Breakthrough!!
IIT researchers develop neuromorphic hardware with magnetic materials
A team of researchers from the Indian Institutes of Technology (IIT) Delhi and Bombay built a new neuromorphic hardware using magnetic materials.
While interactive voice assistants like Alexa and Siri have made it very easy to obtain a response to a query or get some tasks done online, the procedure in the background is quite intricate. It involves delivering the speech data to the cloud over the internet, processing the data in the cloud, and then providing the response to the device.
But due to slow internet speed, power outages, and other factors, speech data is frequently lost. This may change with the advent of neuromorphic computing, where local hardware with built-in memory executes all operations, hence preventing data loss while simultaneously reducing power consumption.
The new neuromorphic hardware developed using magnetic materials is capable of storing data even when the power is off. It functions similar to a synapse in the neurological system, said the research team led by Prof. Pranaba Kishor Muduli at IIT Delhi and Prof. Debanjan Bhowmik at IIT Bombay.
This is the first experimental demonstration of a neuromorphic device using magnetic materials from India, according to the researchers. They detailed their device in the ACS Applied Electronic Materials journal.
"We have high hopes that the study contributes significantly to both the India Semiconductor mission as well as Make-in-India mission of the government of India," said Prof Pranaba Kishor Muduli, Department of Physics, IIT Delhi, in a statement.
The device consists of an ultrathin layer of cobalt. Thickness of the layer used is in the nano-metre range, which is 80,000-100,000 times smaller than human hair. The fabrication needs enormous precision; hence it is carried out in an ultra-high vacuum chamber containing nearly no air molecules, the team said.
For more details:
Demonstration of Synaptic Behavior in a Heavy-Metal-Ferromagnetic-Metal-Oxide-Heterostructure-Based Spintronic Device for On-Chip Learning in Crossbar-Array-Based Neural Networks, Ram S. Yadav, Pankhuri Gupta, Amod Holla, Kacho Imtiyaz Ali Khan, P. K. Muduli, and Debanjan Bhowmik, ACS Appl. Electron. Mater., 5, 484 (2023).
Indian Researchers demonstrates DMI interaction using Graphene
A team of Indian researchers from Indian Institute of Technology, Delhi and S. N. Bose National Centre for Basic Sciences, Kolkata—has found a new type of interaction, called Dzyaloshinskii–Moriya interactions (DMIs) when graphene is combined with the commonly used magnetic material permalloy ( 80 % Ni and 20 % Fe).
The DMI is the essential ingredient for formation of swirling patterns at the interface known as “Skyrmions”, which are considered potential way for for storing and erasing data on magnetic media as well as for logic and neuromorphic computing.
The team report a new and rather unusual type of DMI in graphene/Py system. They showed that the DMI in their system arises due to the defect induced extrinsic spin-orbit coupling at the graphene/Py interface.
Recently, Yang et al., [1] had also reported DMI at graphene/Ferromagnet interface. However, Yang et al. explained the iDMI observed in their system using Rashba effect. The iDMI observed by In Yang et al.’s work decays exponentially and vanishes within few monolayer of FM layer. In contrast, the Indian team demonstrated significantly large value of iDMI in rather thicker Py films (up to 8 nm).
The team also used simple and industry friendly sputter deposition for growth of Permalloy as opposed to molecular beam epitaxy used by Yang et al.
The Indian team believes that the discovery is very important for this field of spintronics and in particular for memory, logic and neuromorphic devices using skyrmions.
For more details:
Avinash Kumar Chaurasiya, Akash Kumar, Rahul Gupta, Sujeet Chaudhary, Pranaba Kishor Muduli, and Anjan Barman, Phys. Rev. B 99, 035402 (2019).
References
[1] Hongxin Yang, et al., Nat. Mat. 17, 605 (2018)