540 °C annealed
tW > 8 nm 17 20 19 
tW = 4.5 nm 163 156 107 
As-deposited270 °C annealed540 °C annealed
tW > 8 nm 17 20 19 
tW = 4.5 nm 163 156 107 

To further understand the structure of W/CoFeB/Zr/MgO multilayer after high temperature annealing, the high-resolution transmission electron microscopy (HRTEM) and polarized neutron reflectometry (PNR) measurements were carried out. Figure 2 shows the HRTEM picture of the W/CoFeB/Zr/MgO multilayers annealed at 540 °C. A significant feature is that the fast Fourier transformation (FFT) pattern in the inset of Fig. 2 has a ring-like pattern, which indicates that the 4.5-nm-thick W layer in the W/CoFeB/Zr/MgO multilayer annealed at 540 °C is amorphous-like. For comparison, the 4.5-nm-thick W layer in the W/CoFeB/Zr/MgO multilayer annealed at 270 °C is mainly composed of metastable β-W.23, Figure 3 shows the elemental mapping data of cross-sectional W/CoFeB/Zr/MgO annealed at 540 °C. Even the sample was annealed after 540 °C, the good uniformity and flatness for each layer can be seen. A continuous coverage of Co can be seen on the W layer when the nominal thickness of CoFeB layer is as thick as 1.2 nm. A clear layered structure of the W/CoFeB/Zr/MgO multilayers annealed at 540 °C was also verified by the PNR analysis (see the supplementary material). Although the 0.43-nm-thick Zr layer turns to Zr oxide,23 some of Zr atoms could diffuse into the metastable β-W layer after annealed at 540 °C. Then, the amorphous-like tungsten might be formed. The possible mechanism of the amorphous phase forming in the thin W layer of the high temperature annealed W/CoFeB/Zr/MgO multilayers will be discussed later.

FIG. 2.

The cross-sectional high-resolution TEM micrograph of W/CoFeB/Zr/MgO after annealing at 540 °C. The inset shows the FFT patterns from the selected region in red square.

FIG. 2.

The cross-sectional high-resolution TEM micrograph of W/CoFeB/Zr/MgO after annealing at 540 °C. The inset shows the FFT patterns from the selected region in red square.

Close modal
FIG. 3.

The elemental mapping data of the cross-sectional W/CoFeB/Zr/MgO annealed at 540 °C.

FIG. 3.

The elemental mapping data of the cross-sectional W/CoFeB/Zr/MgO annealed at 540 °C.

Close modal

Figure 4 shows the representative CIMS curves for the W/CoFeB/Zr/MgO multilayers after annealing from 270 to 540 °C. The CIMS curves exhibit typical binary switching and change their polarity when the applied in-plane field reverses. The switching direction of CIMS in these samples is consistent with the previous results.14 By varying the in-plane field (Hx) and the current density, the switching phase diagrams of the samples under Tann from 270 to 540 °C are plotted in Fig. 4(c). It is clear that the annealing stability for robust CIMS over 540 °C is well beyond the previous reports.27,28

FIG. 4.

Current-induced magnetization switching. (a) CIMS curves with in-plane assist fields of Hx = +150 and −150 Oe for the sample annealing at 270 °C. (b) CIMS curves with in-plane assist fields of Hx = +500 and −500 Oe for the sample annealing at 540 °C. (c) Switching phase diagrams for the samples annealed at 270, 360, 450, and 540 °C.

FIG. 4.

Current-induced magnetization switching. (a) CIMS curves with in-plane assist fields of Hx = +150 and −150 Oe for the sample annealing at 270 °C. (b) CIMS curves with in-plane assist fields of Hx = +500 and −500 Oe for the sample annealing at 540 °C. (c) Switching phase diagrams for the samples annealed at 270, 360, 450, and 540 °C.

Close modal
The thermal stability factor can be obtained by using the pulse driven CIMS measurements.29–31 Thus, the average switching current is given by29–32,
Ic=Ic011Δlntpulseτ0,
(1)
where Ic0 is the zero-thermal critical switching current and 1/τ0 is the attempt frequency (τ0 is usually taken to be 1 ns), which characterizes how often magnetization approaches the barrier due to thermal fluctuations. The threshold current Ic for switching is a function of tpulse/τ0, as shown in Figs. 5(a) and 5(b). From linear fits, we obtain Ic0 and Δ of the W/CoFeB/Zr/MgO multilayers. In the previous thermal stability studies using magnetic tunnel junctions,32–34 the energy barrier Eb and Δ strongly depend on the size of devices.33,34 When the device size is smaller than the nucleation diameter Dn, the magnetization switching obeys the single-domain switching model. Δ is proportional to the square of device diameter (D),33 i.e., Δ = Keffπ(D/2)2tCoFeB/kBT. When the device size is large enough, the magnetization switching is mediated by the nucleation and propagation of a domain wall across the device. Δ reflects the energy barrier of the volume with the size equal to the domain wall width [δw = π(As/Keff)1/2, where As is the exchange stiffness constant].34 The thermal stability factor of the device can be expressed as Δ ≈ Keffπ(δDW/2)2tCoFeB/kBT = π3AstCoFeB/4kBT, which is independent of the size of the device.
FIG. 5.

The current pulse width dependence of critical switching current. The linear fitting for the samples annealed at (a) 270 and (b) 540 °C. (c) Annealing temperature dependence of the thermal stability factor Δ.

FIG. 5.

The current pulse width dependence of critical switching current. The linear fitting for the samples annealed at (a) 270 and (b) 540 °C. (c) Annealing temperature dependence of the thermal stability factor Δ.

Close modal

It should be pointed out that Eq. (1) is for coherent switching in the magnetic layer.32 The typical spin–orbit torque switching behavior can be divided into a dynamical and a thermally activated regime, depending on the pulse width applied.19,30 As for the μm-scaled Hall bar device, the thermal reversal of magnetization becomes the primary switching mechanism if the pulse width used in the CIMS measurements (tpulse > 200 µs) is long enough.30 Thus, we could still use Eq. (1) to estimate Δ for a μm-scaled Hall bar device.31,35 Figure 5(c) shows the annealing temperature dependence of Δ. With an increase in Tann, the Δ shows a dramatic enhancement, which is mainly ascribed to the increased effective magnetic anisotropy energy Keff in the annealed films.23 When the sample was annealed at 540 °C, the Δ is about 79, which is greater than that reported for the SOT devices.29–31 When we further increase the Tann up to 600 °C, the value of Δ decreases to about 40, whereas the anisotropy field decreases from 15.5 to 5.4 kOe.23 The results suggest that the interfacial interdiffusion may become serious after annealing above 540 °C, which results in the suppression of the perpendicular anisotropy and the thermal stability factor.

As for the SOT efficiency, the harmonic Hall voltage measurements were carried out under the in-plane magnetic field parallel or perpendicular to the current direction (HL and HT).23 In both measurement configurations, the magnetic field is applied with a small tilting angle (about 2°) to the film plane. Figures 6(a) and 6(c) show the representative first harmonic (Vω) and second harmonic (V) Hall voltages under HL and HT, which are fitted by parabolic and linear functions, respectively. The longitudinal effective field (ΔHL) and the transverse effective field (ΔHT) can be calculated by the slope in Figs. 6(b) and 6(d), respectively, according to36 
ΔHDLFL=2V2ωHLT2VωHLT2.
(2)
FIG. 6.

The harmonic Hall voltage measurement in W/CoFeB/Zr/MgO annealed at 540 °C. (a) First Vω and (b) second V harmonic Hall voltages as a function of longitudinal magnetic field HL. (c) and (d) Curves of Vω and V2ω as a function of transverse magnetic field HT. (e) Current dependence of the longitudinal (transverse) effective field ΔHLHT). Note that +Mz (−Mz) denotes the magnetization pointing along the +z (−z) direction. (f) The annealing temperature dependence of the longitudinal (transverse) effective efficiency βL (βT).

FIG. 6.

The harmonic Hall voltage measurement in W/CoFeB/Zr/MgO annealed at 540 °C. (a) First Vω and (b) second V harmonic Hall voltages as a function of longitudinal magnetic field HL. (c) and (d) Curves of Vω and V2ω as a function of transverse magnetic field HT. (e) Current dependence of the longitudinal (transverse) effective field ΔHLHT). Note that +Mz (−Mz) denotes the magnetization pointing along the +z (−z) direction. (f) The annealing temperature dependence of the longitudinal (transverse) effective efficiency βL (βT).

Close modal

The longitudinal (transverse) torque efficiencies, βL(T), are defined as ΔHL(T)/Jc, where Jc is the applied current density in the W layer. In general, the ΔHL(T) needs to be corrected when the ratio of the planar Hall to the anomalous Hall voltages ξ is not negligible.37,38 However, we can still use Eq. (2) to calculate ΔHL(T) without planar Hall correction,36 when ξ is larger than 0.3 (see the supplementary material). As shown in Fig. 6(e), ΔHL and ΔHT are both linear with the applied current in the W layer (IW), indicating that the contribution of the current-induced Joule heating effect is small; thus, its influence on the magnetic properties can be negligible. Other thermal effects, such as the ordinary Nernst effect, anomalous Nernst effect, and spin Seebeck effect,39 can also be neglected when the alternating current is small (see the supplementary material). Figure 6(f) shows that the βL increases gently with increasing Tann, while the βT decreases rapidly.

Figure 7 shows the Tann dependence of the SOT performance in the W/CoFeB/Zr/MgO multilayers. The damping-like torque efficiency ξDL is calculated as40,
ξDL=2eMStCoFeBΔHL/Jc,
(3)
where tCoFeB is the thickness of CoFeB and is the reduced Planck constant. To simplify, we keep MS = 1060 emu/cm3 for the calculation of ξDL.23 It is surprising that ξDL of the W/CoFeB/Zr/MgO multilayers is independent of Tann and almost keeps a constant value of −0.3, while the ρW has a strong decrease with the increase in Tann. Since the spin transparency of β-W/FM interface is likely to be closer to unity,41 the lower bound of SHC σSH obeys the following relation: σSH = (/2e)ξDL/ρW. In addition, considering that the spin diffusion length of the high resistive W may be quite short if the Elliot–Yafet spin scattering mechanism dominates, ξDL reaches its saturation value when the thickness of W is equal to 4.5 nm. Hence, we can roughly calculate the annealing temperature dependence of σSH by using ξDL. As shown in Fig. 7(c), the |σSH| gradually increases from 765 of β-W to 1420 (ℏ/e)(Ω cm)−1 of amorphous-like W, which will be discussed later.
FIG. 7.

Annealing temperature dependence of the spin Hall effect of tungsten in the W/CoFeB/Zr/MgO multilayers. (a) The resistivity and (b) damping-like torque efficiency ξDL of the tungsten layer. (c) Spin Hall conductance σSH. The lines in (b) and (c) are guides to the eyes.

FIG. 7.

Annealing temperature dependence of the spin Hall effect of tungsten in the W/CoFeB/Zr/MgO multilayers. (a) The resistivity and (b) damping-like torque efficiency ξDL of the tungsten layer. (c) Spin Hall conductance σSH. The lines in (b) and (c) are guides to the eyes.

Close modal

Note that the thin W film (normally below 6 nm) is β-phase (A15 crystal structure) with a resistivity of 100–300 μΩ cm4,14,15,18,19 and high θSH, while the thick W film tends to be purely α-W with a resistivity of about 20 μΩ cm and low θSH. The very thin as-prepared W layer could be amorphous with a quite high resistivity (about 260 μΩ cm).19 From a thermodynamic point of view, the stable structure of tungsten is bcc α-W. The nonequilibrium metastable β-W thin layer can be obtained at the as-deposited state. After annealing, the amorphous state can also be reported due to the incorporation of the light element, such as B,19 in tungsten. In our case, we suggest that the amorphous state can be formed due to the incorporation of B and Zr elements in β-W after high temperature annealing. According to the three empirical rules proposed by Inoue,42 bulk amorphous alloys can be obtained, due to more than three elements, large atomic size difference, and negative heat of mixing. The difference in atomic sizes between W (1.41 Å) and Zr (1.60 Å) is about 12%, and the heat of mixing between W and Zr is −9 kJ/mol. The atomic size of B (0.09 Å) is much small, and the heat of mixing between W and B and between Zr and B is −31 and −71 kJ/mol, respectively. Therefore, the incorporation of B and Zr elements in W can obey all three rules for the amorphous forming, although the composition needs to be further investigated.

The theoretically calculated SHC |σSH| of β-W is 1255 (ℏ/e)(Ω cm)−1.43 In our experiments, the lower bound |σSH| of β-W (in the sample after annealing at 270 °C) is measured as 765 (ℏ/e)(Ω cm)−1. The difference may be due to the sputter-deposited β-W film that usually mixes with α-W.4,14 The significant feature for the SHC of W in the W/CoFeB/Zr/MgO multilayer is that the lower bound |σSH| is 1420 (ℏ/e)(Ω cm)−1 for the mediate resistive amorphous-like tungsten (in the sample after annealing at 540 °C), which is near 20% larger than the theoretically calculated SHC of β-W. Although the intrinsic SHC by the Berry curvature is known to be insensitive to the resistivity, it should be pointed out that ab initio calculations also exhibit that the SHC of β-W can be enhanced by doping.43,44 For instance, the calculated |σSH| of β-W is 1773 (ℏ/e)(Ω cm)−1 by doping with Ta.43 Furthermore, the preliminary experimental results show that the spin Hall angle of β-W stays near 0.2 regardless of the Ta doping percentage,45 but the electrical conductance of the alloy film increases with the Ta doping.

Meanwhile, it cannot exclude the possibility that a small amount of α-W, converted from β-W, could be mixed after annealing at 540 °C, reducing the total resistivity of the W layer after annealing at 540 °C. A large SHC is still kept because a quite large θSH in α-W could also be possible.46 Although the contents and the annealing temperature dependence of B or Zr in W may not be accurately detected, the diffused B or Zr atoms could act as scattering centers. Further detailed transport measurements are necessary to distinguish the contributions of extrinsic mechanisms in SHC.

The fact of the reduction in W resistivity in the annealed W/CoFeB/Zr/MgO multilayer suggests that the structure of the W layer may directly change from β-phase to a mediate resistive amorphous-like structure during annealing. The resistivity of W layer decreases nearly by 40% when the Tann increases up to 540 °C, which is then due to a larger hole doping effect, whereas the damping-like efficiency almost keeps −0.3. Therefore, the SHC enhances in the annealed W/CoFeB/Zr/MgO multilayer. Nevertheless, the power consumption for switching FM electrodes in unit magnetic volume is Pρxx/|ξDL|2.36 Compared to typical HMs, such as W, the power consumption of the W/CoFeB/Zr/MgO samples could have a dramatic reduction, due to the low resistivity after annealing.

To conclude, we have demonstrated that the spin–orbit torque driven magnetization switching in the perpendicularly magnetized W buffered CoFeB/MgO multilayers with a thin Zr layer (about 0.43 nm) is robust against high annealing temperatures up to 540 °C. When the annealing temperature increases from 270 to 540 °C, the structure of the tungsten layer in the W/CoFeB/Zr/MgO multilayer changes from β-phase to a mediate resistive amorphous-like structure, while the damping-like efficiency almost keeps constant (about −0.3). The absolute spin Hall conductance increases from 765 for β-phase to 1420 (ℏ/e)(Ω cm)−1 for this new amorphous-like tungsten. Benefited from the mediate resistive amorphous W with high SHC, the normalized power consumption could be reduced. These results indicate that the W/CoFeB/Zr/MgO multilayers can be a very compelling material for SOT applications and also pave a new way for achieving SOT devices with mature CMOS techniques.

See the supplementary material for the x-ray diffractions, polarized neutron reflectometry, planar Hall effect measurement, and thermal effect analysis.

This study was supported by the National Key Research and Development Program of China (Grant Nos. 2020YFA0406002 and 2021YFA1400300), National Natural Science Foundation of China (Grant Nos. 52130103, U22A20263, and 51871018), Beijing Laboratory of Metallic Materials and Processing for Modern Transportation, Opening Project of Key Laboratory of Microelectronics Devices & Integrated Technology, Institute of Microelectronics of Chinese Academy of Sciences, Beijing Natural Science Foundation (Grant No. Z180014), and Beijing Outstanding Young Scientists Projects (Grant No. BJJWZYJH01201910005018). H.-W.L. was supported by the National Research Foundation of Korea (Grant No. NRF- 2018R1A5A6075964).

The authors have no conflicts to disclose.

Q. X. Guo: Data curation (equal); Formal analysis (lead); Investigation (equal); Writing – original draft (equal). Z. C. Zheng: Data curation (equal); Formal analysis (supporting); Investigation (equal). L. H. Wang: Data curation (equal); Funding acquisition (supporting). K. Wang: Data curation (supporting); Investigation (supporting). X. M. Wang: Funding acquisition (supporting); Supervision (supporting). S. Cheng: Data curation (supporting); Formal analysis (supporting). W. He: Investigation (supporting). G. H. Yu: Funding acquisition (supporting); Project administration (supporting). H.-W. Lee: Formal analysis (supporting); Funding acquisition (supporting); Writing – review & editing (supporting). Y. Q. Guo: Formal analysis (supporting). J. Teng: Funding acquisition (supporting); Investigation (supporting); Supervision (equal). T. Zhu: Conceptualization (lead); Funding acquisition (lead); Project administration (lead); Supervision (equal); Writing – review & editing (lead).

The data that support the findings of this study are available within the article and its supplementary material.

1.
I. M.
Miron
,
K.
Garello
,
G.
Gaudin
,
P.-J.
Zermatten
,
M. V.
Costache
,
S.
Auffret
,
S.
Bandiera
,
B.
Rodmacq
,
A.
Schuhl
, and
P.
Gambardella
, “
Perpendicular switching of a single ferromagnetic layer induced by in-plane current injection
,”
Nature
476
,
189
(
2011
).
2.
L. Q.
Liu
,
O. J.
Lee
,
T. J.
Gudmundsen
,
D. C.
Ralph
, and
R. A.
Buhrman
, “
Current-induced switching of perpendicularly magnetized magnetic layers using spin torque from the spin Hall effect
,”
Phys. Rev. Lett.
109
,
096602
(
2012
).
3.
L. Q.
Liu
,
C.-F.
Pai
,
Y.
Li
,
H.
Tseng
,
D. C.
Ralph
, and
R. A.
Buhrman
, “
Spin-torque switching with the giant spin Hall effect of tantalum
,”
Science
336
,
555
(
2012
).
4.
C.-F.
Pai
,
L. Q.
Liu
,
Y.
Li
,
H. W.
Tseng
,
D. C.
Ralph
, and
R. A.
Buhrman
, “
Spin transfer torque devices utilizing the giant spin Hall effect of tungsten
,”
Appl. Phys. Lett.
101
,
122404
(
2012
).
5.
J.
Kim
,
J.
Sinha
,
M.
Hayashi
,
M.
Yamanouchi
,
S.
Fukami
,
T.
Suzuki
,
S.
Mitani
, and
H.
Ohno
, “
Layer thickness dependence of the current-induced effective field vector in Ta|CoFeB|MgO
,”
Nat. Mater.
12
,
240
(
2013
).
6.
K.
Garello
,
I. M.
Miron
,
C. O.
Avci
,
F.
Freimuth
,
Y.
Mokrousov
,
S.
Blugel
,
S.
Auffret
,
O.
Boulle
,
G.
Gaudin
, and
P.
Gambardella
, “
Symmetry and magnitude of spin-orbit torques in ferromagnetic heterostructures
,”
Nat. Nanotechnol.
8
,
587
(
2013
).
7.
S.
Emori
,
U.
Bauer
,
S. M.
Ahn
,
E.
Martinez
, and
G. S.
Beach
, “
Current-driven dynamics of chiral ferromagnetic domain walls
,”
Nat. Mater.
12
,
611
(
2013
).
8.
G.
Yu
,
U.
Pramey
,
Y.
Fan
,
J. G.
Alzate
,
W.
Jiang
,
K. L.
Wong
,
S.
Takei
,
S. A.
Bender
,
L.-T.
Chang
,
Y.
Jiang
,
M.
Lang
,
J.
Tang
,
Y.
Wang
,
Y.
Tserkovnyak
,
P. K.
Amiri
, and
K. L.
Wang
, “
Switching of perpendicular magnetization by spin-orbit torques in the absence of external magnetic fields
,”
Nat. Nanotechnol.
9
,
548
(
2014
).
9.
D.
Bhowmik
,
L.
You
, and
S.
Salahuddin
, “
Spin Hall effect clocking of nanomagnetic logic without a magnetic field
,”
Nat. Nanotechnol.
9
,
59
(
2014
).
10.
X. P.
Qiu
,
P.
Deorani
,
K.
Narayanapillai
,
K. S.
Lee
,
K. J.
Lee
,
H. W.
Lee
, and
H.
Yang
, “
Angular and temperature dependence of current induced spin-orbit effective fields in Ta/CoFeB/MgO nanowires
,”
Sci. Rep.
4
,
4491
(
2014
).
11.
A.
van den Brink
et al
, “
Field-free magnetization reversal by spin-Hall effect and exchange bias
,”
Nat. Commun.
7
,
10854
(
2016
).
12.
L.
Liu
,
C. H.
Zhou
,
X. Y.
Shu
,
C. J.
Li
,
T. Y.
Zhao
,
W.
Lin
,
J. Y.
Deng
,
Q. D.
Xie
,
S. H.
Chen
,
J.
Zhou
,
R.
Guo
,
H.
Wang
,
J. H.
Yu
,
S.
Shi
,
P.
Yang
,
S.
Pennycook
,
A.
Manchon
, and
J. S.
Chen
, “
Symmetry-dependent field-free switching of perpendicular magnetization
,”
Nat. Nanotechnol.
16
,
277
(
2021
).
13.
S.
Ikeda
,
K.
Miura
,
H.
Yamamoto
,
K.
Mizunuma
,
H. D.
Gan
,
M.
Endo
,
S.
Kanai
,
J.
Hayakawa
,
F.
Matsukura
, and
H.
Ohno
, “
A perpendicular-anisotropy CoFeB–MgO magnetic tunnel junction
,”
Nat. Mater.
9
,
721
(
2010
).
14.
C.-F.
Pai
,
M. H.
Nguyen
,
C.
Belvin
,
L. H.
Vilela-Leao
,
D. C.
Ralph
, and
R. A.
Buhrman
, “
Enhancement of perpendicular magnetic anisotropy and transmission of spin-Hall-effect-induced spin currents by a Hf spacer layer in W/Hf/CoFeB/MgO layer structures
,”
Appl. Phys. Lett.
104
,
082407
(
2014
).
15.
Q.
Hao
and
G.
Xiao
, “
Giant spin Hall effect and switching induced by spin-transfer torque in a W/Co40Fe40B20/MgO structure with perpendicular magnetic anisotropy
,”
Phys. Rev. Appl.
3
,
034009
(
2015
).
16.
K.-U.
Demasius
,
T.
Phung
,
W.
Zhang
,
B. P.
Hughes
,
S.-H.
Yang
,
A.
Kellock
,
W.
Han
,
A.
Pushp
, and
S. S. P.
Parkin
, “
Enhanced spin-orbit torques by oxygen incorporation in tungsten films
,”
Nat. Commun.
7
,
10644
(
2016
).
17.
L.
Neumann
,
D.
Meier
,
J.
Schmalhorst
,
K.
Rott
,
G.
Reiss
, and
M.
Meinert
, “
Temperature dependence of the spin Hall angle and switching current in the nc-W(O)/CoFeB/MgO system with perpendicular magnetic anisotropy
,”
Appl. Phys. Lett.
109
,
142405
(
2016
).
18.
L.
Zhang
,
X. Y.
Zhang
,
M. X.
Wang
,
Z. H.
Wang
,
W. L.
Cai
,
K. H.
Cao
,
D. Q.
Zhu
,
H. W.
Yang
, and
W. S.
Zhao
, “
Size dependence of the spin-orbit torque induced magnetic reversal in W/CoFeB/MgO nanostructures
,”
Appl. Phys. Lett.
112
,
142410
(
2018
).
19.
G. W.
Kim
,
A. S.
Samardak
,
Y. J.
Kim
,
I. H.
Cha
,
A. V.
Ognev
,
A. V.
Sadovnikov
,
S. A.
Nikitov
, and
Y. K.
Kim
, “
Role of the heavy metal’s crystal phase in oscillations of perpendicular magnetic anisotropy and the interfacial Dzyaloshinskii–Moriya interaction in W/Co–Fe–B/MgO films
,”
Phys. Rev. Appl.
9
,
064005
(
2018
).
20.
X.
Li
,
S.-J.
Lin
,
M.
Dc
,
Y.-C.
Liao
,
C.
Yao
,
A.
Naeemi
,
W.
Tsai
, and
S. X.
Wang
, “
Materials requirements of high-speed and low-power spin-orbit-torque magnetic random-access memory
,”
IEEE J. Electron Devices Soc.
8
,
674
(
2020
).
21.
K. L.
Wang
,
J. G.
Alzate
, and
P.
Khalili Amiri
, “
Low-power non-volatile spintronic memory: STT-RAM and beyond
,”
J. Phys. D: Appl. Phys.
46
,
074003
(
2013
).
22.
L.
Thomas
,
G.
Jan
,
J.
Zhu
,
H.
Liu
,
Y.-J.
Lee
,
S.
Le
,
R.-Y.
Tong
,
K.
Pi
,
Y.-J.
Wang
,
D.
Shen
,
R.
He
,
J.
Haq
,
J.
Teng
,
V.
Lam
,
K.
Huang
,
T.
Zhong
,
T.
Torng
, and
P.-K.
Wang
, “
Perpendicular spin transfer torque magnetic random access memories with high spin torque efficiency and thermal stability for embedded applications (invited)
,”
J. Appl. Phys.
115
,
172615
(
2014
).
23.
Q. X.
Guo
,
K.
Wang
,
H.
Bai
,
Z. C.
Zheng
,
L. H.
Wang
,
X. M.
Wang
,
W.
He
,
G. H.
Yu
,
J.
Teng
, and
T.
Zhu
, “
Ultra-high thermal stability of perpendicular magnetic anisotropy in the W buffered CoFeB/MgO stacks with Zr dusting layers
,”
Appl. Phys. Lett.
120
,
022402
(
2022
).
24.
Q.
Hao
,
W.
Chen
, and
G.
Xiao
, “
Beta (β) tungsten thin films: Structure, electron transport, and giant spin Hall effect
,”
Appl. Phys. Lett.
106
,
182403
(
2015
).
25.
C.
Zhang
,
S.
Fukami
,
K.
Watanabe
,
A.
Ohkawara
,
S.
DuttaGupta
,
H.
Sato
,
F.
Matsukura
, and
H.
Ohno
, “
Critical role of W deposition condition on spin-orbit torque induced magnetization switching in nanoscale W/CoFeB/MgO
,”
Appl. Phys. Lett.
109
,
192405
(
2016
).
26.
P.
Sheng
,
Y.
Sakuraba
,
Y.-C.
Lau
,
S.
Takahashi
,
S.
Mitani
, and
M.
Hayashi
, “
The spin Nernst effect in tungsten
,”
Sci. Adv.
3
,
e1701503
(
2017
).
27.
X.
Li
,
G. Q.
Yu
,
H.
Wu
,
P. V.
Ong
,
K.
Wong
,
Q.
Hu
,
F.
Ebrahimi
,
P.
Upadhyaya
,
M.
Akyol
,
N.
Kioussis
,
X. F.
Han
,
P.
Khalili Amiri
, and
K. L.
Wang
, “
Thermally stable voltage-controlled perpendicular magnetic anisotropy in Mo|CoFeB|MgO structures
,”
Appl. Phys. Lett.
107
,
142403
(
2015
).
28.
D.
Wu
,
G. Q.
Yu
,
Q. M.
Shao
,
X.
Li
,
H.
Wu
,
K. L.
Wong
,
Z. Z.
Zhang
,
X. F.
Han
,
P.
Khalili Amiri
, and
K. L.
Wang
,
Appl. Phys. Lett.
108
,
212406
(
2016
).
29.
T.-Y.
Chen
,
H.-I.
Chan
,
W.-B.
Liao
, and
C.-F.
Pai
, “
Current-induced spin-orbit torque and field-free switching in Mo-based magnetic heterostructures
,”
Phys. Rev. Appl.
10
,
044038
(
2018
).
30.
W.-B.
Liao
,
T.-Y.
Chen
,
Y.-C.
Hsiao
, and
C.-F.
Pai
, “
Pulse-width and temperature dependence of memristive spin-orbit torque switching
,”
Appl. Phys. Lett.
117
,
182402
(
2020
).
31.
T.
Fan
,
N. H. D.
Khang
,
S.
Nakano
, and
P. N.
Hai
, “
Ultrahigh efficient spin orbit torque magnetization switching in fully sputtered topological insulator and ferromagnet multilayers
,”
Sci. Rep.
12
,
2998
(
2022
).
32.
R. H.
Koch
,
J. A.
Katine
, and
J. Z.
Sun
, “
Time-resolved reversal of spin-transfer switching in a nanomagnet
,”
Phys. Rev. Lett.
92
,
088302
(
2004
).
33.
H.
Sato
,
M.
Yamanouchi
,
K.
Miura
,
S.
Ikeda
,
H. D.
Gan
,
K.
Mizunuma
,
R.
Koizumi
,
F.
Matsukura
, and
H.
Ohno
, “
Junction size effect on switching current and thermal stability in CoFeB/MgO perpendicular magnetic tunnel junctions
,”
Appl. Phys. Lett.
99
,
042501
(
2011
).
34.
H.
Sato
,
E. C. I.
Enobio
,
M.
Yamanouchi
,
S.
Ikeda
,
S.
Fukami
,
S.
Kanai
,
F.
Matsukura
, and
H.
Ohno
, “
Properties of magnetic tunnel junctions with a MgO/CoFeB/Ta/CoFeB/MgO recording structure down to junction diameter of 11 nm
,”
Appl. Phys. Lett.
105
,
062403
(
2014
).
35.
A. V.
Khvalkovskiy
,
D.
Apalkov
,
S.
Watts
,
R.
Chepulskii
,
R. S.
Beach
,
A.
Ong
,
X.
Tang
,
A.
Driskill-Smith
,
W. H.
Butler
,
P. B.
Visscher
,
D.
Lottis
,
E.
Chen
,
V.
Nikitin
, and
M.
Krounbi
,
J. Phys. D: Appl. Phys.
46
,
139601
(
2013
).
36.
L.
Zhu
,
K.
Sobotkiewich
,
X.
Ma
,
X. Q.
Li
,
D. C.
Ralph
, and
R. A.
Buhrman
, “
Strong damping-like spin-orbit torque and tunable Dzyaloshinskii–Moriya interaction generated by low-resistivity Pd1−xPtx alloys
,”
Adv. Funct. Mater.
29
,
1805822
(
2019
).
37.
M.
Hayashi
,
J.
Kim
,
M.
Yamanouchi
, and
H.
Ohno
, “
Quantitative characterization of the spin-orbit torque using harmonic Hall voltage measurements
,”
Phys. Rev. B
89
,
144425
(
2014
).
38.
J.
Torrejon
,
J.
Kim
,
J.
Sinha
,
S.
Mitani
,
M.
Hayashi
,
M.
Yamanouchi
, and
H.
Ohno
, “
Interface control of the magnetic chirality in CoFeB/MgO heterostructures with heavy-metal underlayers
,”
Nat. Commun.
5
,
4655
(
2014
).
39.
N.
Roschewsky
,
E. S.
Walker
,
P.
Gowtham
,
S.
Muschinske
,
F.
Hellman
,
S. R.
Bank
, and
S.
Salahuddin
, “
Spin-orbit torque and Nernst effect in Bi–Sb/Co heterostructures
,”
Phys. Rev. B
99
,
195103
(
2019
).
40.
Z. C.
Zheng
,
Q. X.
Guo
,
D.
Jo
,
D.
Go
,
L. H.
Wang
,
H. C.
Chen
,
W.
Yin
,
X. M.
Wang
,
G. H.
Yu
,
W.
He
,
H.-W.
Lee
,
J.
Teng
, and
T.
Zhu
, “
Magnetization switching driven by current-induced torque from weakly spin-orbit coupled Zr
,”
Phys. Rev. Res.
2
,
013127
(
2020
).
41.
C.-F.
Pai
,
Y.
Ou
,
L. H.
Vilela-Leão
,
D.
Ralph
, and
R.
Buhrman
, “
Dependence of the efficiency of spin Hall torque on the transparency of Pt/ferromagnetic layer interfaces
,”
Phys. Rev. B
92
,
064426
(
2015
).
42.
A.
Inoue
, “
Stabilization of metallic supercooled liquid and bulk amorphous alloys
,”
Acta Mater.
48
,
279
(
2000
).
43.
X. L.
Sui
,
C.
Wang
,
J. W.
Kim
,
J. F.
Wang
,
S. H.
Rhim
,
W. H.
Duan
, and
N.
Kioussis
, “
Giant enhancement of the intrinsic spin Hall conductivity in β-tungsten via substitutional doping
,”
Phys. Rev. B
96
(
24
),
241105
(
2017
).
44.
E.
Derunova
,
Y.
Sun
,
C.
Felser
,
S. S. P.
Parkin
,
B.
Yan
, and
M. N.
Ali
, “
Giant intrinsic spin Hall effect in W3Ta and other A15 superconductors
,”
Sci. Adv.
5
,
eaav8575
(
2019
).
45.
Y.
Saito
,
N.
Tezuka
,
S.
Ikeda
,
H.
Sato
, and
T.
Endoh
, “
Increase in spin-Hall effect and influence of anomalous Nernst effect on spin-Hall magnetoresistance in β-phase and α-phase W100−xTax/CoFeB systems
,”
Appl. Phys. Express
12
,
053008
(
2019
).
46.
W. B.
Liao
,
T. Y.
Chen
,
Y.
Ferrante
,
S. S. P.
Parkin
, and
C. F.
Pai
, “
Current‐induced magnetization switching by the high spin Hall conductivity α‐W
,”
Phys. Status Solidi RRL
13
,
1900408
(
2019
).

Supplementary Material