Magnetic Random Access Memory Research Articles

The central role of tilted anisotropy for field-free spin–orbit torque switching of perpendicular magnetization

The discovery of efficient magnetization switching upon device activation by spin Hall effect (SHE)-induced spin–orbit torque (SOT) changed the course of magnetic random-access memory (MRAM) research and development. However, for electronic systems with perpendicular magnetic anisotropy (PMA), the use of SOT is still hampered by the necessity of a longitudinal magnetic field to break magnetic symmetry and achieve deterministic switching. In this work, we demonstrate that robust and tunable field-free current-driven SOT switching of perpendicular magnetization can be controlled by the growth protocol in Pt-based magnetic heterostructures. We further elucidate that such growth-dependent symmetry breaking originates from the laterally tilted magnetic anisotropy of the ferromagnetic layer with PMA, a phenomenon that has been largely neglected in previous studies. We show experimentally and in simulation that in a PMA system with tilted anisotropy, the deterministic field-free switching exhibits a conventional SHE-induced damping-like torque feature, and the resulting current-induced effective field shows a nonlinear dependence on the applied current density. This relationship could be potentially misattributed to an unconventional SOT origin.

Spin orbit magnetic random access memory based binary CNN in-memory accelerator (BIMA) with sense amplifier

The research tends to suggest a spin-orbit torque magnetic random access memory (SOT-MRAM)-based Binary CNN In-Memory Accelerator (BIMA) to minimize power utilization and suggests an In-Memory Computing (IMC) for AdderNet-based BIMA to further enhance performance by fully utilizing the benefits of IMC as well as a low current consumption configuration employing SOT-MRAM. And recommended an IMC-friendly computation pipeline for AdderNet convolution at the algorithm level. Additionally, the suggested sense amplifier is not only capable of the addition operation but also typical Boolean operations including subtraction etc. The architecture suggested in this research consumes less power than its spin-orbit torque (STT) MRAM and resistive random access memory (ReRAM)-based counterparts in the Modified National Institute of Standards and Technology (MNIST) data set, according to simulation results. Based to evaluation outcomes, the pre-sented strategy outperforms the in-memory accelerator in terms of speedup and energy efficiency by 17.13× and 18.20×, respectively.

Interfacial modulation of magnetic relaxation and electrical characteristic in RuO2/CrO2 antiferromagnet-half metal bilayer

Half-metallic chromium dioxide (CrO2) is a promising candidate in magnetic random-access memory (MRAM) devices due to its nearly full spin polarization and ultralow Gilbert damping at room temperature. In this study, we succeeded in fabricating (110) and (100)-oriented RuO2/CrO2 bilayers with different thicknesses of CrO2 film using two-step method. The CrO2 film epitaxially grown on a RuO2 layer exhibits a perfect in-plane magnetic uniaxial anisotropy. The magnetic relaxation of CrO2 films can be strongly modulated via the proximity to RuO2 layer, which may relate to the enhancement of the two-magnon scattering due to effects of antiferromagnetic order in RuO2 film. In addition, electronic transport measurements indicate that the electronic characteristic is strongly related to the film growth orientation, and a linear magnetoresistance is observed stemming from intergrain tunneling magnetoresistance and double-exchange mechanism. Angle-dependent ferromagnetic resonance measurements indicates that the uniaxial anisotropy in CrO2 film is strongly related to the strain effect, which is consistent with reciprocal space mapping results. This work lays a foundation for current-drive MRAM based on RuO2/CrO2 bilayer and provides a feasible damping modulation method for CrO2 film.

Thermodynamic properties and switching dynamics of perpendicular shape anisotropy MRAM

The power consumption of modern random access memory (RAM) has been a motivation for the development of low-power non-volatile magnetic RAM (MRAM). Based on a CoFeB/MgO magnetic tunnel junction, MRAM must satisfy high thermal stability and a low writing current while being scaled down to a sub-20 nm size to compete with the densities of current RAM technology. A recent development has been to exploit perpendicular shape anisotropy along the easy axis by creating tower structures, with the free layers’ thickness (along the easy axis) being larger than its width. Here we use an atomistic model to explore the temperature dependent properties of thin cylindrical MRAM towers of 5 nm diameter while scaling down the free layer from 48 to 8 nm thick. We find thermal fluctuations are a significant driving force for the switching mechanism at operational temperatures by analysing the switching field distribution from hysteresis data. We find that a reduction of the free layer thickness below 18 nm rapidly loses shape anisotropy, and consequently stability, even at 0 K. Additionally, there is a change in the switching mechanism as the free layer is reduced to 8 nm. Coherent rotation is observed for the 8 nm free layer, while all taller towers demonstrate incoherent rotation via a propagated domain wall.

Single-nanometer CoFeB/MgO magnetic tunnel junctions with high-retention and high-speed capabilities

Making magnetic tunnel junctions (MTJs) smaller while meeting performance requirements is critical for future electronics with spin-transfer torque magnetoresistive random access memory (STT-MRAM). However, it is challenging in the conventional MTJs using a thin CoFeB free layer capped with an MgO layer because of increasing difficulties in satisfying the required data retention and switching speed at smaller scales. Here we report single-nanometer MTJs using a free layer consisting of CoFeB/MgO multilayers, where the number of CoFeB/MgO interfaces and/or the CoFeB thicknesses are engineered to tailor device performance to applications requiring high-data retention or high-speed capability. We fabricate ultra-small MTJs down to 2.0 nm and show high data retention (over 10 years) and high-speed switching at 10 ns or below in sub-5-nm MTJs. The stack design proposed here proves that ultra-small CoFeB/MgO MTJs hold the potential for high-performance and high-density STT-MRAM.

Tailoring Interlayer Chiral Exchange by Azimuthal Symmetry Engineering.

Recent theoretical and experimental studies of the interlayer Dzyaloshinskii-Moriya interaction (DMI) have sparked great interest in its implementation into practical magnetic random-access memory (MRAM) devices, due to its capability to mediate long-range chiral spin textures. So far, experimental reports focused on the observation of interlayer DMI, leaving the development of strategies to control interlayer DMI's magnitude unaddressed. Here, we introduce an azimuthal symmetry engineering protocol capable of additive/subtractive tuning of interlayer DMI through the control of wedge deposition of separate layers and demonstrate its capability to mediate field-free spin-orbit torque (SOT) magnetization switching in both orthogonally magnetized and synthetic antiferromagnetically coupled systems. Furthermore, we showcase that the spatial inhomogeneity brought about by wedge deposition can be suppressed by specific azimuthal engineering design, ideal for practical implementation. Our findings provide guidelines for effective manipulations of interlayer DMI strength, beneficial for the future design of SOT-MRAM or other spintronic devices utilizing interlayer DMI.

Advanced hybrid MRAM based novel GPU cache system for graphic processing with high efficiency

With the rapid development of portable computing devices and users’ demand for high-quality graphics rendering, embedded Graphics Processing Units (GPU) systems for graphics processing are increasingly turning into a key component of computer architecture to enhance computability. The cache system based on traditional static random access memory (SRAM) plays a crucial role in GPUs. But high leakage, low lifetime and poor integration problems deeply plague the science and engineering field. In the paper, a novel magnetic random access memory (MRAM) based cache architecture of GPU systems is proposed for highly efficient graphics processing and computing accelerating, with the merits of high speed, long endurance, strong interference resistance, and ultra-low power consumption. Spin transfer torque-MRAM and spin orbit torque-MRAM are utilized in off-chip and on-chip caches, respectively. A controller design scheme with prefetching modules and optimized cache coherency protocols are adopted. After testing and evaluating with multiple loads, neural network models and datasets, the simulation results show that the proposed system can achieve up to 28%, 56%, and 66.45% optimizations mostly in terms of speed, energy and leakage power, respectively.

Impact of Technology Scaling and Back-End-of-the-Line Technology Solutions on Magnetic Random-Access Memories

Impact of Technology Scaling and Back-End-of-the-Line Technology Solutions on Magnetic Random-Access Memories

A High-Speed, Low-Power, High-Reliability and Fully Single Event Double Node Upset Tolerant Design for Magnetic Random Access Memory

A High-Speed, Low-Power, High-Reliability and Fully Single Event Double Node Upset Tolerant Design for Magnetic Random Access Memory

Novel CPU cache architecture based on two-dimensional MTJ device with ferromagnetic Fe3GeTe2

With the development of Artificial Intelligence (AI) in recent years, the fields of computer, biology, medicine, and aerospace have demanded higher requirements for the processing and storage of information. In this paper, a novel Magnetic Tunnel Junction (MTJ) based Spin-Orbital Torque Magnetic Random Access Memory (SOT-MRAM) composed of Fe3GeTe2 (FGT) is employed as a storage medium in the computer architecture. On the basis of the analysis of the fundamentals, model configuration, characteristics and performance advantages of the FGT based SOT device, a hybrid storage (L1, L2, Last Level Cache) is constructed, with FGT-SOT-MRAM, conventional SOT-MRAM and STT-MRAM replacing the original static random access memory (SRAM) in the novel triple-level CPU cache architecture. This can override the increasing leakage problem of SRAM, while opening up the application of two-dimensional van der Waals ferromagnets in computer systems at the L1 cache level. Meanwhile, an innovative cache optimization scheme is put forward for magnetic memory to better match the performance of FGT-SOT-MRAM to CPU. The simulation results demonstrate that the FGT-based MRAM can achieve up to 38.03% IPC optimization and 53.41% power optimization in the CPU cache system in contrast to the conventional ones.

Field-free magnetic switching dependence on lateral interfaces in synthetic antiferromagnets by ion implantation

Field-free spin–orbit torque switching in synthetic antiferromagnets (SAF) holds significant promise for high-density spintronic memory and logic devices. In this paper, we realize the field-free magnetization switching in SAFs due to the local ion implantation-induced 45° lateral interface and symmetry breaking. Moreover, the magnetization switching ratio is enlarged by the lateral interface owing to the superimposition of a damping-like effective field and a symmetry-breaking effective field. Our work is significant for the development of magnetic random-access memory technology with high-speed and anti-interference ability.

Magnetization switching driven by spin-orbit torque of Weyl semimetal WTe<sub>2</sub>

The Wely semimetal WTe<sub>2</sub> exhibits significant spin-orbit coupling characteristics and can generate unconventional spin current with out-of-plane polarization, which has become a hotspot in recent years. Meanwhile, WTe<sub>2</sub> also has high charge-spin conversion efficiency, allowing perpendicular magnetization to be switched deterministically without the assistance of an external magnetic field, which is critical for the high-density integration of low-power magnetic random-access memories. The purpose of this paper is to review the recent advances in the research on spin orbit torque in heterostructures composed of WTe<sub>2</sub> and ferromagnetic layers, focusing on progress of research on the detection and magnetization switching in the spin orbit torque of heterojunctions composed of WTe<sub>2</sub> prepared by different methods (e.g. mechanical exfoliation and chemical vapor deposition) and ferromagnetic layers such as conventional magnets (e.g, FeNi and CoFeB, etc.) and two-dimensional magnets (e.g. Fe<sub>3</sub>GeTe<sub>2</sub>, etc.). Finally, the prospect of related research is discussed.

External fields effectively switch the spin channels of transition metal-doped β-phase tellurene from first principles.

Non-volatile magnetic random-access memories have proposed the need for spin channel switching. However, this presents a challenge as each spin channel reacts differently to the external field. Tellurene is a semiconductor with a tunable bandgap, excellent stability, and high carrier concentration, but its lack of magnetic properties has hindered its application in spintronics. In this work, the influence of an external field on transition metal (TM)-doped β-tellurene is systematically analysed from first principles. First, the active-learning moment-tensor-potential (MTP) is used to verify the thermal stability of the V-doped system with the MTP proving to be 900 times faster than the traditional method. Subsequently, under biaxial strain ranging from -2% to 10%, the V-doped system undergoes a gradual transition from a magnetic semiconductor to a spin-gapless semiconductor, and further to a half-metal and magnetic metal. The band structure can be maintained under an electric field. By examining the magnetic anisotropy energy, the lattice changes profoundly impact the electromagnetic properties, particularly with the TMs being sensitive to strain. Moreover, the band structure is reflected in the spin resolution current of the magnetic tunnel junction. This work investigates the response of doped β-Te to external fields, revealing its potential applications in spintronics.

Skyrmion Based Synaptic Device Having Perpendicular Magnetic-Tunnel-Junction Spin-Valve Sensor

Perpendicular spin-transfer-torque magnetic random access memory (p-STT MRAM) has been reported as an application of binary neural networks (BNN) [1]. In addition, magnetic skyrmions have been reported as information carriers for neuromorphic devices due to their small nanoscale diameter, topological stability, and complete linearity [2-3]. A previous study has demonstrated the synaptic characteristics of skyrmion based synapses using a Hall bar measurement [4]. However, skyrmion based synapses using the Hall bar measurement were not able to detect electrically a synaptic behavior. Thus, for the first time, the skyrmion based synaptic device detected resistance change via sensing the perpendicular magnetic-tunnel-junction (p-MTJ) spin-valve, as shown in Fig.1(a). The detailed synaptic behavior and mechanism will be presented.In this study, we introduced the skyrmion based synaptic device having p-MTJ spin-valve sensor designed by CoFeB free layer, MgO tunneling barrier, CoFeB pinned layer, SyAF layer, and interface reducing W layer between the CoFeB free layer and MgO tunneling barrier, as shown in Fig.1(a). The spin direction of the CoFeB free layer was changed by the generated skyrmions, and the resistance of the synaptic device changed depending on the parallel or anti-parallel state between CoFeB free and pinned layer. In particular, the skyrmion generation was determined critically by the combination of DMI and PMA (i.e., saturation magnetization (Ms), squareness), and the interface reducing W layer was designed to reduce PMA. The M-H loop of the synaptic device with the interface reducing W layer thickness of 0.110 nm showed reduced saturation magnetization (Ms ~ 60 μemu) and squareness (Hc ~ 8 Oe) via a vibrating-sample magnetometer (VSM), as shown in [Figs.1(b), 1(c), and 1(d)]. In addition, uniform state, generation of skyrmions, and stripe domains in the synaptic device were achieved at the external magnetic field of 0, 3.5, and 7 Oe, respectively, as shown in [Figs.1(e), 1(f), and 1(g)]. Moreover, we observed that the number of skyrmions in the synaptic device with a cell diameter of 5μm changed depending on the external magnetic field via the magneto-optical Kerr effect (MOKE) image, as shown in [Figs.1(h), 1(i), and 1(j)]. Finally, we demonstrated a multi-level operation in which the electrical resistance of the synaptic device changed as the number of skyrmions. In our presentation, we will review in detail the synaptic behavior and mechanism for the synaptic device. Acknowledgements This research was supported by National R&D Program through the National Research Foundation of Korea (NRF) funded by Ministry of Science and ICT (2021M3F3A2A01037733) References [1] Thi-Nhan Pham et al., IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS, 12, 569-579 (2022)[2] Nagaosa, N. & Tokura, Y. Nat. Nanotechnol. 8, 899–911 (2013).[3] C Back, V Cros et al., J. phys. D. Appl. phys. 53, 363001 (2020).[4] Kyung Mee Song et al., Nature Electronics. 3, 148–155 (2020) Figure 1

A novel time-domain in-memory computing unit using STT-MRAM

Advancements in technologies like Big Data, IoT, and AI have revealed a bottleneck in traditional von-Neumann architecture, resulting in high energy consumption and limited memory bandwidth. In-memory computing (IMC) presents a promising solution by enabling computations directly within the memory, enhancing energy-efficient computing. The existing time-domain (TD)-based IMC computations either require multiple cycles for computation through a successive read/write approach or contribute to the complexity of the peripheral circuit by adopting a cumulative delay approach. In this paper, we present a novel array architecture that utilizes spin transfer torque magnetic random access memory (STT-MRAM) bit-cells, mitigating source degeneration issue. By leveraging this advanced technology and employing a TD computing scheme, we have successfully implemented various arithmetic operations, alongside a comprehensive set of Boolean logic operations. Our design demonstrates improved area and energy efficiency compared to other existing TD computing schemes. Furthermore, despite the higher delay, our parameter-driven optimization approach efficiently minimizes it. To validate our proposal, we performed simulations using the 45 nm CMOS process and the Verilog-A based magnetic tunnel junction (MTJ) compact model. Through meticulous Monte-Carlo simulations, considering CMOS variations, the results demonstrate enhanced computational accuracy with increasing Tunnel Magnetoresistance (TMR) ratio, showcasing the potential of our architecture in advancing the field of computing.

Large enhancement of spin–orbit torques under a MHz modulation due to phonon–magnon coupling

The discovery of spin–orbit torques (SOTs) generated through the spin Hall or Rashba effects provides an alternative write approach for magnetic random-access memory (MRAM), igniting the development of spin orbitronics in recent years. Quantitative characterization of SOTs highly relies on the SOT-driven ferromagnetic resonance (ST-FMR), where a modulated microwave current is used to generate ac SOTs and the modulation frequency is usually less than 100 kHz (the limit of conventional lock-in amplifiers). Here, we have investigated the SOT of typical SOT material/ferromagnet bilayers in an extended modulation-frequency range, up to MHz, by developing the ST-FMR measurement. Remarkably, we found that the measured SOTs are enhanced about three times in the MHz range, which cannot be explained according to the present SOT theory. We attribute the enhancement of SOT to additional magnon excitations due to phonon–magnon coupling, which is also reflected in the slight changes of resonant field and linewidth in the acquired ST-FMR spectra, corresponding to the modifications of effective magnetization and damping constant, respectively. Our results indicate that the write current of SOT-MRAM may be reduced with the assistance of phonon–magnon coupling.

Magnetic random-access memory with disturb-free read

Magnetic random-access memory with disturb-free read

Demonstration of a manufacturable SOT-MRAM multiplexer array towards industrial applications

We have successfully demonstrated a 1 Kb spin-orbit torque (SOT) magnetic random-access memory (MRAM) multiplexer (MUX) array with remarkable performance. The 1 Kb MUX array exhibits an in-die function yield of over 99.6%. Additionally, it provides a sufficient readout window, with a TMR/R P_sigma% value of 21.4. Moreover, the SOT magnetic tunnel junctions (MTJs) in the array show write error rates as low as 10−6 without any ballooning effects or back-hopping behaviors, ensuring the write stability and reliability. This array achieves write operations in 20 ns and 1.2 V for an industrial-level temperature range from −40 to 125 °C. Overall, the demonstrated array shows competitive specifications compared to the state-of-the-art works. Our work paves the way for the industrial-scale production of SOT-MRAM, moving this technology beyond R&D and towards widespread adoption.

Large voltage-controlled magnetic anisotropy effect in magnetic tunnel junctions prepared by deposition at cryogenic temperatures

We investigated the influence of the buffer material and a cryogenic temperature deposition process on the voltage-controlled magnetic anisotropy (VCMA) effect for an ultrathin CoFeB layer in bottom-free type MgO-based magnetic tunnel junctions prepared by a mass production sputtering process. We used Ta and TaB buffers and compared the differences between them. The TaB buffer enabled us to form a flat and less-contaminated CoFeB/MgO interface by suppressing the diffusion of Ta with maintaining a stable amorphous phase. Furthermore, the introduction of cryogenic temperature deposition for the ultrathin CoFeB layer on the TaB buffer improved the efficiency of the VCMA effect and its annealing tolerance. Combining this with interface engineering employing an Ir layer for doping and a CoFe termination layer, a large VCMA coefficient of −138 ± 3 fJ/Vm was achieved. The developed techniques for the growth of ultrathin ferromagnet and oxide thin films using cryogenic temperature deposition will contribute to the development of high-performance spintronic devices, such as voltage-controlled magnetoresistive random access memories.

IMGA: Efficient In-Memory Graph Convolution Network Aggregation With Data Flow Optimizations

Aggregating features from neighbor vertices is a fundamental operation in Graph Convolution Network (GCN). However, the sparsity in graph data creates poor spatial and temporal locality, causing dynamic and irregular memory access patterns and limiting the performance of aggregation on the Von Neumann architecture. The emerging processing-in-memory (PIM) architecture is based on emerging non-volatile memory (NVM), like Spin-orbit torque Magnetic RAM (SOT-MRAM), and demonstrates promising prospects in alleviating the Von Neumann bottleneck. However, the limited memory capacity of PIM medium still incurs non-negligible data movements between PIM architecture and external memory. To solve this challenge, we propose a SOT-MRAM based in-memory computing architecture, called IMGA, for efficient in-situ graph aggregation. Specifically, we design adaptive data flow management strategies that reuse vertex data in MRAM when processing graphs of different scales and adopt edge data as the control signal source to utilize the graph’s structural information. A reordering optimization strategy leveraging hardware-software co-design principle is proposed to further reduce the costly data movement. Experimental results demonstrate that IMGA achieves an average 2523x and 21x speedup, and 1.03E+6 and 1.04E+3 energy efficiency compared with CPU and GPU, respectively.