Giant Spin-Orbit Magnetic State Readout Enhanced by a Magnetic Tunnel Junction

Giant Spin-Orbit Magnetic State Readout Enhanced by a Magnetic Tunnel Junction | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Giant Spin-Orbit Magnetic State Readout Enhanced by a Magnetic Tunnel Junction Weisheng Zhao, Yan Huang, Kun Zhang, Guo Liu, Xiaobai Ning, Shiyang Lu, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7680431/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Magnetoelectric spin-orbit (MESO) logic, composed of a voltage-controlled magnetoelectric writing module and a spin-orbit readout module, is highly expected to substitute the silicon-based transistors and enables energy-efficient and scalable computing. Nevertheless, the output voltage of readout module based on spin-to-charge conversion is far less than the minimum magnetoelectric writing voltage, which greatly restricts the cascading function of MESO logic. Here, we first propose a magnetic tunnel junction (MTJ)-enhanced MESO logic to implement giant readout signal. Up to 1.5 mV output voltage is obtained, marking a significant improvement of approximately two orders of magnitude compared to previous findings. We ascribe the substantial enhancement to junction resistance modulation and the spin filtering effect of MgO-single-crystal MTJ. Moreover, the naturally integrated MTJ and MESO enables instantaneous and nonvolatile data exchange between computing module and external unit. Our work not only enhances output signal of readout module for direct cascading of MESO logic but also refines the design architecture, marking a pivotal stride forward in propelling MESO technology toward practical applications. Physical sciences/Nanoscience and technology/Nanoscale devices/Electronic devices Physical sciences/Nanoscience and technology/Nanoscale devices/Magnetic devices Magnetoelectric spin-orbit logic magnetic tunnel junction spin-to-charge conversion spin-orbit coupling spin polarization Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction With technology nodes shrinking, conventional complementary metal-oxide-semiconductor (CMOS) technology faces a growing hurdle in scaling down due to leakage current in the OFF state, imposing a limit on the overall system energy efficiency 1 , 2 . As a solution, emerging nonvolatile memories, featuring normally-OFF/instantly-ON computation, can overcome the static power brought by leakage current 3 – 7 . However, because these technologies are not supported to generate an effective driving force, they are hard to drive the subsequent data writing and construct cascaded logic circuits 2 , 7 . Magnetoelectric spin-orbit (MESO) logic, utilizing voltage-controlled magnetoelectric (ME) effect for data writing and spin-orbit coupling (SOC) for magnetic state reading, can be directly cascaded and is regarded as a compelling alternative to CMOS toward 100-mV logic 8 , 9 . In a MESO logic, a single magnetic layer is used to store bit data, as well as couple writing block (ME block) and reading block (SO block) 8 , 10 . To date, the quasi-static ferroelectric switching of ME writing module has been achieved down to about 150 mV using ultrathin La-doped BiFeO 3 , with a pathway to get down to 100 mV 8,11,12 . However, there is a notable challenge that the spin-to-charge conversion (SCC) voltage of readout module is far less than 100 mV, which greatly restricts the cascading function of MESO logic. Typically, this signal measures in the tens or hundreds of milliohms, which results in an output voltage of less than 20 µV 7,13 . The stark discrepancy between the voltages required for writing and those available for reading presents a wide gap to the practical implementation of MESO logics. Although it is predicted using emerging quantum materials such as topological insulators (TIs) and two-dimensional electron gases (2DEGs) could enlarge the SCC signal due to their high SCC efficiency 14 – 19 , experimental results at room temperature (R.T.), to date, have not met the anticipated expectations 20 , 21 . In addition, the processing compatibility for massive production of these quantum materials has not been totally addressed. Therefore, more brand-new enhancement mechanism of output signal of MESO, different from utilizing high SOC materials, is highly desired to promote the MESO technology towards practical applications. Here, we propose for the first time the integration of magnetic tunnel junction (MTJ) and MESO by sharing one ferromagnetic (FM) layer as illustrated in Fig. 1 a. On the one hand, the injection current modulated by MTJ can amplify the output voltage. As shown in Fig. 1 b, the injection tunneling current experiences a sudden increase upon the transition of MTJ from antiparallel (AP) to parallel (P) state under a constant voltage supply V supply , i.e. , from I AP to I P . This current alteration can induce an extra voltage signal superposed to the SCC output signal, and amplify the output signal of MESO. On the other hand, the high spin polarization (SP) rate effectuated by the spin filtering mechanism inherent in the single-crystal CoFeB/MgO/CoFeB sandwiched-structure can improve the SCC output voltage. As shown in Fig. 1c , Δ 1 band electrons in sandwiched CoFeB/MgO/CoFeB tunnel easily when the magnetization of two CoFeB layers are parallel alignment, and hard when they are antiparallel, while Δ 2 and Δ 5 band electrons count little 22 – 25 . Consequently, tunneling current across an MTJ could be highly polarized to significantly enhance SP rate. Our experimental findings at R.T. reveal a giant output signal, which is significantly enhanced to an impressive 1.5 mV, about two orders of magnitude greater than that reported in previous literatures 7 – 10 , 13 , 21 . Furthermore, the integrated MTJ can function as a cache memory to provide infinite bandwidth for MESO logic as shown in Fig. 1 d, addressing the data exchange issue of computing unit and external memory from the perspective of architecture level, a problem that has been overlooked in previous researches. The proposed “MESO + MTJ” scheme can enhance the output signal and optimize the design architecture, offering a promising direction for future development of MESO technology. 2. Results 2.1 Structural, magnetic and electrical properties At the outset of our experiments, we examined the impact of the MgO layer thickness on the tunneling magnetoresistance (TMR) ratio, which reveals film quality and SP rate. The functional stacks depicted in Fig. 2 a, consisting of W (3.5)/CoFeB (1.9)/MgO (t)/CoFeB (1.9)/CoFe (0.5)/Ru (0.8)/CoFe (2)/IrMn (7.5)/Ru (5), are deposited on bottom electrode Ta (25)/CuN (20) (numbers in parentheses denotes film thickness in nm) with varying MgO thickness t . After post annealing at 400 ℃, we determined resistance-area product (RA) and TMR by using current in-plane tunneling (CIPT) measurement at R.T.. The results are shown in Fig. 2 b. The RA value increases with increasing MgO thickness, while the TMR ratio exhibits an initial increase followed by a decrease, reaching its maximum when the MgO thickness is 1.1 nm. This optimal TMR value suggests a higher quality of the film and a greater degree of SP 26 . Note that, TMR is expected to appear oscillations as t keeps increasing due to coherent tunneling effect 27 , 28 . We then selected a 1.1 nm MgO layer for subsequent experiments, and deposited MTJ stacks directly on thermally oxidized silicon substrate without Ta (25)/CuN (20). Figure 2 c illustrates the hysteresis loop of full functional stacks deposited on thermally-oxidized silicon substrate using vibrating sample magnetometer. By applying an in-plane magnetic field along x direction, distinct magnetic behaviors were observed. The minor loop inset in Fig. 2 c indicates that the coercive field of the free layer is approximately 10 Oe. After a series of electron beam and ultraviolet lithography, as well as etching steps, we have successfully fabricated a junction device featuring multiple electrodes. Figure 2 d illustrates the schematic of proposed MTJ-enhanced MESO readout block. The device features with functional layers of two FM, a SOC, and a tunneling barrier layers. A cross-shape bottom electrode is used to generate transverse output voltage which can be measured between lead 2 and 3, with power supply across lead 0 and 1 7 . Generally, this reading process can be divided into two distinct phases: the spin polarization phase and the SCC phase. During the spin polarization phase, a constant supply source, denoted as V supply or I supply , is applied perpendicularly to the device stacks and injecting a spin polarized current 7 . The injection efficiency is characterized by SP rate. Typically, SP rates are approximately 0.37 for Ni, 0.45 for Co, and 0.43 for Fe, with NiFe and CoFeB exhibiting higher values of 0.48 and 0.53, respectively 29 – 31 . By applying an external magnetic field and sweeping it along x axis, vertical voltage V 1 (V MTJ ) and transverse voltage V 2 (V xy ) are both supposed to appear abrupt changes when free layer magnetization changes, indicating TMR and SCC output signals. In Fig. 2 e, the upper panel shows the top view a junction with an area about 60 nm × 130 nm, while the lower panel gives the side view of junction in which the MgO barrier can be recognized clearly, indicating good crystallization. TMR and SCC signals are shown in Fig. 2 f. The low resistance state, corresponding to the alignment of the two magnetic layers at P state, is measured to be around R MTJ,P = 4.6 kΩ. Meanwhile, the high resistance state exhibits a resistance of around R MTJ,AP = 8 kΩ, corresponding to AP state. Consequently, TMR ratio is calculated to be 74% according to the equation of TMR = (R MTJ,AP - R MTJ,P )/R MTJ,P . The TMR loss compared with CIPT results can be attributed to damage during nanofabrication. Notably, there are some anomalous points when the external field is around ± 20 Oe, which may be caused by irregular shapes and probable defects and pinnings at the edge of the junction. Meanwhile, we measure transverse voltage V xy versus H x (V xy -H x loop) under a constant voltage supply of 300 mV. As expected, there is abrupt changes in transverse signal V xy , precisely at the points where free layer magnetization reverses. The output voltage ∆V xy is defined as the difference between two magnetic states. By utilizing MTJ modulation and high SP rate, the observed output voltage ∆V xy at 300 mV supply voltage is approximately 1.5 mV. This result is at least two orders larger than those in previous reports (less than 15 µV in ref 7 and 1.5 µV in ref 13 ). 2.2 SCC signal under current source We performed output signal under a constant current source to precisely evaluate the SCC signal without resistance modulation. The SCC result is intrinsically guaranteed by the anomalous spin-polarized velocity arising from a momentum-space Berry phase of Bloch electrons 32 . Reciprocally, charge-to-spin conversion (CSC), the inverse process of the SCC, can also be assessed using the same device. The charge flowing along W channel would cause spin accumulation and diffusion at and cross interfaces, known as direct spin Hall effect (SHE) 32 – 35 , which can be detected by magnetic electrodes 7 , 36 . Through applying current across lead 2 and 4, and measuring voltage between lead 0 and 1, SHE can be evaluated as CSC signal. In Fig. 3 a, we present the SCC curve \(\:\partial\:{V}_{24}/\partial\:{I}_{01}\) and its inverse, CSC curve \(\:\partial\:{V}_{01}/\partial\:{I}_{24}\) at 5 µA supply current. The CSC curve appears with opposite polarity to SCC curve, consistent with previous reports 7 , 36 . This indicates a giant SCC signal 2∆R SCC around 14.5 Ω, which is nearly 50 times larger than previous reports (0.3 Ω) 7 . However, the CSC signal is about 11 Ω, slightly smaller than SCC signal. We attribute this difference to the spin diffusion and relaxation in z-direction, compared with spin drift in SCC situation 7 , 36 . The measurement set-up in Fig. 2 d also brings various Hall signals mixing in the transverse voltage V xy , such as conventional direct Hall effect, planar Hall effect and anomalous Hall effect 37 , 38 . Details are discussed in Supplementary Materials S1 . These Hall signals count less than 0.1 Ω in our nanodot structure, and affect little for the evaluation of final output signal. The spin injection phase and the SCC phase are both theoretically bias dependent. At first, the spin injection facilitated by an MTJ could be modulated by its bias-dependent SP 24 . Moreover, considering the ISHE mechanism in SCC phase, the transverse voltage loop of V 2 shall invert polarity when the supply current direction is reversed, that is, V 2 (-I supply , H x ) = -V 2 (I supply , H x ), resulting no polarity change in resistance versus field curve, that is, R xy (-I supply , H x ) = R xy (I supply , H x ) 39 . As a result, if one obtains R xy (-I supply , H x ) = -R xy (I supply , H x ), this result may be dominated by a current-symmetric term rather than SCC. The second is normally neglected by most researches, but it plays a significant role in SCC signal assessment. Therefore, we measured different SCC signals by applying currents of varying amplitude and direction, to investigate the bias impact. The SCC signals at ± 30 µA appear with the same polarity as expected, but are approximately 2.5 Ω lower than that at 1 µA, as shown in Fig. 3 b. Moreover, the signal at + 30 µA seems larger than signal at -30 µA. A detailed examination reveals a significant degradation in the SCC signal as supply current is increased, as depicted in Fig. 3 c. Interestingly, similar phenomena have been previously reported in W/MgO/CoFeB tri-layer spin-tunneling structure 36 . The maximium SCC signal of 2∆R SCC is around 14.5 Ω at 5 µA supply. Obviously, when the current amplitudes are equal, the signal obtained under negative bias is slightly smaller than that under positive bias. These observations may imply a bias-dependent SP rate. Consequently, we explored the bias current dependence of the junction resistance R MTJ , and transverse resistance R xy . We use R MTJ,P (R MTJ,AP ) and R xy,P (R xy,AP ) to represent the junction and transverse resistances when MTJ is at P (AP) state, respectively. The findings are summarized in Fig. 3 d, where it is noted that while R MTJ,AP exhibits a pronounced variation with bias current, R MTJ,P remains relatively unchanged. For the transverse resistance R xy , both R xy,P and R xy,AP exhibit monotonically decreasing trend with the increasing supply current, suggesting an additional mechanism between the SCC and TMR signals. Generally, the SCC signal in our MTJ-enhanced readout module contain two components of information, i.e. energetic electrons contribution from reference layer and equilibrium electrons mostly from free layer 36 . The energetic electrons are more energy-sensitive and thus attribute to the bias-dependent signal. These analyses collectively suggest that the utilization of highly spin-polarized electrons tunneling from CoFeB/MgO/CoFeB structures holds greater promise than those from a single FM layer in the development of spin-injection-based devices. 2.3 MTJ-enhanced output signals under voltage source Then, we compare the output voltage ∆V xy under current source and voltage source. When the applied current source is substituted with a constant voltage V supply , the injection current would experience a sudden decrease upon the transition of MTJ from P to AP state.This alteration consequently affects the output voltage ∆V xy . As illustrated in the inset of Fig. 4 a, the maximum observed output voltage ∆V xy is approximately 1.5 mV corresponding to a 300 mV supply voltage, while the largest ∆V xy for current source is about 0.4 mV for 40 µA supply. For a supply voltage of 300 mV, the tunneling current under the P state measures approximately 65 µA. Given that ∆V xy is proportional to the amplitude of the injection tunneling current, under a supply current of 65 µA, ∆V xy is estimated to be around 0.65 mV, also significantly lower than 1.5 mV. These findings effectively validate the regulation impact of TMR effect when operates under a voltage source. A comprehensive list of output voltages for various applied currents and voltages is provided in Fig. 4 a. It is observed that larger applied source generates larger output voltage, and the voltage source can yield larger output voltage than current source when the current flow are comparable. Moreover, we could calculate the transverse resistance difference under voltage source by the formula, ∆R xy = R xy,P - R xy,AP = V xy,P /I P - V xy,AP /I AP . As shown in Fig. 4 b, the relationship between ∆R xy and voltage supply exhibits no regularity, quite different from that depicted in Fig. 3 c, meaning an extra term brought by TMR signal, in agreement with our analysis. To clarify the regulation impact of MTJ, we calculate the relationship between ∆V xy and MTJ parameters under a voltage source as follows, $$\:\varDelta\:{\text{V}}_{\text{x}\text{y}}={\text{V}}_{\text{x}\text{y},\text{P}}-{\text{V}}_{\text{x}\text{y},\text{A}\text{P}}=[\left(1+\text{T}\text{M}\text{R}\right){\text{R}}_{\text{x}\text{y},\text{P}}-{\text{V}}_{\text{x}\text{y},\text{A}\text{P}}]\frac{{\text{V}}_{\text{s}\text{u}\text{p}\text{p}\text{l}\text{y}}}{{\text{R}}_{\text{M}\text{T}\text{J},\text{A}\text{P}}}\:$$ 1 The detailed formular derivation process can be found in Supplementary Material S2 . For an MTJ, the parameter I AP ​ is directly determined by the thickness of the MgO barrier. Meanwhile, the TMR ratio critically depends on the overall quality of the MTJ structure. Therefore, to achieve a substantial increase in ∆V xy under a constant voltage source, it is essential to optimize two key factors: 1) the MgO barrier must be sufficiently thin to enhance tunneling efficiency, and reduce R MTJ,AP ; 2) the crystalline quality of the MTJ must be carefully optimized to ensure large TMR ratio. Figure 4 c lists some impressive progress in MESO-like readout modules reported so far, containing different SOC materials (heavy metals, TIs and 2DEGs) 7 , 10 , 13 , 20 , 21 , 40 , 41 . Some of them provide output resistance signal ∆R xy with a given current I supply , but no output voltages ∆V xy . As ∆V xy is more critical and essential for practical application, We derive their voltage signals as ∆V xy = I supply ×∆R xy and make a comparison. Although TIs and 2DEGs may remain theoritical superiority thanks to topological surface states and Rashba surface, experimental outputs have no advantages compared with W or Pt based device. One possible reason is their topological state are not stable at R.T.. W-based device outputs 3-times larger than Pt-based device as expected, when they are patterned with same size 41 . Interestingly, Pt-based SO device has been reported the most, yet its results vary greatly. The most possible reason is Pt’s resistivity in reports varies a lot, from tens of µΩ·cm to hundreds of µΩ·cm 7 . Among all of these proposals, our work not only outperforms all others by a significant margin, but also enhances the output voltage from the microvolt level to the millivolt level, marking a new record. 2.3 Mechanism Explorations In this part, we explored possible mechanisms and factors to induce the giant readout signal. At first, we compared the output transverse signal in proposed MTJ-enhanced device and traditional FM/SOC-bilayer device. We designed some control experiments using microdot devices with the same channel width of 3 µm. The output signals at I supply = 10 µA of these two samples are shown in Fig. 5 a. Sample 1 is a bilayer device composed of W (6)/CoFeB (5), yielding about 2.5 mΩ output signal. Sample 2 is a MTJ-enhanced device with the above-mentioned film stack structure, which presents approximately 1.2 Ω signal. This difference is quite reasonable according to the numerical model 7 : 1) the SP rates of sample 2 is about several times larger that of sample1; 2) Because there is a phase change when the thickness of W changes from 6 nm to 3.5 nm, the W resistivity in sample 1 is about 60 µΩ·cm and that in sample 2 is about 367 µΩ·cm; 3) The θ SH decreases for W from 0.26 at 3.5 nm to 0.04 at 6 nm 42,43 . The characterization of θ SH of W are presented in Supplementary Materials S3 . These factors give about hundreds times of difference between these two scenarios, as depicted in Fig. 5 b, quite matching the experiments. More details about the numerical model are presented in Supplementary Materials S4. To further explain the point 1) about SP rates, we established W/CoFe/W and W/CoFe/MgO/CoFe/W atomic-level structure to perform calculation as shown in Fig. 5 c and 5 d. K-resolved transmission spectra in Fig. 5 e and 5 f illustrate the majority-to-majority tunneling and minority-to-minority tunneling in W/CoFe/W, both generally spreading over the Brillouin zone. In contrast, majority-to-majority tunneling in W/CoFe/MgO/CoFe/W (P state) is concentrated within a square region around Γ (Fig. 5 g), vanishing in the minority-to-minority spectrum (Fig. 5 h). Given the definition of spin polarization rate \(\:\text{S}\text{P}=({\text{G}}_{\uparrow\:\uparrow\:}^{\uparrow\:}-{\text{G}}_{\uparrow\:\uparrow\:}^{\downarrow\:})/({\text{G}}_{\uparrow\:\uparrow\:}^{\uparrow\:}+{\text{G}}_{\uparrow\:\uparrow\:}^{\downarrow\:})\) , where \(\:{G}_{\uparrow\:\uparrow\:}^{\uparrow\:}\) is the majority-to-majority spin conductance and \(\:{G}_{\uparrow\:\uparrow\:}^{\downarrow\:}\) is the minority-to-minority spin conductance 44 , the calculated value gives about 92% SP rate for W/CoFe/MgO/CoFe/W structure, about 4 times larger than that for W/CoFe/W. It is evident that several monolayer MgO acts as spin filter to enlarge the difference between majority and minority spin transmission. 3. Conclusion and Outlook Our proposed "MESO + MTJ" scheme effectively overcomes key bottlenecks in MESO logic systems by simultaneously amplifying output signals and redefining data-exchange architectures. This dual advancement manifests in two pivotal aspects: On the one hand, the output voltage can be further improved to 100 mV to enable the direct cascading of multiple MESO devices. According to Eq. 1 , output voltage conversion efficiency ∆V xy /V supply is highly related to MTJ properties. With higher TMR, and lower MTJ resistance at AP state, comes higher conversion efficiency. With proper stack design and interface optimization, TMR at R.T. of MgO-single-crystal MTJ can be increased high up to 631% 28 . Meanwhile, high TMR ratio means high spin polarization of injection current. By adapting strong SOC material and scaling down the device size, the SCC process can be further improved. These methods are promising to improve the output voltages to the order of 100 mV. On the other hand, the "MESO + MTJ" design addresses the critical issue of data exchange between computing unit and peripheral memory from system level. A one-bit full adder composed of "MESO + MTJ" units is schematically shown in Supplementary Fig. S3 of Supplementary Materials . This scheme can operate as a logic unit with a intrinsically-integrated high-speed cache memory to resolve the bandwidth mismatch between logic unit and external memory 45 . Moreover, this innovative design enables the extraction of any intermediate process results while simultaneously monitoring its resistance state during computation, which can improve the computation efficiency and relibility. In conclusion, our proposal offers a viable pathway toward advancing the real-world implementation of MESO logic technology. Methods Film growth and characterization : All the films used in this work were deposited on 8-inch thermally oxidized silicon substrate, using ultrahigh vacuum dc/rf magnetron sputtering system. The substrate was first pre-cleaned for 60 s by Ar + ion beam. The CoFe, CoFeB, and IrMn denote Co 70 Fe 30 , Co 60 Fe 20 B 20 , and Ir 20 Mn 80 alloy with nominal target compositions, respectively. After deposition, the stacks were in-situ annealed at 400 ℃ for 1 h under an external field of 1 T along x direction to define the reference layer direction and form highly oriented crystalline CoFeB/MgO/CoFeB MTJs. The static in-plane anisotropy magnetization behavior was examined using a vibrating sample magnetometer system. Device fabrication : For a nanoscale MTJ-enhanced spin-orbit device, we first patterned 500 nm × 2.5 µm cross-shape channels using standard electron beam lithography (EBL) and Ar + ion beam etching (IBE). During milling, we monitored the secondary-ion mass spectra. After that, a second round of EBL and IBE procedures were used to define elliptical junctions. Thanks to high-resolution and high-aspect-ratio negative tone photoresist, we successfully formed a photoresist pillar with tens-nanometer length and hundreds-nanometer height. The junction etching was divided into two phases: first, we defined its appearance by small angle etching until the bottom 3.5-nm-thick W signal appears; second, large angle etching was utilized to meticulously clean the sidewall re-depositions across the barrier. Rather than removing the photoresist immediately, we covered the pillars with a 50-nm-thick silicon nitride (Si 3 N 4 ) passivation layer, by using chemical vapor deposition, and then followed by lift-off process. Vias through bottom electrodes were formed by standard ultraviolet lithography and inductively coupled plasma etching. Ti (10nm)/ Au (100nm) electrodes are finally formed by E-beam evaporation (EBE). For the microscale SO device, i.e. , CoFeB/W readout device and CoFeB/MgO/CoFeB/W MTJ-enhanced readout device, standard ultraviolet lithography and etching process were used to pattern the bottom electrode and the microdot pillar. Then, 50-nm-thick Si 3 N 4 passivation layer covered the pillars, and followed by lift-off process. Finally, Ti (10 nm)/Au (100 nm) electrodes are formed by EBE. For bilayer ST-FMR measurement, the device was formed by standard ultraviolet lithography and etching process, with Ti(10 nm)/ Au(100 nm) electrodes using EBE and lift-off process. Device characterization : Electronic transport measurements of MTJ-enhanced readout device were conducted on our self-developed probe station, using a Keithley 2182 nanovoltmeter and a Keysight B1500A semiconductor device parameter analyzer under varying in-plane magnetic field. The B1500A is used for direct resistance measurement while 2182 nanovoltmeter is more sensitive and used for transverse voltage measurement. ST-FMR measurement of the bilayer heterojunction was performed using a Keysight MXG N5183B analog signal generator and SR830 lock-in amplifier. ab-initio simulation : The atomic structures are relaxed by Vienna ab-initio simulation package (VASP) so that the residual forces are minimized under 0.01 eV/Å 46 . Next, transport calculations are performed by NanoDCAL, which is based on density functional theory (DFT) and Keldysh non-equilibrium Greens function (NEGF) 47 . After a self-consistent calculation of 20×20×1 k-point mesh, the spin-resolved conductance calculation of 500×500×1 was performed to calculate TMR and polarization. The Perdew-Burke-Ernzerhof generalized gradient approximation (PBE-GGA) is selected to describe the exchange-correlation potential, and the cutoff energy of the real space grids is fixed as 3000 eV. Declarations Acknowledgements This work was supported by the National Natural Science Foundation of China (92164206, and 52261145694) and Beijing Natural Science Foundation (L234081). We thank Y. Han (Center of Nanofabrication, Tsinghua University) for micro-nano fabrication technical assistance. Data availability The data that support the plots within this paper and the other findings of this study are available from the corresponding author upon reasonable request. References Tan C, Tang J, Gao X, Xue C, Peng H (2025) 2D bismuth oxyselenide semiconductor for future electronics. Nat Rev Electr Eng 2:494–513 Incorvia JAC, Xiao TP, Zogbi N, Naeemi A, Adelmann C, Catthoor F, Tahoori M, Casanova F, Becherer M, Prenat G, Couet S (2024) Spintronics for achieving system-level energy-efficient logic. Nat Rev Electr Eng 1:700–713 Chen M, Ma Y, Aslam N, Liu C, Chen Y, Luo L, Zhang X, Mai K, Xiao H, Zhu K, Alharbi O, Zheng D, Xu X, Liao H, Yang Y, Wang H, Zhou Z, Wang H, Tian B, Li J, He X, Chang K, Wan Y, Shamim A, Alshareef HN, Lanza M, Anthopoulos TD, Han Z, Xue F, Zhang X (2025) Ultrawide-bandwidth boron nitride photonic memristors. Nat Nanotechnol 1–8 Guo Z, Yin J, Bai Y, Zhu D, Shi K, Wang G, Cao K, Zhao W (2021) Spintronics for energy-efficient computing: an overview and outlook. Proc. IEEE 109, 1398–1417 Khwa W-S, Wen T-H, Hsu H-H, Huang W-H, Chang Y-C, Chiu T-C, Ke Z-E, Chin Y-H, Wen H-J, Hsu W-T, Lo C-C, Liu R-S, Hsieh C-C, Tang K-T, Ho M-S, Lele AS, Teng S-H, Chou C-C, Chih Y-D, Chang T-YJ, Chang (2025) M.-F. A mixed-precision memristor and SRAM compute-in-memory AI processor. Nature 639:617–623 Huang Y, Cao K, Zhang K, Wang J, Shi K, Hao Z, Cai W, Du A, Yin J, Yang Q, Li J, Gao J, Zhao C, Zhao W (2023) Implementation of 16 Boolean logic operations based on one basic cell of spin-transfer-torque magnetic random access memory. Sci China Inf Sci 66:1–8 van Pham T, Groen I, Manipatruni S, Choi WY, Nikonov DE, Sagasta E, Lin C-C, Gosavi TA, Marty A, Hueso LE, Young IA, Casanova F (2020) Spin–orbit magnetic state readout in scaled ferromagnetic/heavy metal nanostructures. Nat Electron 3:309–315 Manipatruni S, Nikonov DE, Lin C-C, Gosavi TA, Liu H, Prasad B, Huang Y-L, Bonturim E, Ramesh R, Young IA (2019) Scalable energy-efficient magnetoelectric spin-orbit logic. Nature 565:35–42 Vaz DC, Lin C-C, Plombon JJ, Choi WY, Groen I, Arango IC, Chuvilin A, Hueso LE, Nikonov DE, Li H, Debashis P, Clendenning SB, Gosavi TA, Huang Y-L, Prasad B, Ramesh R, Vecchiola A, Bibes M, Bouzehouane K, Fusil S, Garcia V, Young IA, Casanova F (2024) Voltage-based magnetization switching and reading in magnetoelectric spin-orbit nanodevices. Nat Commun 15:1902 Lin C-C, Gosavi T, Nikonov D, Huang Y-L, Prasad B, Choi W, Pham VT, Groen I, Chen J-Y, Liu MDCH, Oguz K, Walker ES, Plombon J, Buford B, Naylor CH, Wang J-P, Casanova F, Ramesh R & I. A. Young. Experimental demonstration of integrated magneto-electric and spin-orbit building blocks implementing energy-efficient logic. In 2019 IEEE Int Electron Devices Meeting (IEDM), 37.3.1–37.3.4 (IEEE 2019) Debashis P, Plombon JJ, Lin C-C, Liao Y-C, Li H, Nikonov DE, Adams D, Rogan C, Dc M, Radosavljevic M, Clendenning SB, Young IA Low-voltage and high-speed switching of a magnetoelectric element for energy efficient compute. In 2022 IEEE Int Electron Devices Meeting (IEDM), 36.4.1–36.4.4 (IEEE 2022) Luo B, Will-Cole AR, Dong C, He Y, Liu X, Lin H, Huang R, Shi X, McConney M, Page M, Sanghadasa M, Ramesh R (2024) Sun, N. X. Magnetoelectric microelectromechanical and nanoelectromechanical systems for the IoT. Nat Rev Electr Eng 1:317–334 Vaz DC, Lin C-C, Plombon J, Choi WY, Groen I, Arango I, Pham VT, Nikonov DE, Li H, Debashis P, Clendenning SB, Gosavi TA, Garcia V, Fusil S, Bibes M, Huang Y-L, Prasad B, Ramesh R, Casanova F & I. A. Young. Functional demonstration of a fully integrated magnetoelectric spin-orbit device. In 2021 IEEE Int Electron Devices Meeting (IEDM), 32.4.1–32.4.4 (IEEE 2021) Grezes C, Kandazoglou A, Cosset-Cheneau M, Arche LMV, Noël P, Sgarro P, Auffret S, Garello K, Bibes M, Vila L, Attané J-P (2023) Non-volatile electric control of spin-orbit torques in an oxide two-dimensional electron gas. Nat Commun 14:2590 Ding J, Liu C, Kalappattil V, Zhang Y, Mosendz O, Erugu U, Yu R, Tian J, DeMann A, Field SB, Yang X, Ding H, Tang J, Terris B, Chen FERTA, H., Wu M (2021) Switching of a magnet by spin-orbit torque from a topological dirac semimetal. Adv Mater 33:2005909 Fan Y, Upadhyaya P, Kou X, Lang M, Takei S, Wang Z, Tang J, He L, Chang L-T, Montazeri M, Yu G, Jiang W, Nie T, Schwartz RN, Tserkovnyak Y, Wang KL (2014) Magnetization switching through giant spin-orbit torque in a magnetically doped topological insulator heterostructure. Nat Mater 13:699–704 Mellnik AR, Lee JS, Richardella A, Grab JL, Mintun PJ, Fischer MH, Vaezi A, Manchon A, Kim E-A, Samarth N, Ralph DC (2014) Spin-transfer torque generated by a topological insulator. Nature 511:449–451 Wu H, Chen A, Zhang P, He H, Nance J, Guo C, Sasaki J, Shirokura T, Hai PN, Fang B, Razavi SA, Wong K, Wen Y, Ma Y, Yu G, Carman GP, Han X, Zhang X, Wang KL (2021) Magnetic memory driven by topological insulators. Nat Commun 12:6251 Vicente-Arche LM, Bréhin J, Varotto S, Cosset-Cheneau M, Mallik S, Salazar R, Noël P, Vaz DC, Trier F, Bhattacharya S, Sander A, Le Fèvre P, Bertran F, Saiz G, Ménard G, Bergeal N, Barthélémy A, Li H, Lin C-C, Nikonov DE, Young IA, Rault JE, Vila L, Attané J-P, Bibes M (2021) Spin-charge interconversion in KTaO 3 2D electron gases. Adv Mater 33:2102102 Choi WY, Arango IC, van Pham T, Vaz DC, Yang H, Groen I, Lin C-C, Kabir ES, Oguz K, Debashis P, Plombon JJ, Li H, Nikonov DE, Chuvilin A, Hueso LE, Young IA, Casanova F (2022) All-electrical spin-to-charge conversion in sputtered Bi x Se 1-x . Nano Lett 22:7992–7999 Gallego F, Trier F, Mallik S, Bréhin J, Varotto S, Vicente-Arche M, Gosavy L, Lin T, Coudevylle C-C, Iglesias J-R, Casanova L, Young F, Vila I, Attané L, J.-P., Bibes M (2024) All-electrical detection of the spin-charge conversion in nanodevices based on SrTiO 3 2-D electron gases. Adv Func Mater 34:2307474 Parkin SSP, Kaiser C, Panchula A, Rice PM, Hughes B, Samant M, Yang S-H (2004) Giant tunnelling magnetoresistance at room temperature with MgO (100) tunnel barriers. Nat Mater 3:862–867 Yuasa S, Djayaprawira DD (2007) Giant tunnel magnetoresistance in magnetic tunnel junctions with a crystalline MgO(001) barrier. J Phys D: Appl Phys 40:R337–R354 Yuasa S, Nagahama T, Fukushima A, Suzuki Y, Ando K (2004) Giant room-temperature magnetoresistance in single-crystal Fe/MgO/Fe magnetic tunnel junctions. Nat Mater 3:868–871 Mathon J, Umerski A (2001) Theory of tunneling magnetoresistance of an epitaxial Fe/MgO/Fe(001) junction. Phys Rev B 63:220403 M.Julliere (1975) Tunneling between ferromagnetic films. Phys Lett A 54:225–226 Wang W, Liu E, Kodzuka M, Sukegawa H, Wojcik M, Jedryka E, Wu GH, Inomata K, Mitani S, Hono K (2010) Coherent tunneling and giant tunneling magnetoresistance in Co 2 FeAl/MgO/CoFe magnetic tunneling junctions. Phys Rev B 81 Scheike T, Wen Z, Sukegawa H, Mitani S (2023) 631% room temperature tunnel magnetoresistance with large oscillation effect in CoFe/MgO/CoFe(001) junctions. Appl Phys Lett 122 Strijkers GJ, Ji Y, Yang FY, Chien CL, Byers JM (2001) Andreev reflections at metal/superconductor point contacts: measurement and analysis. Phys Rev B 63 Kaiser C, Parkin SSP (2004) Spin polarization in ferromagnet/insulator/superconductor structures with the superconductor on top of the barrier. Appl Phys Lett 84:3582–3584 Paluskar PV, Attema JJ, de Wijs GA, Fiddy S, Snoeck E, Kohlhepp JT, Swagten HJM, de Groot RA, Koopmans B (2008) Spin tunneling in junctions with disordered ferromagnets. Phys Rev Lett 100:57205 Sinova J, Valenzuela SO, Wunderlich J, Back CH, Jungwirth T (2015) Spin Hall effects. Rev Mod Phys 87:1213–1260 Peng S, Zheng X, Li S, Lao B, Han Y, Liao Z, Zheng H, Yang Y, Yu T, Liu P, Sun Y, Chen X-Q, Peng S, Zhao W, Li R-W, Wang Z (2025) Unconventional scaling of the orbital Hall effect. Nat Mater 1–7 Zheng Z, Jia L, Zhang Z, Shen Q, Zhou G, Cui Z, Ren L, Chen Z, Jamaludin NF, Zhao T, Xiao R, Zhang Q, Du Y, Liu L, Gradečak S, Novoselov KS, Zhao W, Xu X, Zhang Y, Chen J (2025) All-electrical perpendicular switching of chiral antiferromagnetic order. Nat Mater 1–7 Takeuchi Y, Sato Y, Yamane Y, Yoon J-Y, Kanno Y, Uchimura T, de Zoysa KV, Han J, Kanai S, Ieda J, Ohno H, Fukami S (2025) Electrical coherent driving of chiral antiferromagnet. Science 389:830–834 Liu L, Chen C-T, Sun JZ (2014) Spin Hall effect tunnelling spectroscopy. Nat Phys 10:561–566 Groen I, van Pham T, Leo N, Marty A, Hueso LE, Casanova F (2021) Disentangling spin, anomalous, and planar Hall effects in ferromagnet–heavy-metal nanostructures. Phys Rev Appl 15:044010 Mizuno H, Isshiki H, Kondou K, Zhu Y, Otani Y (2021) Influence of planar Hall effect on the output signal in a T-shaped spin conversion device. Appl Phys Let 119:0061147 Pashenkin IY, Sapozhnikov MV, Gusev NS, Karashtin EA, Fraerman AA (2022) Extrinsic tunnel Hall effect in MgO-based tunnel junctions. Phys Rev B 106:L220408 Teresi S, Sebe N, Patterson J, Frottier T, Kandazoglou A, Noël P, Sgarro P, Térébénec D, Bernier N, Hippert F, Attané J-P, Vila L, Noé P, Cosset-Chéneau M (2023) Spin-orbit readout using thin films of topological insulator Sb 2 Te 3 deposited by industrial magnetron sputtering. Adv Func Mater 33:02303878 Lin Y-L, Wu Y-H, Lee S-K, Cheng C-W, Chen J-W, Chung C-H, Song M-Y, Bao X, Tseng Y-C (2024) Scalable spin-to-charge conversion device with nanopatterning. ACS Appl Electron Mater 6:6657–6664 Pai C-F, Liu L, Li Y, Tseng HW, Ralph DC, Buhrman RA (2012) Spin transfer torque devices utilizing the giant spin Hall effect of tungsten. Appl Phys Lett 101:122404 Sriram K, Mondal R, Pradhan J, Haldar A, Murapaka C (2023) Structural phase engineering of (α + β)-W for a large spin Hall angle and spin diffusion length. J Phys Chem C 127:22704–22712 Sun H-H, Zhang K-W, Hu L-H, Li C, Wang G-Y, Ma H-Y, Xu Z-A, Gao C-L, Guan D-D, Li Y-Y, Liu C, Qian D, Zhou Y, Fu L, Li S-C, Zhang F-C, Jia J-F (2016) Majorana zero mode detected with spin selective Andreev reflection in the vortex of a topological superconductor. Phys Rev Lett 116:257003 McKee SA (2004) Reflections on the memory wall. In 2004 Computer Frontiers Conference , p. 162ACM Press Hafner J (2008) Ab-initio simulations of materials using VASP: Density-functional theory and beyond. J Comput Chem 29:2044–2078 Taylor J, Guo H, Wang J (2001) Ab initio modeling of quantum transport properties of molecular electronic devices. Phys Rev B 63:245407 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryInformationV3.docx Supplementary materials Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7680431","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":521077357,"identity":"950afb74-eb1b-4f5f-a6c6-56201f749d7f","order_by":0,"name":"Weisheng Zhao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABD0lEQVRIiWNgGAWjYFACxgZmCIMHiCsOINhEajlDlBYGBoQWxjYitBgcb27+XFBxx66B/+zBDx/n3Unsbz/A+OBtG4O8OS4tZw62Sc848yy5geFcsuTMbc8SZ5xJYDac28ZguLMBuxazG4ltzLxth5MZGHsMpHm3HU5sOJDAJs3bxpBgcACHlvsPmz/z/gNqYeYx/s0753Di/PMP2H/j1XKDsUGat+GwHQMbjxmIkbjhRgIbMz4t9mcS26R5jh1OYOPhMbOcceyZ8cYbD5sl55yTMNyAQ4tk+/HHn3lqDtvz858xvvGh5o7svPPJBz+8KbORx2ULDCS2IdiMDUBCAr96kAMJqhgFo2AUjIKRCwA0a2HlLaXMqAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-8088-0404","institution":"Beihang University","correspondingAuthor":true,"prefix":"","firstName":"Weisheng","middleName":"","lastName":"Zhao","suffix":""},{"id":521077358,"identity":"f9aed573-8f76-4ede-81f9-a40cb980364b","order_by":1,"name":"Yan Huang","email":"","orcid":"","institution":"Fert Beijing Research Institute, Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Huang","suffix":""},{"id":521077359,"identity":"1e25f735-ad43-4ca7-9e8b-4d4068ec69b6","order_by":2,"name":"Kun Zhang","email":"","orcid":"","institution":"Fert Beijing Research Institute, Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Kun","middleName":"","lastName":"Zhang","suffix":""},{"id":521077360,"identity":"63634d02-bbc4-44aa-98ff-c88d4f4a5539","order_by":3,"name":"Guo Liu","email":"","orcid":"","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Guo","middleName":"","lastName":"Liu","suffix":""},{"id":521077361,"identity":"9469402c-402c-4abb-af5c-54593c445c67","order_by":4,"name":"Xiaobai Ning","email":"","orcid":"https://orcid.org/0009-0004-0206-0522","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Xiaobai","middleName":"","lastName":"Ning","suffix":""},{"id":521077362,"identity":"16a6983b-788b-4b6e-b67c-3805f7cf7667","order_by":5,"name":"Shiyang Lu","email":"","orcid":"","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Shiyang","middleName":"","lastName":"Lu","suffix":""},{"id":521077363,"identity":"aafa2c10-95bb-44ce-b551-0e1b057c3465","order_by":6,"name":"Wenlong Cai","email":"","orcid":"","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Wenlong","middleName":"","lastName":"Cai","suffix":""},{"id":521077364,"identity":"87520ca2-4331-4108-b02b-d0221ae78334","order_by":7,"name":"Yuxuan Yao","email":"","orcid":"https://orcid.org/0000-0003-0478-1601","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Yuxuan","middleName":"","lastName":"Yao","suffix":""},{"id":521077365,"identity":"7ad0947c-96a3-4cb6-b23b-612a98564d0c","order_by":8,"name":"Yu He","email":"","orcid":"","institution":"Fert Beijing Research Institute, Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"He","suffix":""},{"id":521077366,"identity":"a71d4060-e6fd-4cdf-90a2-f1c3bc97da84","order_by":9,"name":"Jinkai Wang","email":"","orcid":"","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Jinkai","middleName":"","lastName":"Wang","suffix":""},{"id":521077367,"identity":"5e0b407e-20d8-42a4-b550-c61fbdb8f886","order_by":10,"name":"Qing Yang","email":"","orcid":"","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Qing","middleName":"","lastName":"Yang","suffix":""},{"id":521077368,"identity":"e9cfaafe-ef76-4dcc-a86a-b68127766761","order_by":11,"name":"Bo Li","email":"","orcid":"","institution":"Fert Beijing Research Institute, Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Bo","middleName":"","lastName":"Li","suffix":""},{"id":521077369,"identity":"6a3cee98-cbf2-4482-802b-18443a14d56d","order_by":12,"name":"Haozhe Yang","email":"","orcid":"","institution":"Fert Beijing Institute, School of Integrated Circuit Science and Engineering, Beihang University, Beijing, 100191, China","correspondingAuthor":false,"prefix":"","firstName":"Haozhe","middleName":"","lastName":"Yang","suffix":""},{"id":521077370,"identity":"a02ea301-5c51-4758-a556-0fdb4d8a3b79","order_by":13,"name":"Kewen Shi","email":"","orcid":"","institution":"School of Integrated Circuit Science and Engineering, Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Kewen","middleName":"","lastName":"Shi","suffix":""},{"id":521077371,"identity":"92e9e3c8-a7c4-43a5-9aff-98ae6fe0e65d","order_by":14,"name":"Kaihua Cao","email":"","orcid":"","institution":"Fert Beijing Institute, BDBC, and School of Electronic and Information Engineering, Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Kaihua","middleName":"","lastName":"Cao","suffix":""},{"id":521077372,"identity":"95e288e0-6831-4004-be56-07759f8d2036","order_by":15,"name":"Chao Zhao","email":"","orcid":"","institution":"Institute of Microelectronics, Chinese Academy of Sciences, Beijing, 100029, China","correspondingAuthor":false,"prefix":"","firstName":"Chao","middleName":"","lastName":"Zhao","suffix":""},{"id":521077373,"identity":"b5da1a7f-c22d-4ba1-b0f5-81fd1deeceec","order_by":16,"name":"Yue Zhang","email":"","orcid":"https://orcid.org/0000-0001-6893-7199","institution":"Beihang University","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-09-22 15:38:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7680431/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7680431/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94581412,"identity":"6f8f5c05-988e-4ade-b624-c176b8477a00","added_by":"auto","created_at":"2025-10-28 18:12:31","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2412879,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConcept of MTJ-enhanced MESO logic\u003c/strong\u003e. a) Structure schematic of MTJ-enhanced MESO. The device formed with a magnetoelectric (ME) layer, two ferromagnetic (FM) layers, a spin injection (SI) layer, a spin-orbit-coupling (SOC) layer, a tunnel barrier and metal connect (MC). b) Principles of tunneling current modulation by MTJ resistance. The tunneling current shrinks when MTJ changes form P state to AP state. c) Principles of spin filtering effect, where ∆\u003csub\u003e1\u003c/sub\u003e band electron with nearly 100% spin polarization is allowed to be injected. d) Data-exchange architecture of “MTJ+MESO” schemes.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/0f015fc878988489b07c73a1.png"},{"id":94579319,"identity":"509a61ac-5312-4639-9435-0d2408f8012e","added_by":"auto","created_at":"2025-10-28 18:11:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1605674,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStructure optimization and basic device properties\u003c/strong\u003e.\u003cstrong\u003e \u003c/strong\u003ea) Schematic of multilayer film stacks. All numbers are in nm with varied MgO thickness t. b) RA and extracted film TMR ratio as a function of MgO thickness t. c) In-plane hysteresis loop of stacks in (a) with 1.1 nm MgO. Inset: minor loop of free layer behavior. d) Schematic of MTJ-enhanced reading module and device measurement set-up, where TMR and SCC signal aresimultaneously obtained. e) Top view (upper panel) of a 60 nm×130 nm junction and side view (lower panel). The well-crystallized MgO can be easily recognized. f) Junction resistance under 1 μA supply current and transverse voltage under 300 mV supply voltage versus x axis magnetic field along.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/61b2be4762b6a399d7b36a0a.png"},{"id":94580351,"identity":"f34a969e-4219-416b-bb20-742619ebfe4e","added_by":"auto","created_at":"2025-10-28 18:11:57","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":360082,"visible":true,"origin":"","legend":"\u003cp\u003eOutput signal characterization under current source. a) SCC signal ∂V\u003csub\u003e24\u003c/sub\u003e/∂I\u003csub\u003e01\u003c/sub\u003e and its reciprocal CSC signal ∂V\u003csub\u003e01\u003c/sub\u003e/∂I\u003csub\u003e24\u003c/sub\u003e as a function of external field with 5 μA supply current. b) Normalized transverse resistance versus external field under different supply current. c) 2∆RSCC as a function of supply current. d) MTJ resistance and transverse output resistance at different state as a function of supply current.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/f94077bc06fcddb5e016e495.png"},{"id":94580305,"identity":"7927c58c-c278-4010-9b5a-d1086ef555f7","added_by":"auto","created_at":"2025-10-28 18:11:56","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":569974,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOutput signals under different sources.\u003c/strong\u003e a) Output voltage comparison between voltage supply and current supply. Inset: Transverse voltage under 300 mV v.s. that under 40 μA. b) Resistance outputs at different supply voltages. c) Benchmark of output signals of MESO-like device for different materials and configurations until now.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/73833850e57ca47ffa902000.png"},{"id":94581230,"identity":"c475db75-4622-401d-a9d9-94869b797865","added_by":"auto","created_at":"2025-10-28 18:12:23","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1309589,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eControl samples and mechanism analysis\u003c/strong\u003e. a) Output resistance signal comparisons between W/MTJ device and W/CoFeB-bilayer device using microdot device. b) Numerical simulations for the bilayer and multilayer structure using physical parameters in experiments. c) and d) Layer structures of W/CoFe/W and W/CoFe/MgO/CoFe/W for ab-initio calculations. e) and f) Majority-to-majority and minority-to-minority transport for W/CoFe/W. g) and h) Majority-to-majority and minority-to-minority transport for W/CoFe/MgO/CoFe/W.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/1c647fe27c2fba0893afcda9.png"},{"id":94595226,"identity":"466d29a2-0c8b-4c35-b8ed-52072e54ff13","added_by":"auto","created_at":"2025-10-28 18:32:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5405474,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/e193f33d-bf56-4df3-8967-5be4b9440b13.pdf"},{"id":94580547,"identity":"4f2a1b9d-073b-4399-ad5f-30982472826a","added_by":"auto","created_at":"2025-10-28 18:12:03","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":767161,"visible":true,"origin":"","legend":"Supplementary materials","description":"","filename":"SupplementaryInformationV3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7680431/v1/2f48156f88a7a19155c8a107.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Giant Spin-Orbit Magnetic State Readout Enhanced by a Magnetic Tunnel Junction","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eWith technology nodes shrinking, conventional complementary metal-oxide-semiconductor (CMOS) technology faces a growing hurdle in scaling down due to leakage current in the OFF state, imposing a limit on the overall system energy efficiency\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. As a solution, emerging nonvolatile memories, featuring normally-OFF/instantly-ON computation, can overcome the static power brought by leakage current\u003csup\u003e\u003cspan additionalcitationids=\"CR4 CR5 CR6\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. However, because these technologies are not supported to generate an effective driving force, they are hard to drive the subsequent data writing and construct cascaded logic circuits\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eMagnetoelectric spin-orbit (MESO) logic, utilizing voltage-controlled magnetoelectric (ME) effect for data writing and spin-orbit coupling (SOC) for magnetic state reading, can be directly cascaded and is regarded as a compelling alternative to CMOS toward 100-mV logic \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. In a MESO logic, a single magnetic layer is used to store bit data, as well as couple writing block (ME block) and reading block (SO block)\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. To date, the quasi-static ferroelectric switching of ME writing module has been achieved down to about 150 mV using ultrathin La-doped BiFeO\u003csub\u003e3\u003c/sub\u003e, with a pathway to get down to 100 mV\u003csup\u003e8,11,12\u003c/sup\u003e. However, there is a notable challenge that the spin-to-charge conversion (SCC) voltage of readout module is far less than 100 mV, which greatly restricts the cascading function of MESO logic. Typically, this signal measures in the tens or hundreds of milliohms, which results in an output voltage of less than 20 \u0026micro;V\u003csup\u003e7,13\u003c/sup\u003e. The stark discrepancy between the voltages required for writing and those available for reading presents a wide gap to the practical implementation of MESO logics. Although it is predicted using emerging quantum materials such as topological insulators (TIs) and two-dimensional electron gases (2DEGs) could enlarge the SCC signal due to their high SCC efficiency\u003csup\u003e\u003cspan additionalcitationids=\"CR15 CR16 CR17 CR18\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e, experimental results at room temperature (R.T.), to date, have not met the anticipated expectations\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. In addition, the processing compatibility for massive production of these quantum materials has not been totally addressed. Therefore, more brand-new enhancement mechanism of output signal of MESO, different from utilizing high SOC materials, is highly desired to promote the MESO technology towards practical applications.\u003c/p\u003e\u003cp\u003eHere, we propose for the first time the integration of magnetic tunnel junction (MTJ) and MESO by sharing one ferromagnetic (FM) layer as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea. On the one hand, the injection current modulated by MTJ can amplify the output voltage. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb, the injection tunneling current experiences a sudden increase upon the transition of MTJ from antiparallel (AP) to parallel (P) state under a constant voltage supply V\u003csub\u003esupply\u003c/sub\u003e, \u003cem\u003ei.e.\u003c/em\u003e, from I\u003csub\u003eAP\u003c/sub\u003e to I\u003csub\u003eP\u003c/sub\u003e. This current alteration can induce an extra voltage signal superposed to the SCC output signal, and amplify the output signal of MESO. On the other hand, the high spin polarization (SP) rate effectuated by the spin filtering mechanism inherent in the single-crystal CoFeB/MgO/CoFeB sandwiched-structure can improve the SCC output voltage. As shown in \u003cb\u003eFig.\u0026nbsp;1c\u003c/b\u003e, Δ\u003csub\u003e1\u003c/sub\u003e band electrons in sandwiched CoFeB/MgO/CoFeB tunnel easily when the magnetization of two CoFeB layers are parallel alignment, and hard when they are antiparallel, while Δ\u003csub\u003e2\u003c/sub\u003e and Δ\u003csub\u003e5\u003c/sub\u003e band electrons count little\u003csup\u003e\u003cspan additionalcitationids=\"CR23 CR24\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Consequently, tunneling current across an MTJ could be highly polarized to significantly enhance SP rate. Our experimental findings at R.T. reveal a giant output signal, which is significantly enhanced to an impressive 1.5 mV, about two orders of magnitude greater than that reported in previous literatures\u003csup\u003e\u003cspan additionalcitationids=\"CR8 CR9\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Furthermore, the integrated MTJ can function as a cache memory to provide infinite bandwidth for MESO logic as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed, addressing the data exchange issue of computing unit and external memory from the perspective of architecture level, a problem that has been overlooked in previous researches. The proposed \u0026ldquo;MESO\u0026thinsp;+\u0026thinsp;MTJ\u0026rdquo; scheme can enhance the output signal and optimize the design architecture, offering a promising direction for future development of MESO technology.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"2. Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Structural, magnetic and electrical properties\u003c/h2\u003e\u003cp\u003eAt the outset of our experiments, we examined the impact of the MgO layer thickness on the tunneling magnetoresistance (TMR) ratio, which reveals film quality and SP rate. The functional stacks depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, consisting of W (3.5)/CoFeB (1.9)/MgO (t)/CoFeB (1.9)/CoFe (0.5)/Ru (0.8)/CoFe (2)/IrMn (7.5)/Ru (5), are deposited on bottom electrode Ta (25)/CuN (20) (numbers in parentheses denotes film thickness in nm) with varying MgO thickness \u003cem\u003et\u003c/em\u003e. After post annealing at 400 ℃, we determined resistance-area product (RA) and TMR by using current in-plane tunneling (CIPT) measurement at R.T.. The results are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb. The RA value increases with increasing MgO thickness, while the TMR ratio exhibits an initial increase followed by a decrease, reaching its maximum when the MgO thickness is 1.1 nm. This optimal TMR value suggests a higher quality of the film and a greater degree of SP\u003csup\u003e26\u003c/sup\u003e. Note that, TMR is expected to appear oscillations as \u003cb\u003et\u003c/b\u003e keeps increasing due to coherent tunneling effect\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eWe then selected a 1.1 nm MgO layer for subsequent experiments, and deposited MTJ stacks directly on thermally oxidized silicon substrate without Ta (25)/CuN (20). Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec illustrates the hysteresis loop of full functional stacks deposited on thermally-oxidized silicon substrate using vibrating sample magnetometer. By applying an in-plane magnetic field along x direction, distinct magnetic behaviors were observed. The minor loop inset in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec indicates that the coercive field of the free layer is approximately 10 Oe.\u003c/p\u003e\u003cp\u003eAfter a series of electron beam and ultraviolet lithography, as well as etching steps, we have successfully fabricated a junction device featuring multiple electrodes. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed illustrates the schematic of proposed MTJ-enhanced MESO readout block. The device features with functional layers of two FM, a SOC, and a tunneling barrier layers. A cross-shape bottom electrode is used to generate transverse output voltage which can be measured between lead 2 and 3, with power supply across lead 0 and 1\u003csup\u003e7\u003c/sup\u003e. Generally, this reading process can be divided into two distinct phases: the spin polarization phase and the SCC phase. During the spin polarization phase, a constant supply source, denoted as V\u003csub\u003esupply\u003c/sub\u003e or I\u003csub\u003esupply\u003c/sub\u003e, is applied perpendicularly to the device stacks and injecting a spin polarized current\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. The injection efficiency is characterized by SP rate. Typically, SP rates are approximately 0.37 for Ni, 0.45 for Co, and 0.43 for Fe, with NiFe and CoFeB exhibiting higher values of 0.48 and 0.53, respectively\u003csup\u003e\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. By applying an external magnetic field and sweeping it along x axis, vertical voltage V\u003csub\u003e1\u003c/sub\u003e (V\u003csub\u003eMTJ\u003c/sub\u003e) and transverse voltage V\u003csub\u003e2\u003c/sub\u003e (V\u003csub\u003exy\u003c/sub\u003e) are both supposed to appear abrupt changes when free layer magnetization changes, indicating TMR and SCC output signals.\u003c/p\u003e\u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee, the upper panel shows the top view a junction with an area about 60 nm \u0026times; 130 nm, while the lower panel gives the side view of junction in which the MgO barrier can be recognized clearly, indicating good crystallization. TMR and SCC signals are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ef. The low resistance state, corresponding to the alignment of the two magnetic layers at P state, is measured to be around R\u003csub\u003eMTJ,P\u003c/sub\u003e = 4.6 kΩ. Meanwhile, the high resistance state exhibits a resistance of around R\u003csub\u003eMTJ,AP\u003c/sub\u003e = 8 kΩ, corresponding to AP state. Consequently, TMR ratio is calculated to be 74% according to the equation of TMR = (R\u003csub\u003eMTJ,AP\u003c/sub\u003e - R\u003csub\u003eMTJ,P\u003c/sub\u003e)/R\u003csub\u003eMTJ,P\u003c/sub\u003e. The TMR loss compared with CIPT results can be attributed to damage during nanofabrication. Notably, there are some anomalous points when the external field is around \u0026plusmn;\u0026thinsp;20 Oe, which may be caused by irregular shapes and probable defects and pinnings at the edge of the junction. Meanwhile, we measure transverse voltage V\u003csub\u003exy\u003c/sub\u003e versus H\u003csub\u003ex\u003c/sub\u003e (V\u003csub\u003exy\u003c/sub\u003e-H\u003csub\u003ex\u003c/sub\u003e loop) under a constant voltage supply of 300 mV. As expected, there is abrupt changes in transverse signal V\u003csub\u003exy\u003c/sub\u003e, precisely at the points where free layer magnetization reverses. The output voltage ∆V\u003csub\u003exy\u003c/sub\u003e is defined as the difference between two magnetic states. By utilizing MTJ modulation and high SP rate, the observed output voltage ∆V\u003csub\u003exy\u003c/sub\u003e at 300 mV supply voltage is approximately 1.5 mV. This result is at least two orders larger than those in previous reports (less than 15 \u0026micro;V in ref \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e and 1.5 \u0026micro;V in ref \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 SCC signal under current source\u003c/h2\u003e\u003cp\u003eWe performed output signal under a constant current source to precisely evaluate the SCC signal without resistance modulation. The SCC result is intrinsically guaranteed by the anomalous spin-polarized velocity arising from a momentum-space Berry phase of Bloch electrons\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. Reciprocally, charge-to-spin conversion (CSC), the inverse process of the SCC, can also be assessed using the same device. The charge flowing along W channel would cause spin accumulation and diffusion at and cross interfaces, known as direct spin Hall effect (SHE)\u003csup\u003e\u003cspan additionalcitationids=\"CR33 CR34\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e, which can be detected by magnetic electrodes\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. Through applying current across lead 2 and 4, and measuring voltage between lead 0 and 1, SHE can be evaluated as CSC signal.\u003c/p\u003e\u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea, we present the SCC curve \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\partial\\:{V}_{24}/\\partial\\:{I}_{01}\\)\u003c/span\u003e\u003c/span\u003e and its inverse, CSC curve \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\partial\\:{V}_{01}/\\partial\\:{I}_{24}\\)\u003c/span\u003e\u003c/span\u003e at 5 \u0026micro;A supply current. The CSC curve appears with opposite polarity to SCC curve, consistent with previous reports\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. This indicates a giant SCC signal \u003cem\u003e2∆R\u003c/em\u003e\u003csub\u003e\u003cem\u003eSCC\u003c/em\u003e\u003c/sub\u003e around 14.5 Ω, which is nearly 50 times larger than previous reports (0.3 Ω)\u003csup\u003e7\u003c/sup\u003e. However, the CSC signal is about 11 Ω, slightly smaller than SCC signal. We attribute this difference to the spin diffusion and relaxation in z-direction, compared with spin drift in SCC situation\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. The measurement set-up in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed also brings various Hall signals mixing in the transverse voltage V\u003csub\u003exy\u003c/sub\u003e, such as conventional direct Hall effect, planar Hall effect and anomalous Hall effect\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. Details are discussed in \u003cb\u003eSupplementary Materials S1\u003c/b\u003e. These Hall signals count less than 0.1 Ω in our nanodot structure, and affect little for the evaluation of final output signal.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe spin injection phase and the SCC phase are both theoretically bias dependent. At first, the spin injection facilitated by an MTJ could be modulated by its bias-dependent SP\u003csup\u003e24\u003c/sup\u003e. Moreover, considering the ISHE mechanism in SCC phase, the transverse voltage loop of V\u003csub\u003e2\u003c/sub\u003e shall invert polarity when the supply current direction is reversed, that is, V\u003csub\u003e2\u003c/sub\u003e (-I\u003csub\u003esupply\u003c/sub\u003e, H\u003csub\u003ex\u003c/sub\u003e) = -V\u003csub\u003e2\u003c/sub\u003e (I\u003csub\u003esupply\u003c/sub\u003e, H\u003csub\u003ex\u003c/sub\u003e), resulting no polarity change in resistance versus field curve, that is, R\u003csub\u003exy\u003c/sub\u003e (-I\u003csub\u003esupply\u003c/sub\u003e, H\u003csub\u003ex\u003c/sub\u003e)\u0026thinsp;=\u0026thinsp;R\u003csub\u003exy\u003c/sub\u003e (I\u003csub\u003esupply\u003c/sub\u003e, H\u003csub\u003ex\u003c/sub\u003e)\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. As a result, if one obtains R\u003csub\u003exy\u003c/sub\u003e (-I\u003csub\u003esupply\u003c/sub\u003e, H\u003csub\u003ex\u003c/sub\u003e) = -R\u003csub\u003exy\u003c/sub\u003e (I\u003csub\u003esupply\u003c/sub\u003e, H\u003csub\u003ex\u003c/sub\u003e), this result may be dominated by a current-symmetric term rather than SCC. The second is normally neglected by most researches, but it plays a significant role in SCC signal assessment.\u003c/p\u003e\u003cp\u003eTherefore, we measured different SCC signals by applying currents of varying amplitude and direction, to investigate the bias impact. The SCC signals at \u0026plusmn;\u0026thinsp;30 \u0026micro;A appear with the same polarity as expected, but are approximately 2.5 Ω lower than that at 1 \u0026micro;A, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb. Moreover, the signal at +\u0026thinsp;30 \u0026micro;A seems larger than signal at -30 \u0026micro;A. A detailed examination reveals a significant degradation in the SCC signal as supply current is increased, as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec. Interestingly, similar phenomena have been previously reported in W/MgO/CoFeB tri-layer spin-tunneling structure\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. The maximium SCC signal of \u003cem\u003e2∆R\u003c/em\u003e\u003csub\u003e\u003cem\u003eSCC\u003c/em\u003e\u003c/sub\u003e is around 14.5 Ω at 5 \u0026micro;A supply. Obviously, when the current amplitudes are equal, the signal obtained under negative bias is slightly smaller than that under positive bias. These observations may imply a bias-dependent SP rate.\u003c/p\u003e\u003cp\u003eConsequently, we explored the bias current dependence of the junction resistance R\u003csub\u003eMTJ\u003c/sub\u003e, and transverse resistance R\u003csub\u003exy\u003c/sub\u003e. We use R\u003csub\u003eMTJ,P\u003c/sub\u003e (R\u003csub\u003eMTJ,AP\u003c/sub\u003e) and R\u003csub\u003exy,P\u003c/sub\u003e (R\u003csub\u003exy,AP\u003c/sub\u003e) to represent the junction and transverse resistances when MTJ is at P (AP) state, respectively. The findings are summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed, where it is noted that while R\u003csub\u003eMTJ,AP\u003c/sub\u003e exhibits a pronounced variation with bias current, R\u003csub\u003eMTJ,P\u003c/sub\u003e remains relatively unchanged. For the transverse resistance R\u003csub\u003exy\u003c/sub\u003e, both R\u003csub\u003exy,P\u003c/sub\u003e and R\u003csub\u003exy,AP\u003c/sub\u003e exhibit monotonically decreasing trend with the increasing supply current, suggesting an additional mechanism between the SCC and TMR signals. Generally, the SCC signal in our MTJ-enhanced readout module contain two components of information, \u003cem\u003ei.e.\u003c/em\u003e energetic electrons contribution from reference layer and equilibrium electrons mostly from free layer\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. The energetic electrons are more energy-sensitive and thus attribute to the bias-dependent signal. These analyses collectively suggest that the utilization of highly spin-polarized electrons tunneling from CoFeB/MgO/CoFeB structures holds greater promise than those from a single FM layer in the development of spin-injection-based devices.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 MTJ-enhanced output signals under voltage source\u003c/h2\u003e\u003cp\u003eThen, we compare the output voltage ∆V\u003csub\u003exy\u003c/sub\u003e under current source and voltage source. When the applied current source is substituted with a constant voltage V\u003csub\u003esupply\u003c/sub\u003e, the injection current would experience a sudden decrease upon the transition of MTJ from P to AP state.This alteration consequently affects the output voltage ∆V\u003csub\u003exy\u003c/sub\u003e. As illustrated in the inset of Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea, the maximum observed output voltage ∆V\u003csub\u003exy\u003c/sub\u003e is approximately 1.5 mV corresponding to a 300 mV supply voltage, while the largest ∆V\u003csub\u003exy\u003c/sub\u003e for current source is about 0.4 mV for 40 \u0026micro;A supply. For a supply voltage of 300 mV, the tunneling current under the P state measures approximately 65 \u0026micro;A. Given that ∆V\u003csub\u003exy\u003c/sub\u003e is proportional to the amplitude of the injection tunneling current, under a supply current of 65 \u0026micro;A, ∆V\u003csub\u003exy\u003c/sub\u003e is estimated to be around 0.65 mV, also significantly lower than 1.5 mV. These findings effectively validate the regulation impact of TMR effect when operates under a voltage source.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eA comprehensive list of output voltages for various applied currents and voltages is provided in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea. It is observed that larger applied source generates larger output voltage, and the voltage source can yield larger output voltage than current source when the current flow are comparable. Moreover, we could calculate the transverse resistance difference under voltage source by the formula, ∆R\u003csub\u003exy\u003c/sub\u003e = R\u003csub\u003exy,P\u003c/sub\u003e - R\u003csub\u003exy,AP\u003c/sub\u003e = V\u003csub\u003exy,P\u003c/sub\u003e/I\u003csub\u003eP\u003c/sub\u003e - V\u003csub\u003exy,AP\u003c/sub\u003e/I\u003csub\u003eAP\u003c/sub\u003e. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb, the relationship between ∆R\u003csub\u003exy\u003c/sub\u003e and voltage supply exhibits no regularity, quite different from that depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, meaning an extra term brought by TMR signal, in agreement with our analysis.\u003c/p\u003e\u003cp\u003eTo clarify the regulation impact of MTJ, we calculate the relationship between ∆V\u003csub\u003exy\u003c/sub\u003e and MTJ parameters under a voltage source as follows,\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:\\varDelta\\:{\\text{V}}_{\\text{x}\\text{y}}={\\text{V}}_{\\text{x}\\text{y},\\text{P}}-{\\text{V}}_{\\text{x}\\text{y},\\text{A}\\text{P}}=[\\left(1+\\text{T}\\text{M}\\text{R}\\right){\\text{R}}_{\\text{x}\\text{y},\\text{P}}-{\\text{V}}_{\\text{x}\\text{y},\\text{A}\\text{P}}]\\frac{{\\text{V}}_{\\text{s}\\text{u}\\text{p}\\text{p}\\text{l}\\text{y}}}{{\\text{R}}_{\\text{M}\\text{T}\\text{J},\\text{A}\\text{P}}}\\:$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe detailed formular derivation process can be found in \u003cb\u003eSupplementary Material S2\u003c/b\u003e. For an MTJ, the parameter I\u003csub\u003eAP\u003c/sub\u003e ​ is directly determined by the thickness of the MgO barrier. Meanwhile, the TMR ratio critically depends on the overall quality of the MTJ structure. Therefore, to achieve a substantial increase in ∆V\u003csub\u003exy\u003c/sub\u003e under a constant voltage source, it is essential to optimize two key factors: 1) the MgO barrier must be sufficiently thin to enhance tunneling efficiency, and reduce R\u003csub\u003eMTJ,AP\u003c/sub\u003e; 2) the crystalline quality of the MTJ must be carefully optimized to ensure large TMR ratio.\u003c/p\u003e\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec lists some impressive progress in MESO-like readout modules reported so far, containing different SOC materials (heavy metals, TIs and 2DEGs)\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Some of them provide output resistance signal ∆R\u003csub\u003exy\u003c/sub\u003e with a given current I\u003csub\u003esupply\u003c/sub\u003e, but no output voltages ∆V\u003csub\u003exy\u003c/sub\u003e. As ∆V\u003csub\u003exy\u003c/sub\u003e is more critical and essential for practical application, We derive their voltage signals as ∆V\u003csub\u003exy\u003c/sub\u003e = I\u003csub\u003esupply\u003c/sub\u003e\u0026times;∆R\u003csub\u003exy\u003c/sub\u003e and make a comparison. Although TIs and 2DEGs may remain theoritical superiority thanks to topological surface states and Rashba surface, experimental outputs have no advantages compared with W or Pt based device. One possible reason is their topological state are not stable at R.T.. W-based device outputs 3-times larger than Pt-based device as expected, when they are patterned with same size\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Interestingly, Pt-based SO device has been reported the most, yet its results vary greatly. The most possible reason is Pt\u0026rsquo;s resistivity in reports varies a lot, from tens of \u0026micro;Ω\u0026middot;cm to hundreds of \u0026micro;Ω\u0026middot;cm\u003csup\u003e7\u003c/sup\u003e. Among all of these proposals, our work not only outperforms all others by a significant margin, but also enhances the output voltage from the microvolt level to the millivolt level, marking a new record.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Mechanism Explorations\u003c/h2\u003e\u003cp\u003eIn this part, we explored possible mechanisms and factors to induce the giant readout signal. At first, we compared the output transverse signal in proposed MTJ-enhanced device and traditional FM/SOC-bilayer device. We designed some control experiments using microdot devices with the same channel width of 3 \u0026micro;m. The output signals at I\u003csub\u003esupply\u003c/sub\u003e = 10 \u0026micro;A of these two samples are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea. Sample 1 is a bilayer device composed of W (6)/CoFeB (5), yielding about 2.5 mΩ output signal. Sample 2 is a MTJ-enhanced device with the above-mentioned film stack structure, which presents approximately 1.2 Ω signal. This difference is quite reasonable according to the numerical model\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e: 1) the SP rates of sample 2 is about several times larger that of sample1; 2) Because there is a phase change when the thickness of W changes from 6 nm to 3.5 nm, the W resistivity in sample 1 is about 60 \u0026micro;Ω\u0026middot;cm and that in sample 2 is about 367 \u0026micro;Ω\u0026middot;cm; 3) The θ\u003csub\u003eSH\u003c/sub\u003e decreases for W from 0.26 at 3.5 nm to 0.04 at 6 nm\u003csup\u003e42,43\u003c/sup\u003e. The characterization of θ\u003csub\u003eSH\u003c/sub\u003e of W are presented in \u003cb\u003eSupplementary Materials S3\u003c/b\u003e. These factors give about hundreds times of difference between these two scenarios, as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb, quite matching the experiments. More details about the numerical model are presented in \u003cb\u003eSupplementary Materials S4.\u003c/b\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo further explain the point 1) about SP rates, we established W/CoFe/W and W/CoFe/MgO/CoFe/W atomic-level structure to perform calculation as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed. K-resolved transmission spectra in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ee and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef illustrate the majority-to-majority tunneling and minority-to-minority tunneling in W/CoFe/W, both generally spreading over the Brillouin zone. In contrast, majority-to-majority tunneling in W/CoFe/MgO/CoFe/W (P state) is concentrated within a square region around Γ (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eg), vanishing in the minority-to-minority spectrum (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eh). Given the definition of spin polarization rate \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{S}\\text{P}=({\\text{G}}_{\\uparrow\\:\\uparrow\\:}^{\\uparrow\\:}-{\\text{G}}_{\\uparrow\\:\\uparrow\\:}^{\\downarrow\\:})/({\\text{G}}_{\\uparrow\\:\\uparrow\\:}^{\\uparrow\\:}+{\\text{G}}_{\\uparrow\\:\\uparrow\\:}^{\\downarrow\\:})\\)\u003c/span\u003e\u003c/span\u003e, where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{G}_{\\uparrow\\:\\uparrow\\:}^{\\uparrow\\:}\\)\u003c/span\u003e\u003c/span\u003e is the majority-to-majority spin conductance and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{G}_{\\uparrow\\:\\uparrow\\:}^{\\downarrow\\:}\\)\u003c/span\u003e\u003c/span\u003e is the minority-to-minority spin conductance\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e, the calculated value gives about 92% SP rate for W/CoFe/MgO/CoFe/W structure, about 4 times larger than that for W/CoFe/W. It is evident that several monolayer MgO acts as spin filter to enlarge the difference between majority and minority spin transmission.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Conclusion and Outlook","content":"\u003cp\u003eOur proposed \"MESO + MTJ\" scheme effectively overcomes key bottlenecks in MESO logic systems by simultaneously amplifying output signals and redefining data-exchange architectures. This dual advancement manifests in two pivotal aspects:\u003c/p\u003e\u003cp\u003eOn the one hand, the output voltage can be further improved to 100 mV to enable the direct cascading of multiple MESO devices. According to Eq.\u0026nbsp;\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, output voltage conversion efficiency ∆V\u003csub\u003exy\u003c/sub\u003e/V\u003csub\u003esupply\u003c/sub\u003e is highly related to MTJ properties. With higher TMR, and lower MTJ resistance at AP state, comes higher conversion efficiency. With proper stack design and interface optimization, TMR at R.T. of MgO-single-crystal MTJ can be increased high up to 631%\u003csup\u003e28\u003c/sup\u003e. Meanwhile, high TMR ratio means high spin polarization of injection current. By adapting strong SOC material and scaling down the device size, the SCC process can be further improved. These methods are promising to improve the output voltages to the order of 100 mV.\u003c/p\u003e\u003cp\u003eOn the other hand, the \"MESO + MTJ\" design addresses the critical issue of data exchange between computing unit and peripheral memory from system level. A one-bit full adder composed of \"MESO + MTJ\" units is schematically shown in \u003cb\u003eSupplementary Fig. S3\u003c/b\u003e of \u003cb\u003eSupplementary Materials\u003c/b\u003e. This scheme can operate as a logic unit with a intrinsically-integrated high-speed cache memory to resolve the bandwidth mismatch between logic unit and external memory\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. Moreover, this innovative design enables the extraction of any intermediate process results while simultaneously monitoring its resistance state during computation, which can improve the computation efficiency and relibility.\u003c/p\u003e\u003cp\u003eIn conclusion, our proposal offers a viable pathway toward advancing the real-world implementation of MESO logic technology.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cem\u003eFilm growth and characterization\u003c/em\u003e:\u003c/p\u003e\u003cp\u003eAll the films used in this work were deposited on 8-inch thermally oxidized silicon substrate, using ultrahigh vacuum \u003cem\u003edc/rf\u003c/em\u003e magnetron sputtering system. The substrate was first pre-cleaned for 60 s by Ar\u003csup\u003e+\u003c/sup\u003e ion beam. The CoFe, CoFeB, and IrMn denote Co\u003csub\u003e70\u003c/sub\u003eFe\u003csub\u003e30\u003c/sub\u003e, Co\u003csub\u003e60\u003c/sub\u003eFe\u003csub\u003e20\u003c/sub\u003eB\u003csub\u003e20\u003c/sub\u003e, and Ir\u003csub\u003e20\u003c/sub\u003eMn\u003csub\u003e80\u003c/sub\u003e alloy with nominal target compositions, respectively. After deposition, the stacks were in-situ annealed at 400 ℃ for 1 h under an external field of 1 T along x direction to define the reference layer direction and form highly oriented crystalline CoFeB/MgO/CoFeB MTJs. The static in-plane anisotropy magnetization behavior was examined using a vibrating sample magnetometer system.\u003c/p\u003e\u003cp\u003e\u003cem\u003eDevice fabrication\u003c/em\u003e:\u003c/p\u003e\u003cp\u003eFor a nanoscale MTJ-enhanced spin-orbit device, we first patterned 500 nm × 2.5 µm cross-shape channels using standard electron beam lithography (EBL) and Ar\u003csup\u003e+\u003c/sup\u003e ion beam etching (IBE). During milling, we monitored the secondary-ion mass spectra. After that, a second round of EBL and IBE procedures were used to define elliptical junctions. Thanks to high-resolution and high-aspect-ratio negative tone photoresist, we successfully formed a photoresist pillar with tens-nanometer length and hundreds-nanometer height. The junction etching was divided into two phases: first, we defined its appearance by small angle etching until the bottom 3.5-nm-thick W signal appears; second, large angle etching was utilized to meticulously clean the sidewall re-depositions across the barrier. Rather than removing the photoresist immediately, we covered the pillars with a 50-nm-thick silicon nitride (Si\u003csub\u003e3\u003c/sub\u003eN\u003csub\u003e4\u003c/sub\u003e) passivation layer, by using chemical vapor deposition, and then followed by lift-off process. Vias through bottom electrodes were formed by standard ultraviolet lithography and inductively coupled plasma etching. Ti (10nm)/ Au (100nm) electrodes are finally formed by E-beam evaporation (EBE).\u003c/p\u003e\u003cp\u003eFor the microscale SO device, \u003cem\u003ei.e.\u003c/em\u003e, CoFeB/W readout device and CoFeB/MgO/CoFeB/W MTJ-enhanced readout device, standard ultraviolet lithography and etching process were used to pattern the bottom electrode and the microdot pillar. Then, 50-nm-thick Si\u003csub\u003e3\u003c/sub\u003eN\u003csub\u003e4\u003c/sub\u003e passivation layer covered the pillars, and followed by lift-off process. Finally, Ti (10 nm)/Au (100 nm) electrodes are formed by EBE. For bilayer ST-FMR measurement, the device was formed by standard ultraviolet lithography and etching process, with Ti(10 nm)/ Au(100 nm) electrodes using EBE and lift-off process.\u003c/p\u003e\u003cp\u003e\u003cem\u003eDevice characterization\u003c/em\u003e:\u003c/p\u003e\u003cp\u003eElectronic transport measurements of MTJ-enhanced readout device were conducted on our self-developed probe station, using a Keithley 2182 nanovoltmeter and a Keysight B1500A semiconductor device parameter analyzer under varying in-plane magnetic field. The B1500A is used for direct resistance measurement while 2182 nanovoltmeter is more sensitive and used for transverse voltage measurement. ST-FMR measurement of the bilayer heterojunction was performed using a Keysight MXG N5183B analog signal generator and SR830 lock-in amplifier.\u003c/p\u003e\u003cp\u003e\u003cem\u003eab-initio simulation\u003c/em\u003e:\u003c/p\u003e\u003cp\u003eThe atomic structures are relaxed by Vienna ab-initio simulation package (VASP) so that the residual forces are minimized under 0.01 eV/Å\u003csup\u003e46\u003c/sup\u003e. Next, transport calculations are performed by NanoDCAL, which is based on density functional theory (DFT) and Keldysh non-equilibrium Greens function (NEGF)\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. After a self-consistent calculation of 20×20×1 k-point mesh, the spin-resolved conductance calculation of 500×500×1 was performed to calculate TMR and polarization. The Perdew-Burke-Ernzerhof generalized gradient approximation (PBE-GGA) is selected to describe the exchange-correlation potential, and the cutoff energy of the real space grids is fixed as 3000 eV.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgements\u003c/h2\u003e\u003cp\u003eThis work was supported by the National Natural Science Foundation of China (92164206, and 52261145694) and Beijing Natural Science Foundation (L234081). We thank Y. Han (Center of Nanofabrication, Tsinghua University) for micro-nano fabrication technical assistance.\u003c/p\u003e\u003ch2\u003eData availability\u003c/h2\u003e\u003cp\u003eThe data that support the plots within this paper and the other findings of this study are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eTan C, Tang J, Gao X, Xue C, Peng H (2025) 2D bismuth oxyselenide semiconductor for future electronics. Nat Rev Electr Eng 2:494\u0026ndash;513\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIncorvia JAC, Xiao TP, Zogbi N, Naeemi A, Adelmann C, Catthoor F, Tahoori M, Casanova F, Becherer M, Prenat G, Couet S (2024) Spintronics for achieving system-level energy-efficient logic. Nat Rev Electr Eng 1:700\u0026ndash;713\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen M, Ma Y, Aslam N, Liu C, Chen Y, Luo L, Zhang X, Mai K, Xiao H, Zhu K, Alharbi O, Zheng D, Xu X, Liao H, Yang Y, Wang H, Zhou Z, Wang H, Tian B, Li J, He X, Chang K, Wan Y, Shamim A, Alshareef HN, Lanza M, Anthopoulos TD, Han Z, Xue F, Zhang X (2025) Ultrawide-bandwidth boron nitride photonic memristors. Nat Nanotechnol 1\u0026ndash;8\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuo Z, Yin J, Bai Y, Zhu D, Shi K, Wang G, Cao K, Zhao W (2021) Spintronics for energy-efficient computing: an overview and outlook. \u003cem\u003eProc. IEEE\u003c/em\u003e 109, 1398\u0026ndash;1417\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhwa W-S, Wen T-H, Hsu H-H, Huang W-H, Chang Y-C, Chiu T-C, Ke Z-E, Chin Y-H, Wen H-J, Hsu W-T, Lo C-C, Liu R-S, Hsieh C-C, Tang K-T, Ho M-S, Lele AS, Teng S-H, Chou C-C, Chih Y-D, Chang T-YJ, Chang (2025) M.-F. A mixed-precision memristor and SRAM compute-in-memory AI processor. Nature 639:617\u0026ndash;623\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang Y, Cao K, Zhang K, Wang J, Shi K, Hao Z, Cai W, Du A, Yin J, Yang Q, Li J, Gao J, Zhao C, Zhao W (2023) Implementation of 16 Boolean logic operations based on one basic cell of spin-transfer-torque magnetic random access memory. Sci China Inf Sci 66:1\u0026ndash;8\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Pham T, Groen I, Manipatruni S, Choi WY, Nikonov DE, Sagasta E, Lin C-C, Gosavi TA, Marty A, Hueso LE, Young IA, Casanova F (2020) Spin\u0026ndash;orbit magnetic state readout in scaled ferromagnetic/heavy metal nanostructures. Nat Electron 3:309\u0026ndash;315\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eManipatruni S, Nikonov DE, Lin C-C, Gosavi TA, Liu H, Prasad B, Huang Y-L, Bonturim E, Ramesh R, Young IA (2019) Scalable energy-efficient magnetoelectric spin-orbit logic. Nature 565:35\u0026ndash;42\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVaz DC, Lin C-C, Plombon JJ, Choi WY, Groen I, Arango IC, Chuvilin A, Hueso LE, Nikonov DE, Li H, Debashis P, Clendenning SB, Gosavi TA, Huang Y-L, Prasad B, Ramesh R, Vecchiola A, Bibes M, Bouzehouane K, Fusil S, Garcia V, Young IA, Casanova F (2024) Voltage-based magnetization switching and reading in magnetoelectric spin-orbit nanodevices. Nat Commun 15:1902\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin C-C, Gosavi T, Nikonov D, Huang Y-L, Prasad B, Choi W, Pham VT, Groen I, Chen J-Y, Liu MDCH, Oguz K, Walker ES, Plombon J, Buford B, Naylor CH, Wang J-P, Casanova F, Ramesh R \u0026amp; I. A. Young. Experimental demonstration of integrated magneto-electric and spin-orbit building blocks implementing energy-efficient logic. In 2019 IEEE Int Electron Devices Meeting (IEDM), 37.3.1\u0026ndash;37.3.4 (IEEE 2019)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDebashis P, Plombon JJ, Lin C-C, Liao Y-C, Li H, Nikonov DE, Adams D, Rogan C, Dc M, Radosavljevic M, Clendenning SB, Young IA Low-voltage and high-speed switching of a magnetoelectric element for energy efficient compute. In 2022 IEEE Int Electron Devices Meeting (IEDM), 36.4.1\u0026ndash;36.4.4 (IEEE 2022)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLuo B, Will-Cole AR, Dong C, He Y, Liu X, Lin H, Huang R, Shi X, McConney M, Page M, Sanghadasa M, Ramesh R (2024) Sun, N. X. Magnetoelectric microelectromechanical and nanoelectromechanical systems for the IoT. Nat Rev Electr Eng 1:317\u0026ndash;334\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVaz DC, Lin C-C, Plombon J, Choi WY, Groen I, Arango I, Pham VT, Nikonov DE, Li H, Debashis P, Clendenning SB, Gosavi TA, Garcia V, Fusil S, Bibes M, Huang Y-L, Prasad B, Ramesh R, Casanova F \u0026amp; I. A. Young. Functional demonstration of a fully integrated magnetoelectric spin-orbit device. In 2021 IEEE Int Electron Devices Meeting (IEDM), 32.4.1\u0026ndash;32.4.4 (IEEE 2021)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGrezes C, Kandazoglou A, Cosset-Cheneau M, Arche LMV, No\u0026euml;l P, Sgarro P, Auffret S, Garello K, Bibes M, Vila L, Attan\u0026eacute; J-P (2023) Non-volatile electric control of spin-orbit torques in an oxide two-dimensional electron gas. Nat Commun 14:2590\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDing J, Liu C, Kalappattil V, Zhang Y, Mosendz O, Erugu U, Yu R, Tian J, DeMann A, Field SB, Yang X, Ding H, Tang J, Terris B, Chen FERTA, H., Wu M (2021) Switching of a magnet by spin-orbit torque from a topological dirac semimetal. Adv Mater 33:2005909\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFan Y, Upadhyaya P, Kou X, Lang M, Takei S, Wang Z, Tang J, He L, Chang L-T, Montazeri M, Yu G, Jiang W, Nie T, Schwartz RN, Tserkovnyak Y, Wang KL (2014) Magnetization switching through giant spin-orbit torque in a magnetically doped topological insulator heterostructure. Nat Mater 13:699\u0026ndash;704\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMellnik AR, Lee JS, Richardella A, Grab JL, Mintun PJ, Fischer MH, Vaezi A, Manchon A, Kim E-A, Samarth N, Ralph DC (2014) Spin-transfer torque generated by a topological insulator. Nature 511:449\u0026ndash;451\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu H, Chen A, Zhang P, He H, Nance J, Guo C, Sasaki J, Shirokura T, Hai PN, Fang B, Razavi SA, Wong K, Wen Y, Ma Y, Yu G, Carman GP, Han X, Zhang X, Wang KL (2021) Magnetic memory driven by topological insulators. Nat Commun 12:6251\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVicente-Arche LM, Br\u0026eacute;hin J, Varotto S, Cosset-Cheneau M, Mallik S, Salazar R, No\u0026euml;l P, Vaz DC, Trier F, Bhattacharya S, Sander A, Le F\u0026egrave;vre P, Bertran F, Saiz G, M\u0026eacute;nard G, Bergeal N, Barth\u0026eacute;l\u0026eacute;my A, Li H, Lin C-C, Nikonov DE, Young IA, Rault JE, Vila L, Attan\u0026eacute; J-P, Bibes M (2021) Spin-charge interconversion in KTaO\u003csub\u003e3\u003c/sub\u003e 2D electron gases. Adv Mater 33:2102102\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChoi WY, Arango IC, van Pham T, Vaz DC, Yang H, Groen I, Lin C-C, Kabir ES, Oguz K, Debashis P, Plombon JJ, Li H, Nikonov DE, Chuvilin A, Hueso LE, Young IA, Casanova F (2022) All-electrical spin-to-charge conversion in sputtered Bi\u003csub\u003ex\u003c/sub\u003eSe\u003csub\u003e1-x\u003c/sub\u003e. Nano Lett 22:7992\u0026ndash;7999\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGallego F, Trier F, Mallik S, Br\u0026eacute;hin J, Varotto S, Vicente-Arche M, Gosavy L, Lin T, Coudevylle C-C, Iglesias J-R, Casanova L, Young F, Vila I, Attan\u0026eacute; L, J.-P., Bibes M (2024) All-electrical detection of the spin-charge conversion in nanodevices based on SrTiO\u003csub\u003e3\u003c/sub\u003e 2-D electron gases. Adv Func Mater 34:2307474\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParkin SSP, Kaiser C, Panchula A, Rice PM, Hughes B, Samant M, Yang S-H (2004) Giant tunnelling magnetoresistance at room temperature with MgO (100) tunnel barriers. Nat Mater 3:862\u0026ndash;867\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYuasa S, Djayaprawira DD (2007) Giant tunnel magnetoresistance in magnetic tunnel junctions with a crystalline MgO(001) barrier. J Phys D: Appl Phys 40:R337\u0026ndash;R354\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYuasa S, Nagahama T, Fukushima A, Suzuki Y, Ando K (2004) Giant room-temperature magnetoresistance in single-crystal Fe/MgO/Fe magnetic tunnel junctions. Nat Mater 3:868\u0026ndash;871\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMathon J, Umerski A (2001) Theory of tunneling magnetoresistance of an epitaxial Fe/MgO/Fe(001) junction. Phys Rev B 63:220403\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eM.Julliere (1975) Tunneling between ferromagnetic films. Phys Lett A 54:225\u0026ndash;226\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang W, Liu E, Kodzuka M, Sukegawa H, Wojcik M, Jedryka E, Wu GH, Inomata K, Mitani S, Hono K (2010) Coherent tunneling and giant tunneling magnetoresistance in Co\u003csub\u003e2\u003c/sub\u003eFeAl/MgO/CoFe magnetic tunneling junctions. Phys Rev B 81\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eScheike T, Wen Z, Sukegawa H, Mitani S (2023) 631% room temperature tunnel magnetoresistance with large oscillation effect in CoFe/MgO/CoFe(001) junctions. Appl Phys Lett 122\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStrijkers GJ, Ji Y, Yang FY, Chien CL, Byers JM (2001) Andreev reflections at metal/superconductor point contacts: measurement and analysis. Phys Rev B 63\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKaiser C, Parkin SSP (2004) Spin polarization in ferromagnet/insulator/superconductor structures with the superconductor on top of the barrier. Appl Phys Lett 84:3582\u0026ndash;3584\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePaluskar PV, Attema JJ, de Wijs GA, Fiddy S, Snoeck E, Kohlhepp JT, Swagten HJM, de Groot RA, Koopmans B (2008) Spin tunneling in junctions with disordered ferromagnets. Phys Rev Lett 100:57205\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSinova J, Valenzuela SO, Wunderlich J, Back CH, Jungwirth T (2015) Spin Hall effects. Rev Mod Phys 87:1213\u0026ndash;1260\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePeng S, Zheng X, Li S, Lao B, Han Y, Liao Z, Zheng H, Yang Y, Yu T, Liu P, Sun Y, Chen X-Q, Peng S, Zhao W, Li R-W, Wang Z (2025) Unconventional scaling of the orbital Hall effect. Nat Mater 1\u0026ndash;7\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZheng Z, Jia L, Zhang Z, Shen Q, Zhou G, Cui Z, Ren L, Chen Z, Jamaludin NF, Zhao T, Xiao R, Zhang Q, Du Y, Liu L, Gradečak S, Novoselov KS, Zhao W, Xu X, Zhang Y, Chen J (2025) All-electrical perpendicular switching of chiral antiferromagnetic order. Nat Mater 1\u0026ndash;7\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTakeuchi Y, Sato Y, Yamane Y, Yoon J-Y, Kanno Y, Uchimura T, de Zoysa KV, Han J, Kanai S, Ieda J, Ohno H, Fukami S (2025) Electrical coherent driving of chiral antiferromagnet. Science 389:830\u0026ndash;834\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu L, Chen C-T, Sun JZ (2014) Spin Hall effect tunnelling spectroscopy. Nat Phys 10:561\u0026ndash;566\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGroen I, van Pham T, Leo N, Marty A, Hueso LE, Casanova F (2021) Disentangling spin, anomalous, and planar Hall effects in ferromagnet\u0026ndash;heavy-metal nanostructures. Phys Rev Appl 15:044010\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMizuno H, Isshiki H, Kondou K, Zhu Y, Otani Y (2021) Influence of planar Hall effect on the output signal in a T-shaped spin conversion device. Appl Phys Let 119:0061147\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePashenkin IY, Sapozhnikov MV, Gusev NS, Karashtin EA, Fraerman AA (2022) Extrinsic tunnel Hall effect in MgO-based tunnel junctions. Phys Rev B 106:L220408\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTeresi S, Sebe N, Patterson J, Frottier T, Kandazoglou A, No\u0026euml;l P, Sgarro P, T\u0026eacute;r\u0026eacute;b\u0026eacute;nec D, Bernier N, Hippert F, Attan\u0026eacute; J-P, Vila L, No\u0026eacute; P, Cosset-Ch\u0026eacute;neau M (2023) Spin-orbit readout using thin films of topological insulator Sb\u003csub\u003e2\u003c/sub\u003eTe\u003csub\u003e3\u003c/sub\u003e deposited by industrial magnetron sputtering. Adv Func Mater 33:02303878\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin Y-L, Wu Y-H, Lee S-K, Cheng C-W, Chen J-W, Chung C-H, Song M-Y, Bao X, Tseng Y-C (2024) Scalable spin-to-charge conversion device with nanopatterning. ACS Appl Electron Mater 6:6657\u0026ndash;6664\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePai C-F, Liu L, Li Y, Tseng HW, Ralph DC, Buhrman RA (2012) Spin transfer torque devices utilizing the giant spin Hall effect of tungsten. Appl Phys Lett 101:122404\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSriram K, Mondal R, Pradhan J, Haldar A, Murapaka C (2023) Structural phase engineering of (α\u0026thinsp;+\u0026thinsp;β)-W for a large spin Hall angle and spin diffusion length. J Phys Chem C 127:22704\u0026ndash;22712\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSun H-H, Zhang K-W, Hu L-H, Li C, Wang G-Y, Ma H-Y, Xu Z-A, Gao C-L, Guan D-D, Li Y-Y, Liu C, Qian D, Zhou Y, Fu L, Li S-C, Zhang F-C, Jia J-F (2016) Majorana zero mode detected with spin selective Andreev reflection in the vortex of a topological superconductor. Phys Rev Lett 116:257003\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcKee SA (2004) Reflections on the memory wall. In \u003cem\u003e2004 Computer Frontiers Conference\u003c/em\u003e, p. 162ACM Press\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHafner J (2008) Ab-initio simulations of materials using VASP: Density-functional theory and beyond. J Comput Chem 29:2044\u0026ndash;2078\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTaylor J, Guo H, Wang J (2001) Ab initio modeling of quantum transport properties of molecular electronic devices. Phys Rev B 63:245407\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Magnetoelectric spin-orbit logic, magnetic tunnel junction, spin-to-charge conversion, spin-orbit coupling, spin polarization","lastPublishedDoi":"10.21203/rs.3.rs-7680431/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7680431/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMagnetoelectric spin-orbit (MESO) logic, composed of a voltage-controlled magnetoelectric writing module and a spin-orbit readout module, is highly expected to substitute the silicon-based transistors and enables energy-efficient and scalable computing. Nevertheless, the output voltage of readout module based on spin-to-charge conversion is far less than the minimum magnetoelectric writing voltage, which greatly restricts the cascading function of MESO logic. Here, we first propose a magnetic tunnel junction (MTJ)-enhanced MESO logic to implement giant readout signal. Up to 1.5 mV output voltage is obtained, marking a significant improvement of approximately two orders of magnitude compared to previous findings. We ascribe the substantial enhancement to junction resistance modulation and the spin filtering effect of MgO-single-crystal MTJ. Moreover, the naturally integrated MTJ and MESO enables instantaneous and nonvolatile data exchange between computing module and external unit. Our work not only enhances output signal of readout module for direct cascading of MESO logic but also refines the design architecture, marking a pivotal stride forward in propelling MESO technology toward practical applications.\u003c/p\u003e","manuscriptTitle":"Giant Spin-Orbit Magnetic State Readout Enhanced by a Magnetic Tunnel Junction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-28 15:27:29","doi":"10.21203/rs.3.rs-7680431/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"2d1ae50d-4af2-41e5-bc7e-09a91272c1cf","owner":[],"postedDate":"October 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":55391135,"name":"Physical sciences/Nanoscience and technology/Nanoscale devices/Electronic devices"},{"id":55391136,"name":"Physical sciences/Nanoscience and technology/Nanoscale devices/Magnetic devices"}],"tags":[],"updatedAt":"2025-10-28T15:27:30+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-28 15:27:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7680431","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7680431","identity":"rs-7680431","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}