Ultralow power magnetoelectric nonvolatile memory devices for AI computing: A potential pathway to revolutionize the IoT technology
Implementing Organization
Indian Institute of Science
Principal Investigator
Dr. Bhagwati Prasad
Indian Institute of Science
CO-Principal Investigator
Prof. Rajeev Ranjan
Indian Institute of Science
Project Overview
The Internet of Things (IoT) is a rapidly growing field that uses artificial intelligence and machine learning algorithms to enhance various systems and processes. The global IoT market is worth around $400 billion, with a projected capital of $1.8 trillion by 2028. However, current computation architectures are inefficient for processing large amounts of data at high speeds. One potential solution is the use of high-density nonvolatile memory (NVM), such as magnetoresistive random-access memory (MRAM). MRAM technology offers a significant advantage over other emerging memory technologies, as it reduces switching energy down to 1 aJ range even at higher operating speeds. This proposal aims to demonstrate these promises in MRAM devices for the first time. Multiferroic oxide, BiFeO3 (BFO), will be explored as a magnetoelectric vehicle to manipulate MRAM resistance states with voltage. The project will investigate the nature of Dzyalozhinski-Moriya interaction with magnetic films and the voltage-control of magnetism for manipulating resistance states of magnetic tunnel junctions. Tools for oxide and metallic thin film depositions will be required, with oxide films deposited in a pulsed laser deposition tool and metallic films in an ultra-high vacuum sputtering system. This project will provide a solid pathway for AI-based exponential technology, starting with basic fundamental research to prototype device demonstration (TRL-3), and ultimately set potential directions for technology development.