Scalable Three Terminal Memory Devices based on Silicon-Compatible Antiferromagnetic Materials

Northwestern University, Evanston IL

Investigators

Abstract

Antiferromagnetic materials are magnetically ordered materials without a net macroscopic magnetization. They offer characteristics that make them promising for high-density, fast, and low-power nonvolatile memory devices: Firstly, there is no inter-bit dipole interaction between neighboring bits, making it possible to place adjacent bits closer together than with ferromagnetic devices. Secondly, antiferromagnets can potentially have ultrafast write times in the picosecond range, due to their exchange-dominated dynamics. Thirdly, devices based on these materials would be immune to tampering by magnetic fields, making them more secure than ferromagnet-based devices. However, to date, the realization of these devices has been hindered by the difficulty of electrically interacting with antiferromagnetic materials, particularly in industry-relevant materials and at nanoscale dimensions. This project aims to develop three-terminal antiferromagnetic memory devices with full electrical read and write capability, based on silicon-compatible antiferromagnetic materials that can be integrated in existing semiconductor manufacturing processes. This project will have a significant economic impact by enabling magnetic memories to address broader markets than currently possible, including dynamic random-access memory and embedded static random-access memory. In addition to its economic impact, this project will achieve broader impact through the incorporation of significant outreach and education activities. This includes outreach to the broader public through performance arts (science-themed plays and film screenings) through the Engineering Transdisciplinary Outreach Project in the Arts at Northwestern University. Research results will also be integrated into a newly developed course that the PI is teaching at Northwestern University, which focuses on the fundamentals and applications of magnetism and spintronics. The intellectual merit of this proposal is in the device design, modeling, fabrication, and electrical characterization of three-terminal nonvolatile memory devices. The devices will be primarily based on noncollinear conductive antiferromagnetic materials that can be sputter-deposited on silicon substrates (e.g., SnMn3, GaMn3, IrMn3 and related compounds). The magnetic state will be controlled electrically via current-induced spin-orbit torque from an adjacent heavy metal. The project aims to demonstrate scaling of this device operation down to industry-relevant bit diameters. The project will also integrate the electrical write mechanism with electrical readout via a tunneling magnetoresistance structure, translating the Néel vector modification into an electrical resistance change. The separation of read and write paths allows for higher cycling endurance and separate optimization of material parameters for the read and write steps. Thermal budgets, statistical variations induced by device variability, and other effects of the fabrication process on device characteristics will be investigated. The developed memory devices will be fully characterized to compare their performance to existing ferromagnet-based memory devices in terms of write time, write energy, write error rates (i.e., switching probability), read disturb rates, and cycling endurance. The device-level tradeoffs (e.g., between speed and endurance) will be characterized and compared to other existing and emerging memory technologies. Modeling will be performed both at the micromagnetic level to aid in device design, and at the physics-based compact model level, to allow for implementation into circuit design environments. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.