MRS Meetings and Events

 

EL04.06.15 2023 MRS Fall Meeting

The Impact of Sc-Substitution on the Switching Behavior in AlN-Based Memristor Devices

When and Where

Nov 29, 2023
8:00pm - 10:00pm

Hynes, Level 1, Hall A

Presenter

Co-Author(s)

Seokgi Kim1,Jimyeong Yu1,Sungkyu Kim1

Sejong University1

Abstract

Seokgi Kim1,Jimyeong Yu1,Sungkyu Kim1

Sejong University1
Aluminum Nitride (AlN) has garnered significant interest as an active material for memristor devices due to its outstanding characteristics such as a tunable bandgap by doping, chemical/thermal stability, and high electrical resistivity (~ 10<sup>14</sup> Ω-m). [1],[2] Since memristor devices for artificial intelligence learning must ensure low power consumption, fast write speed, and long-term stability, it is important to clearly identify the switching behavior and control the electrical path in active materials. [3] However, the strong electric field is required to generate the set state, and permanent breakdown by Joule heating occurs due to the high barrier of AlN/metal Schottky junction. [4] Therefore, it is necessary to lower the Schottky barrier <i>via</i> elemental doping and form the robust conductive path by soft breakdown for realizing a stable AlN-based memristor with multilevel resistance.<br/>In this study, we fabricated the vertically stacked Au/Al<sub>0.7</sub>Sc<sub>0.3</sub>N/Cu device with a modified active layer in which 30% of Al sites is substituted with Sc to prevent permanent breakdown in Au/AlN/Cu device. The yield of Au/AlN/Cu is lower than Au/Al<sub>0.7</sub>Sc<sub>0.3</sub>N/Cu by 23% because of the excessive Joule heating induced by high forming voltage of AlN causes permanent breakdown. The irreversible migration of nitrogen ions in AlN active layer leads to the degradation of conductance. Compared to AlN-based device, Sc-substituted AlN (AlScN) with a 15% reduced bandgap exhibits a low forming voltage of 6.0 V due to Schottky barrier lowering. In addition, AlScN with a thickness of 30 nm shows stable switching behavior at a reduced set voltage of 2.1 V and a reset voltage of -1.3 V formed by relatively weak soft breakdown. We directly observed conductive path composed of nitrogen vacancies in a AlScN layer using atomic-resolution transmission electron microscopy. In addition, the resistance at low resistive state gradually decrease at the elevating temperature from 300 K to 400 K, indicating typical semiconductor behavior. Under various compliance current conditions ranging from 0.1 mA to 10 mA, all resistance states were maintained for a duration of 10<sup>5</sup> seconds at a read voltage of 0.3 V without noticeable degradation. The analog switching characteristics of AlScN and AlN devices were measured by applying a 10 us positive/negative voltage pulse for potentiation/depression, and statistical values of non-linearity factors (NL) were obtained for its quantification. Compared to the average NL of AlN devices is 0.67, the average NL of AlScN devices is 0.36, which is explained by the reversible movement of nitrogen ions in AlScN matrix. Modification of junction properties by elemental substitution provides a specific method to improve the reliability of AlN-based memristor devices for future neuromorphic computing.<br/><br/>[1] Ravi Prakash et al., Vacuum 143, 102-105 (2017)<br/>[2] Zhiping Zhang et al., IEEE ELECTRON devices LETTERS, VOL. 36, NO. 1, JANUARY (2015)<br/>[3] K. Moon et al., Faraday Discuss., 2019, 213, 421<br/>[4] Eszter Piros et al., Appl. Phys. Lett.117, 013504 (2020)

Keywords

electrical properties | nitride

Symposium Organizers

Simone Fabiano, Linkoping University
Paschalis Gkoupidenis, Max Planck Institute
Zeinab Jahed, University of California, San Diego
Francesca Santoro, Forschungszentrum Jülich/RWTH Aachen University

Symposium Support

Bronze
Kepler Computing

Session Chairs

Paschalis Gkoupidenis
Zeinab Jahed

In this Session

EL04.06.01
Visible Light-Driven IGZO Optoelectronic Synaptic Transistors with Subgap State Enhanced by Sonication

EL04.06.02
Bio-Interface for Actuation and Neuromorphic Devices

EL04.06.03
Enhancing RRAM Device Performance: A Design of Experiments Approach

EL04.06.05
Visible Light Stimulated Optoelectronic Synaptic Transistor via Solution Processed Vertically Diffused Cd Doped IGZO

EL04.06.06
Expanding Dynamic Range of Ionic Liquid Based Physical Reservoirs by Utilizing High Molecular Design Flexibility

EL04.06.07
Neuromorphic Applications Realized by a Free-Standing Multilayer Molybdenum Disulfide Memristor

EL04.06.08
Self-Rectifying and Artificial Synaptic Characteristics of Amorphous Ta2O5 Thin Film Bilayer Memristor

EL04.06.09
Improvement of Information Processing Performance in the Ionic Liquid-Based Physical Reservoir Device by Thermal and Electrical Pretreatment

EL04.06.11
Preparation and Characterization of Hf0.5Zr0.5O2-Based Flexible RRAM Device

EL04.06.12
Crystalline NaNbO3 Thin Films Grown on a Sr2Nb3O10 Seed Layer at Low Temperature for Self-Rectifying and Self-Powered ReRAM Devices

View More »

Publishing Alliance

MRS publishes with Springer Nature