The Potential of MRAM in Advanced Technology Nodes

A recent technical paper titled “Impact of Technology Scaling and Back-End-of-the-Line Technology Solutions on Magnetic Random-Access Memories” has been published by a team of researchers from the Georgia Institute of Technology. The study explores the challenges and potential solutions for implementing magnetic random-access memories (MRAM) at advanced technology nodes, with a particular focus on the impact of scaling and back-end-of-the-line (BEOL) technology solutions at the 7 nm node.

The appeal of MRAM lies in its nonvolatility, fast speeds, and high endurance. However, the team highlights the major hurdles hindering the adoption of MRAM for advanced technology nodes, particularly the increasing resistances of interconnects as devices scale down. Through their research, they have investigated the impact of shrinking interconnect dimensions on MRAM performance across various technology nodes.

Furthermore, the researchers have explored the potential impact of various back-end-of-the-line technology solutions at the 7 nm node. By conducting simulations using technology computer-aided design (TCAD) and experimentally validated/calibrated physical models, they have been able to quantify the potential array-level performance of MRAM using SPICE simulations.

Based on their findings, the team projects that certain BEOL technology solutions have the capability to significantly reduce the write and read energy of MRAM arrays. Specifically, they have observed that some potential solutions can reduce the write energy by up to 34.6% with spin-orbit torque (SOT) MRAM and 29.0% with spin-transfer torque (STT) MRAM. Additionally, there is a potential reduction of up to 21.4% in the read energy of the SOT-MRAM arrays. This represents a significant advancement in addressing the challenges related to scaling MRAM devices.

For those interested in delving deeper into the technical details of the study, the full paper is available in the IEEE Journal on Exploratory Solid-State Computational Devices and Circuits, authored by P. Kumar, D. E. Shim, S. Narla, and A. Naeemi.

This research is timely and holds great promise, especially given the increasing attention garnered by MRAM at the smallest nodes. With its potential to be the memory of choice for leading-edge designs and in automotive applications, the 25-year-old technology is becoming increasingly relevant. Moreover, as known issues are resolved, there is growing interest in Resistive Random-Access Memory (ReRAM) for embedded computing, particularly in automotive applications. However, it should be noted that there is no one-size-fits-all non-volatile memory (NVM) solution.

In conclusion, the pioneering work conducted by the researchers at the Georgia Institute of Technology sheds light on the potential of MRAM in advanced technology nodes and the strides being made to overcome the challenges associated with its scaling. As technology continues to advance, the study serves as a testament to the ongoing innovation and progress within the realm of memory technologies.