⇦Chinese Physics Letters, 2024, Vol. 41, No. 5, Article code 057503 Current-Induced Magnetization Switching Behavior in Perpendicular Magnetized ${\rm L1_{0}}$-MnAl/B2-CoGa Bilayer Hong-Li Sun (孙宏利)1,2, Rong-Kun Han (韩荣坤)1,2, Hong-Rui Qin (秦红蕊)1,2, Xu-Peng Zhao (赵旭鹏)3, Zhi-Cheng Xie (谢志成)1,2, Da-Hai Wei (魏大海)1,2, and Jian-Hua Zhao (赵建华)1,2* Affiliations 1State Key Laboratory of Superlattices and Microstructures, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, China 2Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100190, China 3International School of Microelectronics, Dongguan University of Technology, Dongguan 523808, China Received 20 March 2024; accepted manuscript online 30 April 2024; published online 23 May 2024 *Corresponding author. Email: jhzhao@semi.ac.cn Citation Text: Sun H L, Han R K, Qin H R et al. 2024 Chin. Phys. Lett. 41 057503    Abstract Rare-earth-free Mn-based binary alloy ${\rm L1_{0}}$-MnAl with bulk perpendicular magnetic anisotropy (PMA) holds promise for high-performance magnetic random access memory (MRAM) devices driven by spin-orbit torque (SOT). However, the lattice-mismatch issue makes it challenging to place conventional spin current sources, such as heavy metals, between ${\rm L1_{0}}$-MnAl layers and substrates. In this work, we propose a solution by using the B2-CoGa alloy as the spin current source. The lattice-matching enables high-quality epitaxial growth of 2-nm-thick ${\rm L1_{0}}$-MnAl on B2-CoGa, and the ${\rm L1_{0}}$-MnAl exhibits a large PMA constant of $1.04\times 10^{6}$ J/m$^{3}$. Subsequently, the considerable spin Hall effect in B2-CoGa enables the achievement of SOT-induced deterministic magnetization switching. Moreover, we quantitatively determine the SOT efficiency in the bilayer. Furthermore, we design an ${\rm L1_{0}}$-MnAl/B2-CoGa/Co$_{2}$MnGa structure to achieve field-free magnetic switching. Our results provide valuable insights for achieving high-performance SOT-MRAM devices based on ${\rm L1_{0}}$-MnAl alloy.
cpl-41-5-057503-fig1.png
cpl-41-5-057503-fig2.png
cpl-41-5-057503-fig3.png
cpl-41-5-057503-fig4.png
cpl-41-5-057503-fig5.png
 DOI:10.1088/0256-307X/41/5/057503 © 2024 Chinese Physics Society Article Text Magnetic random-access memory (MRAM) represents a promising avenue for next-generation storage devices within the semiconductor industry.[1,2] In MRAM technologies, modulating local magnetization via electrical methods is essential. Compared to two-terminal spin-transfer torque (STT) MRAM,[3,4] the spin-orbit torque (SOT) MRAM, as a three-terminal device, avoids the passage of large currents through the barrier of magnetic tunnel junction (MTJ), significantly enhancing the reliability and endurance of the device.[5-7] In the typical spin current generator (SCG)/ferromagnet (FM) bilayer, the SOT effect originates from the spin Hall effect (SHE) and interfacial Rashba–Edelstein effect.[8-10] For SHE, when applying a current along the $x$ direction, $y$-polarized spins are separated in the $z$ direction, accumulating at the interface and further influencing the adjacent FM layer. In films with perpendicular magnetic anisotropy (PMA), SOT-induced deterministic magnetization switching typically necessitates an in-plane magnetic field to break the symmetry between up and down magnetization directions.[11,12] However, introducing an external field complicates the device architecture, which hinders high-density integration. Consequently, the pursuit of field-free deterministic magnetization switching remains a crucial area of research. Several solutions have been identified in related studies encompassing interlayer exchange coupling (IEC),[13,14] wedge structure,[15,16] exchange bias,[17,18] competing spin currents,[19] and $z$-direction spin currents,[20,21] amongst other approaches.[22-25] For field-free SOT-induced magnetization switching via IEC, two FM layers, one with in-plane magnetic anisotropy (IMA) and the other with PMA, are separated by a non-magnetic (NM) layer. The IMA layer provides an in-plane equivalent magnetic field for the PMA layer, thereby achieving field-free SOT switching. Moreover, the SOT-induced switching behavior can be modulated by pre-magnetizing the IMA layer.[14] On the other hand, the IMA layer can directly generate $z$-direction spin currents ($\sigma_{z})$ through anomalous SHE, which could contribute to field-free SOT-induced magnetization switching.[26] Rare-earth-free Mn-based binary alloys, including ${\rm L1_{0}}$-MnGa, ${\rm D0_{22}}$-Mn$_{3}$Ga, and ${\rm L1_{0}}$-MnAl, are promising for MRAM electrodes owing to their superior magnetic properties, such as significant PMA constants, high spin polarizations, high Curie temperature, and low magnetic damping constants.[27-32] Compared to other PMA Mn-based binary alloys, the ${\rm L1_{0}}$-MnAl stands out being cost-effective, attributing to the abundant availability of Al, while maintaining substantial magnetic performance.[29,33-36] Furthermore, theoretical studies forecast the half metallicity at a $\varDelta_{1}$ band structure in ${\rm L1_{0}}$-MnAl, with the tunnel magnetoresistance (TMR) ratio in ${\rm L1_{0}}$-MnAl/MgO-based MTJ potentially exceeding 8000%, rendering it a suitable candidate for MTJ electrodes.[37] However, the bulk PMA of ${\rm L1_{0}}$-MnAl, originating from magnetocrystalline anisotropy, necessitates a high-quality single crystalline structure. This requirement means the necessity for lattice-matching buffer layers, which places rigorous demands on growth technology and stacking structure design and imposes constraints on applications of ${\rm L1_{0}}$-MnAl. Current research on the SOT effect in Mn-based binary alloys mainly places traditional SCG, such as heavy metals (HMs), on the top of stacking structure, attributed to their ineffectiveness as buffers because of the lattice-mismatch issue.[38-42] However, in instances where ${\rm L1_{0}}$-MnAl serves as the electrode in SOT-MTJ, employing a lattice-matching buffer as SCG is essential to ensure direct contact between ${\rm L1_{0}}$-MnAl and the MgO barrier to guarantee the high thermal stability and TMR ratios of MTJ. Consequently, there is an urgent requirement for an appropriate material capable of serving both as a buffer and as an SCG layer for ${\rm L1_{0}}$-MnAl alloy. Recent studies have demonstrated the excellent lattice-matching properties between paramagnetic B2-CoGa and Mn-based binary alloys.[43,44] In addition, B2-CoGa has been identified to exhibit considerable SHE arising from the 3$d$–4$p$ orbital hybridization, enabling its function as an SCG layer.[45,46] These findings establish B2-CoGa as a unique material that meets the aforementioned requirements. Furthermore, as previously noted, for the realization of field-free SOT-induced magnetization switching via IEC, it is imperative that both the IMA layer and the SCG layer exhibit lattice matching with ${\rm L1_{0}}$-MnAl. Research has shown that IMA Co-based full Heusler alloys Co$_{2}$MnX (X = Si, Ga, Al, Ge, Sn) can serve as a suitable buffer to enhance the PMA of the ${\rm L1_{0}}$-MnAl layer.[32] These results provide beneficial references for applications of ${\rm L1_{0}}$-MnAl in high-performance and field-free SOT-MRAM. However, studies pertaining to the SOT behavior in the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer remain limited. Moreover, research on field-free SOT-induced magnetization switching behavior in the ${\rm L1_{0}}$-MnAl alloy system is relatively scarce. In this Letter, we systemically investigate the structural and magnetic properties of the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer. An excellent PMA is obtained in the bilayer. Furthermore, we delve into the SOT-induced deterministic magnetization switching in the bilayer and have quantitatively evaluated the SOT efficiency. Moreover, we develop a multilayer stacking structure of ${\rm L1_{0}}$-MnAl/B2-CoGa/Co$_{2}$MnGa, where the top ${\rm L1_{0}}$-MnAl layer retains robust PMA and accomplishes field-free SOT-induced deterministic magnetization switching. Sample preparation was carried out using a VG-V80 molecular-beam epitaxy (MBE) system with two interconnected growth chambers. The multilayers were grown on semi-insulating (SI) GaAs (001) substrates, employing top-bottom stacking structures of Al/${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5)/GaAs and Al/${\rm L1_{0}}$-MnAl(2)/B2-CoGa(3)/Co$_{2}$MnGa(1.5)/GaAs (thicknesses in nanometers). Initially, the GaAs substrate was transferred into one growth chamber for deoxygenation and growth of a 200-nm-thick GaAs buffer layer, aiming to achieve a smooth interface. Subsequently, the substrate was transferred into another growth chamber for the growth of the metal multilayer. The CoGa layer was deposited, maintaining a nominal atomic ratio of 1 : 1 at 350 ℃. After growth, the substrate temperature was lowered to 300 ℃ to grow the MnAl layer. Finally, an Al layer was deposited on the top at room temperature to prevent oxidation. Characterization of the crystalline structure and magnetic properties was performed using x-ray diffraction (XRD) and superconductive quantum interface devices (SQUIDs). The XRD measurements were performed using the Cu $K_{\alpha1}$ with a wavelength of 1.54056 Å and $\theta$–2$\theta$ geometry. Subsequently, the samples were patterned into Hall devices through photolithography and ion beam etching (IBE) to perform the anomalous Hall effect (AHE), SOT-induced magnetization switching, and harmonic Hall voltage response (HHVR) measurements. Electrical characterization was conducted using the physical property measurements system (PPMS), combined with a Keithley 6221 current source, a 2182A nanovoltmeter, and a Stanford SR830 lock-in amplifier.
cpl-41-5-057503-fig1.png
Fig. 1. [(a), (b)] Unit cells of tetragonal ${\rm L1_{0}}$-MnAl and cubic B2-CoGa alloys. (c) Schematic stacking structure of the sample. (d) XRD patterns of ${\rm L1_{0}}$-MnAl(30)/B2-CoGa(1) thin film.
Figures 1(a) and 1(b) illustrate the structure diagrams of tetragonal ${\rm L1_{0}}$-MnAl and cubic B2-CoGa alloys.[47,48] Within the ${\rm L1_{0}}$-MnAl structure, spins of adjacent Mn atoms align parallel to each other along the $z$ axis, exhibiting ferromagnetism. The stacking structure of the samples is displayed in Fig. 1(c). For reference, we grew Al/MnAl(30)/CoGa(1)/GaAs structure for XRD characterization to investigate the structural properties of the MnAl layer. Figure 1(d) reveals distinct GaAs (002) and (004) substrate peaks, along with the (001) superlattice peak and (002) fundamental peak of MnAl, indicating the high (001) orientation. The pronounced intensity of the superlattice (001) peak suggests a high degree of chemical ordering in the ${\rm L1_{0}}$-MnAl thin films, favorable for achieving superior PMA. In addition, the small values of full width at half maximum (FWHM) of the ${\rm L1_{0}}$-MnAl peaks are indicative of a high-quality crystalline structure. The (001) and (002) peak positions have been determined to be 26.02$^{\circ}$ and 53.56$^{\circ}$, respectively. According to Bragg's law, the lattice constant has been calculated to be 3.42 Å, marginally smaller than the bulk state value (3.54 Å) attributed to the strain effect.[30] These findings reveal that ${\rm L1_{0}}$-MnAl/B2-CoGa bilayers are capable of achieving excellent epitaxial growth on GaAs (001) substrates.
cpl-41-5-057503-fig2.png
Fig. 2. (a) Hall device and measure schematic. (b) Hysteresis curve of the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) structure. (c) Out-of-plane AHE curve of the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) structure. (d) In-plane AHE curve of the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) structure.
Following the characterization of the crystalline structures, we further measured the magnetic properties of the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) bilayer. The bilayer was patterned into the Hall device for electrical measurements, as illustrated in Fig. 2(a). The out-of-plane hysteresis loop and the AHE loop are shown in Figs. 2(b) and 2(c), respectively. The loops exhibit a high remanence ratio, indicating the good PMA of the bilayer, which is constant with the structural analysis. Moreover, the saturation magnetization ($M_{\rm s})$ of the bilayer is determined to be 325.79 kA/m. To quantitatively assess the PMA, we characterize the effective perpendicular anisotropy field ($H_{k}^{\rm eff})$ of the bilayer by measuring the in-plane AHE curves. As shown in Fig. 2(d), the $R_{\rm AH}$ value gradually decreases to zero as the in-plane external magnetic field increases, indicating that the perpendicular magnetization is progressively aligned parallel to the field. As a result, the value of $\mu_0 H_{k}^{\rm eff}$ is determined to be 5.97 T. According to the formula $K_{\rm u}=M_{\rm s}^{2}$/2$\mu_0+H_{k}^{\rm eff}M_{\rm s}$/2, the PMA constant ($K_{\rm u}$) of the sample is determined to be $1.04\times 10^{6}$ J/m$^{3}$. In our experiment, a higher PMA constant is achieved in 2-nm-thick ${\rm L1_{0}}$-MnAl with B2-CoGa buffer, surpassing the values in the nanometer-thick ${\rm L1_{0}}$-MnAl/B2-CoAl ($8.5\times 10^{5}$ J/m$^{3})$ and ${\rm L1_{0}}$-MnGa/B2-CoGa ($4\times 10^{5}$ J/m$^{3})$ bilayer.[45,49] We have calculated the critical size $d$ using the typical thermal stability condition $K_{\rm u}d^{3}$/$k_{\rm B}T\ge 60$, where $k_{\rm B}$ and $T$ represent the Boltzmann constant and temperature, respectively. In this study, we have obtained $K_{\rm u}=1.04\times 10^{6}$ J/m$^{3}$. At room temperature ($T=300$ K), the value of $d$ has been calculated to be 6.21 nm. The result represents that the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) bilayer in our experiment is able to support the 6.21 nm diameter MTJ size to be capable of sustaining data integrity at room temperature for over ten years, thereby offering a substantial material foundation for the high-performance MRAM.
cpl-41-5-057503-fig3.png
Fig. 3. (a) Schematic diagram of SOT-driven magnetization switching. (b) Deterministic magnetization switching of ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) under $\mu_{0}H_{x}=\pm 100$ mT. (c) SOT-induced magnetization switching curves of the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) structure for the different magnitude of $\mu_{0}H_{x}$. (d) The variations of switching ratio and $J_{\rm c}$ with $\mu_{0}H_{x}$.
Based on the above results, we investigated the SOT-induced magnetization switching behaviors in the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) bilayer. A series of pulsed currents with 50 µs duration were applied along the $x$-axis of the Hall device. Subsequently, a current of 0.1 mA was employed to measure the anomalous Hall resistance ($R_{\rm AH})$ in the sample. An in-plane assistant magnetic field ($\mu_{0}H_{x}$) was applied to achieve deterministic magnetization switching. As depicted in Fig. 3(b), under a $\mu_{0}H_{x}=+100$ mT, the out-of-plane magnetization switches from down to up as the current sweeps from negative to positive, and it switches back when the current sweeps in the opposite direction. When the direction of $\mu_{0}H_{x}$ is reversed, the switching orientation changes from counterclockwise to clockwise. The reversed switching direction observed under $\mu_{0}H_{x}=\pm 100$ mT confirms the realization of SOT-induced deterministic magnetization switching. Additionally, the counterclockwise switching is exhibited under positive $\mu_{0}H_{x}$, which is consistent with the previous finding that CoGa exhibits a positive effective SOT efficiency.[45] Subsequent investigations focus on the SOT-induced magnetization switching behavior of the samples under different magnitudes of $\mu_{0}H_{x}$, as depicted in Fig. 3(c). Taking into account the shunt effect of the parallel circuit and using resistivity values of $\rho_{\rm CoGa}=175\,µ\Omega\cdot$cm and $\rho_{\rm MnAl}=165\,µ\Omega\cdot$cm, the critical switching current density ($J_{\rm c}$) values of the sample, defined as the current density at which $R_{\rm AH}$ changes by half, have been meticulously calculated. Figures 3(d) and 3(e) illustrate the variation of the switching ratio and the $J_{\rm c}$ with different $\mu_{0}H_{x}$ magnitudes. The value of $J_{\rm c}$ is determined to be $4.63\times 10^{7}$ A/cm$^{2}$ under $\mu_{0}H_{x}=100$ mT, which is higher than the ${\rm L1_{0}}$-MnGa/B2-CoGa bilayer ($1.6\times 10^{7}$ A/cm$^{2})$ due to the significantly larger $K_{\rm u}$ but is lower than the value in the Ta/${\rm L1_{0}}$-MnGa bilayer ($8.5\times 10^{7}$ A/cm$^{2})$.[39,45] Additionally, with the increase of $\mu_{0}H_{x}$, the switching ratio initially increases to a peak value before beginning to decrease. The $J_{\rm c}$ decreases with increasing $\mu_{0}H_{x}$, yet the relationship is not strictly linear, suggesting that the switching process is governed by a magnetic domain correlation mechanism rather than the macro-spin model.[50] The situation is similar under positive $\mu_{0}H_{x}$ changes and will not be reiterated here. In the previous study of the ${\rm L1_{0}}$-MnGa(2)/B2-CoGa(5) bilayer, the damping-like and field-like SOT efficiencies have been identified as $\xi_{\rm DL}=0.05$ and $\xi_{\rm FL}=0.27$, respectively. The damping-like SOT efficiency arises from the SHE induced by the 3$d$–4$p$ orbital hybridization in CoGa alloy, while the substantial field-like SOT efficiency possibly originates from the special band matching at the ${\rm L1_{0}}$-MnGa/B2-CoGa interface.[45] Based on previous results, we have quantitatively determined the $\xi_{\rm DL}$ and $\xi_{\rm FL}$ in the ${\rm L1_{0}}$-MnAl(2)/B2-CoGa(5) system using the HHVR measurements. The schematic diagram of the HHVR measurements is shown in Fig. 4(a). In the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer-based Hall device, applying an alternating current (AC) along the $x$- or $y$-axis induces the current-dependent SOT effective fields, including damping-like and field-like effective fields ($H_{\rm DL}$ and $H_{\rm FL}$). These fields will periodically modulate the magnetization, which results in the generation of high-order harmonic Hall voltage signals. By measuring both the variation of the first and second-order harmonic Hall voltage signals in response to external magnetic field $H_{x}$ and $H_{y}$, the $H_{\rm DL}$ and $H_{\rm FL}$ of the bilayer can be ascertained by the following equation:[41] \begin{equation} H_{\rm{DL(FL)}}=-\frac{2{\partial V_{2\omega}}/{\partial H_{x(y)}}}{{\partial^{2}V_{1\omega}}/{\partial H_{x(y)}^{2}}}. \tag {1} \end{equation} Figure 4(b) displays the variation of the first harmonic Hall voltage signal $V_{1\omega}$ with $H_{x}$ for both $\pm z$-direction magnetizations ($\pm M_{z}$). The variations of the $V_{1\omega}$ with both $H_{x}$ and $H_{y}$ are almost the same. Figures 4(c) and 4(d) illustrate the variations of the second harmonic Hall voltage signal $V_{2\omega}$ with $H_{x}$ and $H_{y}$. The $V_{2\omega}$ signal demonstrates a strong linear correlation with the in-plane fields. Subsequently, we have calculated the damping-like and field-like SOT efficiency ($\xi_{\rm DL}$ and $\xi_{\rm FL})$ of the bilayer using the following formula: \begin{equation} \xi_{\rm{DL(FL)}} =\frac{2e}{\hbar}\frac{\mu_{0}H_{\rm{DL(FL)}} M_{\rm s} t_{\rm FM}}{J_0}, \tag {2} \end{equation} where $e$ is the electron charge, $\hbar$ is the reduced Planck constant, and $t_{\rm FM}$ is the thickness of the FM layer. We obtain $\xi_{\rm DL}=0.042\pm 0.006$ and $\xi_{\rm FL}=0.059\pm 0.002$, with the former being comparable to that in the ${\rm L1_{0}}$-MnGa/B2-CoGa bilayer and the latter being significantly smaller. Previous studies on ${\rm L1_{0}}$-MnGa/B2-CoGa and CoFeB/B2-CoGa revealed that the $\xi_{\rm FL}$ of the former (0.27) is significantly larger than that in the latter (0.03), and this difference was attributed to the sensitivity of $\xi_{\rm FL}$ to the interfacial band matching.[45] In our experiment, we infer that the reduced $\xi_{\rm FL}$ may also result from differences in interfacial band matching with ${\rm L1_{0}}$-MnAl/B2-CoGa and ${\rm L1_{0}}$-MnGa/B2-CoGa, which relies on further investigation. The realization of excellent PMA and SOT-induced deterministic magnetization switching makes the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer an outstanding candidate for the electrode of SOT-MRAM.
cpl-41-5-057503-fig4.png
Fig. 4. (a) Schematic diagram of the HHVR measurement. (b) Variation of the first harmonic Hall voltage signal $V_{1\omega}$ with $\mu_{0}H_{x}$. [(c), (d)] Variations of the second harmonic Hall voltage signal $V_{2\omega}$ with $\mu_{0}H_{x}$ and $\mu_{0}H_{y}$, respectively.
Table 1. The lattice constants of each layer and corresponding mismatch with ${\rm L1_{0}}$-MnAl.
Layer Space group $a$ (Å) $f$ (%)
GaAs $F\bar{4}3m$ 5.65[51] 2.0
B2-CoGa $Pm\bar{3}m$ 2.86[47] 3.1
Co$_{2}$MnGa $Fm\bar{3}m$ 5.77[51] 4.0
${\rm L1_{0}}$-MnAl $P4/mmm$ 2.77[48] –
Furthermore, we designed an ${\rm L1_{0}}$-MnAl-based multilayer employing IEC to realize field-free SOT-induced magnetization switching. The specific stacking structure comprises Al/${\rm L1_{0}}$-MnAl(2)/B2-CoGa(3)/Co$_{2}$MnGa(1.5) [Fig. 5(a)]. As shown in Table 1, the Co-based full Heusler alloy Co$_{2}$MnGa, belonging to the $Fm\bar{3}m$ space group with a lattice constant of 5.77 Å,[51] exhibits small lattice-mismatch with both GaAs, B2-CoGa, and ${\rm L1_{0}}$-MnAl, which is capable of achieving epitaxial growth and providing IEC. The AHE curve, as shown in Fig. 5(b), displays a superposition of IMA and PMA. Notably, Co$_{2}$MnGa, as a Weyl semimetal, may also contribute spin currents with an effective SOT efficiency sign opposite to that of B2-CoGa.[52] To quantitatively estimate the SOT efficiency of the Co$_{2}$MnGa and CoGa layers in our study, we have utilized the following formula:[53] \begin{equation} J_{\rm s} (t_{\rm SCG})/J_{\rm s} (t_{\rm SCG} =\infty)=1-{\rm sech}(t_{\rm SCG} /\lambda_{\rm s}), \tag {3} \end{equation} where $t_{\rm SCG}$ and $\lambda_{\rm s}$ represent the thickness and spin relaxation length of the SCG layer, respectively. By integrating the known SOT efficiencies of CoGa ($t_{\rm CoGa}=5$ nm, $\xi_{\rm DL}=0.042$) and Co$_{2}$MnGa ($t_{\rm Co_{2}MnGa}=10$ nm, $\xi_{\rm DL}=-0.052$),[54] we have estimated the SOT efficiency of 1.5-nm-thick Co$_{2}$MnGa as $-0.006$, assuming $\lambda_{\rm s}$ to be 3 nm.[55] For the 3-nm-thick CoGa layer, the efficiency has been determined to be 0.023. Moreover, theoretical studies have demonstrated that $\sigma_{z}$ can be generated in the IMA layer when the direction of the applied current is parallel to its magnetic moment.[26] In our experiments, $\sigma_{z}$ may be generated in the Co$_{2}$MnGa layer and have an impact on field-free SOT-induced magnetization switching. Since $\sigma_{z}$ is expected to attenuate after passing through the CoGa layer with strong SOC, the actual effect of $\sigma_{z}$ on ${\rm L1_{0}}$-MnAl would be relatively minor. Therefore, we propose that the IEC primarily facilitates field-free SOT-induced magnetization switching in our structure while the $\sigma_{z}$ plays a secondary role. As presented in Fig. 5(c), we have measured the SOT-induced magnetization switching behavior in the multilayer with an in-plane magnetic field changing from 50 to 0 mT. Initially, the sample was pre-magnetized (pre-mag.) with a 50 mT magnetic field applied along the $+x$ direction ($H_{x})$, and then $H_{x}$ was gradually reduced to 0. During the process, Co$_{2}$MnGa remains magnetized in the $+x$ direction, providing an equivalent magnetic field along the $-x$ direction to ${\rm L1_{0}}$-MnAl via the antiferromagnetic IEC. As shown in Fig. 5(c), the switching ratio decreases to 0 when the applied magnetic field reaches 1 mT, indicating that the $H_{x}$ is compensated with the IEC field completely.[14,56] Under varying $H_{x}$, the $\sigma_{z}$ generated in Co$_{2}$MnGa is consistently present and remains basically unchanged, as it is solely dependent on $M$ and $J$ of Co$_{2}$MnGa,[26] yet the $\sigma_{z}$ is not strong enough to independently drive the magnetization switching. Consequently, we have determined the magnitude of the IEC field to be 1 mT.
cpl-41-5-057503-fig5.png
Fig. 5. [(a), (b)] Stacking structure and AHE curve of the multilayer, respectively. (c) SOT switching curves of the samples with in-plane magnetic field $H_{x}$ changing from 50 to 0 mT.
Additionally, since the multilayer designed in our experiment incorporates CoGa and Co$_{2}$MnGa with SOT efficiencies of opposite signs, it is expected that their spin currents will be neutralized entirely through the adjustment of the relative thicknesses of the two layers in subsequent experiments, thereby enabling the study of the competing spin-current mechanism.[19] Increasing the resistivity difference between the IMA and NM layers by modulating their element compositions may facilitate the realization of spin-orbit procession at the interface.[57] In addition, although the SHE strength is weak due to the limited thickness of Co$_{2}$MnGa in our experiment, increasing its thickness is expected to realize the spin-Hall procession (SHP), as the $\sigma_{z}$ generated by the SHP correlates with the SHE in the IMA layer.[58] Moreover, expanding our material system by substituting alloys with comparable lattice constants (e.g., CoGa-CoAl, Co$_{2}$MnGa-Co$_{2}$MnAl) could enable us to investigate additional phenomena. All these designs will advance the understanding of the above critical mechanisms in the ${\rm L1_{0}}$-MnAl-based material system. In summary, we have systematically investigated the structural and magnetic properties as well as the SOT-induced magnetization switching behaviors in the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer. Furthermore, we have demonstrated an ${\rm L1_{0}}$-MnAl-based stacking structure for field-free SOT-induced magnetization switching. By utilizing the B2-CoGa buffer layer, we have achieved large PMA constants $K_{\rm u}$ in 2-nm-thick ${\rm L1_{0}}$-MnAl film, capable of supporting an MTJ size of 6.21 nm in diameter. Subsequently, SOT-induced deterministic magnetization switching in the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer has been achieved, with an exploration into the properties of the switching behaviors under in-plane magnetic fields. The damping-like and field-like SOT efficiencies in the ${\rm L1_{0}}$-MnAl/B2-CoGa bilayer have been ascertained through HHVR measurement, respectively. Furthermore, an ${\rm L1_{0}}$-MnAl/B2-CoGa/Co$_{2}$MnGa multilayer structure has been designed for field-free SOT-induced magnetization switching. These findings offer an effective approach for developing high-performance SOT-MRAM based on ${\rm L1_{0}}$-MnAl. Acknowledgement. The work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDB44000000). References Integration of high-performance spin-orbit torque MRAM devices by 200-mm-wafer manufacturing platformNAND-SPIN-based processing-in-MRAM architecture for convolutional neural network accelerationSpin-transfer torque magnetic random access memory (STT-MRAM)Experimental Demonstration of STT-MRAM-based Nonvolatile Instantly On/Off System for IoT Applications: Case StudiesUltra High-Density SOT-MRAM Design for Last-Level On-Chip Cache ApplicationRobust magnetic field-free switching of a perpendicularly magnetized free layer for SOT-MRAMSpin-Torque Switching with the Giant Spin Hall Effect of TantalumSpin Hall Effect in the Presence of Spin DiffusionSpin Hall EffectSpin polarization of conduction electrons induced by electric current in two-dimensional asymmetric electron systemsCurrent-Induced Switching of Perpendicularly Magnetized Magnetic Layers Using Spin Torque from the Spin Hall EffectCentral role of domain wall depinning for perpendicular magnetization switching driven by spin torque from the spin Hall effectSpin–orbit torque switching without an external field using interlayer exchange couplingAdjustable Current-Induced Magnetization Switching Utilizing Interlayer Exchange CouplingCurrent-Induced Spin-Orbit Torque and Field-Free Switching in Mo -Based Magnetic HeterostructuresVoltage-gated field-free spin–orbit torque switching in Pt/Co/Ir/MgO wedged structuresMagnetization switching by spin–orbit torque in an antiferromagnet–ferromagnet bilayer systemFormation of helical spin alignment in the AFM/FM/AFM trilayers by spin–orbit torque controlled exchange biasSwitching a Perpendicular Ferromagnetic Layer by Competing Spin CurrentsNonlinear spin current generation in noncentrosymmetric spin-orbit coupled systemsField-free magnetization switching in CoPt induced by noncollinear antiferromagnetic Mn 3 Ga Efficient and controllable magnetization switching induced by intermixing-enhanced bulk spin–orbit torque in ferromagnetic multilayersField-free switching of a perpendicular magnetic tunnel junction through the interplay of spin–orbit and spin-transfer torquesCluster magnetic octupole induced out-of-plane spin polarization in antiperovskite antiferromagnetGiant field-like torque by the out-of-plane magnetic spin Hall effect in a topological antiferromagnetAnomalous spin Hall and inverse spin Hall effects in magnetic systemsMultifunctional L10-Mn1.5Ga Films with Ultrahigh Coercivity, Giant Perpendicular Magnetocrystalline Anisotropy and Large Magnetic Energy ProductNanostructured D022 -Mn3 Ga with high coercivitySaturation magnetization and crystalline anisotropy calculations for MnAl permanent magnetPerpendicularly magnetized τ -MnAl (001) thin films epitaxied on GaAsStructural characterization and magnetic properties of perpendicularly magnetized MnAl films grown by molecular-beam epitaxyUltrafast Magnetization Precession in Perpendicularly Magnetized L10 -MnAl Thin Films with Co2 MnSi Buffer LayersElectronic Structure and Magnetocrystalline Anisotropy Energy of MnAlLong-Lived Ultrafast Spin Precession in Manganese Alloys Films with a Large Perpendicular Magnetic AnisotropyElectronic structures and magnetism of CuAu-type MnNi and MnGaElectronic structure and magnetic properties of δ-MnGaFirst-principles study of MnAl for its application in MgO-based perpendicular magnetic tunnel junctionsCurrent-induced spin–orbit torque magnetization switching in a MnGa/Pt film with a perpendicular magnetic anisotropyModulated switching current density and spin-orbit torques in MnGa/Ta films with inserting ferromagnetic layersManipulation of perpendicular magnetic anisotropy and spin–orbit torque switching behavior in ferrimagnetic D 022-Mn3Ga based multilayersSpin-orbit torque induced magnetization switching in ferrimagnetic Heusler alloy D 22-Mn3Ga with large perpendicular magnetic anisotropyEnhanced spin-orbit torques in MnAl/Ta films with improving chemical orderingHighly (001) oriented MnAl thin film fabricated on CoGa buffer layerRoom temperature growth of ultrathin ordered MnGa films on a CoGa buffer layerSpin Hall effect from hybridized 3 d − 4 p orbitalsIn-plane current-induced magnetization switching in CoGa/MnGa/MgO filmsAll-optical probe of sub-THz spin precession in a L 10 MnGa nanolayerEpitaxial ferromagnetic τ‐MnAl films on GaAsNanometer-thin L 1-MnAl film with B 2-CoAl underlayer for high-speed and high-density STT-MRAM: Structure and magnetic propertiesEstimating spin Hall angle in heavy metal/ferromagnet heterostructuresInterfacial, electrical, and spin-injection properties of epitaxial Co2MnGa grown on GaAs(100)Magnetization switching induced by spin–orbit torque from Co2 MnGa magnetic Weyl semimetal thin filmsSpin-diffusion lengths in metals and alloys, and spin-flipping at metal/metal interfaces: an experimentalist’s critical reviewSpin-generation in magnetic Weyl semimetal Co2MnGa across varying degree of chemical orderGigantic Anisotropy of Self-Induced Spin-Orbit Torque in Weyl Ferromagnet Co2 MnGaEnhanced interlayer Dzyaloshinskii–Moriya interaction and field-free switching in magnetic trilayers with orthogonal magnetizationSpin currents and spin–orbit torques in ferromagnetic trilayersGiant charge-to-spin conversion in ferromagnet via spin-orbit coupling
[1] Zhang H C, Ma X Y, Jiang C P, Yin J L, Lyu S Q, Lu S Y, Shang X T, Man B W, Zhang C, Li D D, Li S H, Chen W J, Liu H X, Wang G F, Cao K H, Wang Z H, and Zhao W S 2022 J. Semicond. 43 102501
[2] Zhao Y, Yang J, Li B, Cheng X, Ye X, Wang X, Jia X, Wang Z, Zhang Y, and Zhao W 2023 Sci. Chin. Inf. Sci. 66 142401
[3] Apalkov D, Khvalkovskiy A, Watts S, Nikitin V, Tang X, Lottis D, Moon K, Luo X, Chen E, Ong A, Driskill-Smith A, and Krounbi M 2013 ACM J. Emerging Technol. Comput. Syst. 9 13
[4] Li Y, Kang W, Zhou K, Qiu K, and Zhao W 2023 ACM Trans. Embed. Comput. Syst. 22 29
[5] Seo Y and Kwon K W 2023 Electronics 12 4223
[6] de Orio R L, Makarov A, Selberherr S, Goes W, Ender J, Fiorentini S, and Sverdlov V 2020 Solid-State Electron. 168 107730
[7] Liu L, Pai C F, Li Y, Tseng H W, Ralph D C, and Buhrman R A 2012 Science 336 555
[8] Zhang S 2000 Phys. Rev. Lett. 85 393
[9] Hirsch J E 1999 Phys. Rev. Lett. 83 1834
[10] Edelstein V M 1990 Solid State Commun. 73 233
[11] Liu L, Lee O J, Gudmundsen T J, Ralph D C, and Buhrman R A 2012 Phys. Rev. Lett. 109 096602
[12] Lee O J, Liu L Q, Pai C F, Li Y, Tseng H W, Gowtham P G, Park J P, Ralph D C, and Buhrman R A 2014 Phys. Rev. B 89 024418
[13] Lau Y C, Betto D, Rode K, Coey J M D, and Stamenov P 2016 Nat. Nanotechnol. 11 758
[14] Sheng Y, Edmonds K W, Ma X, Zheng H, and Wang K 2018 Adv. Electron. Mater. 4 1800224
[15] Chen T Y, Chan H I, Liao W B, and Pai C F 2018 Phys. Rev. Appl. 10 044038
[16] Li Y, Zhao X, Liu W, Wu J, Liu L, Song Y, Ma J, and Zhang Z 2023 Appl. Phys. Lett. 123 032403
[17] Fukami S, Zhang C, DuttaGupta S, Kurenkov A, and Ohno H 2016 Nat. Mater. 15 535
[18] Choi W C, Yoon S, Kim H J, Ha J H, Park K J, Baek E, Kim D R, Shin Y, You C Y, Kim J W, and Hong J I 2023 APL Mater. 11 121115
[19] Ma Q, Li Y, Gopman D B, Kabanov Y P, Shull R D, and Chien C L 2018 Phys. Rev. Lett. 120 117703
[20] Hamamoto K, Ezawa M, Kim K W, Morimoto T, and Nagaosa N 2017 Phys. Rev. B 95 224430
[21] Han R K, Zhao X P, Qin H R, Sun H L, Wang H L, Wei D H, and Zhao J H 2023 Phys. Rev. B 107 134422
[22] Zhang K, Chen L, Zhang Y, Hong B, He Y, Lin K, Zhang Z, Zheng Z, Feng X, Zhang Y, Otani Y, and Zhao W 2022 Appl. Phys. Rev. 9 011407
[23] Wang M X, Cai W L, Zhu D Q, Wang Z H, Kan J, Zhao Z Y, Cao K H, Wang Z L, Zhang Y G, Zhang T R, Park C, Wang J P, Fert A, and Zhao W S 2018 Nat. Electron. 1 582
[24] You Y, Bai H, Feng X, Fan X, Han L, Zhou X, Zhou Y, Zhang R, Chen T, Pan F, and Song C 2021 Nat. Commun. 12 6524
[25] Kondou K, Chen H, Tomita T, Ikhlas M, Higo T, MacDonald A H, Nakatsuji S, and Otani Y 2021 Nat. Commun. 12 6491
[26] Wang X R 2021 Commun. Phys. 4 55
[27] Zhu L J, Nie S H, Meng K K, Pan D, Zhao J H, and Zheng H Z 2012 Adv. Mater. 24 4547
[28] El-Gendy A A and Hadjipanayis G 2015 J. Phys. D 48 125001
[29] Park J H, Hong Y K, Bae S, Lee J J, Jalli J, Abo G S, Neveu N, Kim S G, Choi C J, and Lee J G 2010 J. Appl. Phys. 107 09A731
[30] Nie S H, Zhu L J, Lu J, Pan D, Wang H L, Yu X Z, Xiao J X, and Zhao J H 2013 Appl. Phys. Lett. 102 152405
[31] Nie S H, Zhu L J, Pan D, Lu J, and Zhao J H 2013 Acta Phys. Sin. 62 178103 (in Chinese)
[32] Mao S W, Lu J, Yang L, Ruan X Z, Wang H L, Wei D H, Xu Y B, and Zhao J H 2020 Chin. Phys. Lett. 37 058501
[33] Sakuma A 1994 J. Phys. Soc. Jpn. 63 1422
[34] Mizukami S, Wu F, Sakuma A, Walowski J, Watanabe D, Kubota T, Zhang X, Naganuma H, Oogane M, Ando Y, and Miyazaki T 2011 Phys. Rev. Lett. 106 117201
[35] Sakuma A 1998 J. Magn. Magn. Mater. 187 105
[36] Yang Z X, Li J, Wang D s, Zhang K M, and Xie X D 1998 J. Magn. Magn. Mater. 182 369
[37] Zhang X, Tao L L, Zhang J, Liang S H, Jiang L, and Han X F 2017 Appl. Phys. Lett. 110 252403
[38] Ranjbar R, Suzuki K Z, Sasaki Y, Bainsla L, and Mizukami S 2016 Jpn. J. Appl. Phys. 55 120302
[39] Meng K, Miao J, Xu X, Wu Y, Xiao J, Zhao J, and Jiang Y 2016 Sci. Rep. 6 38375
[40] Zhao X P, Sun H L, Tong S C, Han R K, Qin H R, and Zhao J H 2023 Appl. Phys. Lett. 123 042407
[41] Zhao X P, Lu J, Mao S W, Yu Z F, Wei D H, and Zhao J H 2019 Appl. Phys. Lett. 115 142405
[42] Meng K K, Miao J, Xu X G, Wu Y, Zhao X P, Zhao J H, and Jiang Y 2017 Appl. Phys. Lett. 110 142401
[43] Oshima D, Kato T, and Iwata S 2020 AIP Adv. 10 025012
[44] Suzuki K Z, Ranjbar R, Sugihara A, Miyazaki T, and Mizukami S 2016 Jpn. J. Appl. Phys. 55 010305
[45] Lau Y C, Lee H, Qu G X, Nakamura K, and Hayashi M 2019 Phys. Rev. B 99 064410
[46] Takikawa M, Suzuki K Z, Ranjbar R, and Mizukami S 2017 Appl. Phys. Express 10 073004
[47] Mizukami S, Suzuki K Z, and Miura Y 2019 Appl. Phys. Express 12 043003
[48] Sands T, Harbison J P, Leadbeater M L, Allen S J, Hull G W, Ramesh R, and Keramidas V G 1990 Appl. Phys. Lett. 57 2609
[49] Takeuchi Y, Okuda R, Igarashi J, Jinnai B, Saino T, Ikeda S, Fukami S, and Ohno H 2022 Appl. Phys. Lett. 120 052404
[50] Deng Y, Yang M, Ji Y, and Wang K 2020 J. Magn. Magn. Mater. 496 165920
[51] Damsgaard C D, Hickey M C, Holmes S N, Feidenhans'l R, Mariager S O, Jacobsen C S, and Hansen J B 2009 J. Appl. Phys. 105 124502
[52] Tang K, Wen Z, Lau Y C, Sukegawa H, Seki T, and Mitani S 2021 Appl. Phys. Lett. 118 062402
[53] Bass J and Pratt W P 2007 J. Phys.: Condens. Matter 19 183201
[54] Safi T S, Chou C T, Hou J T, Han J, and Liu L 2022 Appl. Phys. Lett. 121 092404
[55] Aoki M, Yin Y, Granville S, Zhang Y, Medhekar N V, Leiva L, Ohshima R, Ando Y, and Shiraishi M 2023 Nano Lett. 23 6951
[56] Zhao X, Sun H, Han R, Qin H, Wen L, Wang H, Wei D, and Zhao J 2024 APL Mater. 12 041103
[57] Baek S C, Amin V P, Oh Y W, Go G, Lee S J, Lee G H, Kim K J, Stiles M D, Park B G, and Lee K J 2018 Nat. Mater. 17 509
[58] Hibino Y, Taniguchi T, Yakushiji K, Fukushima A, Kubota H, and Yuasa S 2021 Nat. Commun. 12 6254