The continuous growth in the demand to store, process, and access data is one of the main drivers of spintronic research. Emerging big data and AI computing technologies require writing data in less time and storing them at a smaller scale, and consuming less energy 1. Magnetic memories (MRAM) using spin-transfer torques have proven their place among leading non-volatile memory technologies and are currently in production for eFlash replacement applications 2. In parallel, spin-orbit torques (SOT) have emerged and proven to be an efficient way of writing reliably magnetization at nanosecond time scale, broadening the scope of MRAM to applications that run close to the clock speed of the central processing unit, leading to novel spintronic memory and computing approaches 3, 4. I will review across this presentation the fundamental characteristics of SOT and their use to switch perpendicularly magnetized magnetic tunnel junction (MTJ) devices. I will cover different challenges to take up SOT-MRAM from material and stack optimization to technology large-scale integration and circuit design 5. After describing typical full-scale integration process of SOT-MRAM devices on 300mm wafers using CMOS compatible processes, I will discuss manufacturable field free methods and figures of merit for SOT-based memories. Finally, I will introduce bit-cell design considerations on scaling and performance in the perspective of circuit macro-design architectures 6, and discuss the interest to combine SOT, STT and voltage-gate (VCMA) effects to improve density and leverage performances 7. I will conclude with opportunities to diverse SOT-MRAM application spectra, notably in the field of in-memory computing 8.

References:

1 G. Molas and E. Nowak, “Advances in emerging memory technologies: From data storage to artificial intelligence,” Appl. Sci., vol. 11, no. 23, (2021), 2 B. Dieny et al., “Opportunities and challenges for spintronics in the microelectronics industry,” Nat. Electron., vol. 3, no. 8, pp. 446–459 (2020), 3 Q. Shao et al., “Roadmap of Spin-Orbit Torques,” IEEE Trans. Magn., vol. 57, no. 7 (2021), 4 A. Manchon et al., “Current-induced spin-orbit torques in ferromagnetic and antiferromagnetic systems,” Rev. Mod. Phys., vol. 91, no. 3 (2019), 5 K. Garello, F. Yasin, and G. S. Kar, “Spin-orbit torque MRAM for ultrafast embedded memories: From fundamentals to large scale technology integration,” IEEE IMW 2019, pp. 2019–2022 (2019), 6 M. Gupta et al., “High-density SOT-MRAM technology and design specifications for the embedded domain at 5nm node,” Tech. Dig. - IEDM, vol. 2020-Decem, pp. 24.5.1-24.5.4 (2020), 7 Y. C. Wu et al., “Voltage-Gate-Assisted Spin-Orbit-Torque Magnetic Random-Access Memory for High-Density and Low-Power Embedded Applications,” Phys. Rev. Appl., vol. 15, no. 6, pp. 1–10 (2021), 8 J. Doevenspeck et al., “SOT-MRAM based Analog in-Memory Computing for DNN inference,” Dig. Tech. Pap. - Symp. VLSI Technol., vol. 2020-June, no. 1, pp. 2020–2021 (2020)