ScholarLens Back to search
2023•Unpublished venueRequires access

Fabrication of Domain Wall based Magnetic Tunnel Junction Devices with Intrinsic Neuromorphic Functionality

Thomas M. Leonard, Samuel Liu, Harrison Jin, Joseph S. Friedman, Christopher H. Bennett, Jean Anne C. Incorvia

Open publisher page 1 citations

Abstract

We summarize our progress towards monolithic and analog neuromorphic computing utilizing domain wall-magnetic tunnel junction (DW-MTJ) devices. We have previously shown device performance for binary logic DW-MTJ devices. Here, we expand on that work by demonstrating neuromorphic functionality using shape-dependent tunability. We measure multi-weight synapses and stochastic neurons monolithically fabricated from the same material stack, enabling future integrated neuromorphic circuits. Future work includes fabrication of leaky integrate and fire (LIF) neurons to complete the library of neuromorphic functionality for a full DW-MTJ crossbar array capable of neuromorphic computing.

About this research paper

What this paper is about

We summarize our progress towards monolithic and analog neuromorphic computing utilizing domain wall-magnetic tunnel junction (DW-MTJ) devices. We have previously shown device performance for binary logic DW-MTJ devices. Here, we expand on that work by demonstrating neuromorphic functionality using shape-dependent tunability. We measure multi-weight synapses and stochastic neurons monolithically fabricated from the same material stack, enabling future integrated neuromorphic circuits. Future work includes fabrication of leaky integrate and fire (LIF) neurons to complete the library of neuromorphic functionality for a full DW-MTJ crossbar array capable of neuromorphic computing.

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

We summarize our progress towards monolithic and analog neuromorphic computing utilizing domain wall-magnetic tunnel junction (DW-MTJ) devices. We have previously shown device performance for binary logic DW-MTJ devices. Here, we expand on that work by demonstrating neuromorphic functionality using shape-dependent tunability. We measure multi-weight synapses and stochastic neurons monolithically fabricated from the same material stack, enabling future integrated neuromorphic circuits. Future work includes fabrication of leaky integrate and fire (LIF) neurons to complete the library of neuromorphic functionality for a full DW-MTJ crossbar array capable of neuromorphic computing.

Key concepts: Neuromorphic engineering, Fabrication, Tunnel magnetoresistance, Crossbar switch, Domain (mathematical analysis), Computer science, Memristor, Stack (abstract data type)

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
Fabrication of Domain Wall based Magnetic Tunnel Junction Devices with Intrinsic Neuromorphic Functionality — Research Paper | ScholarLens