Related Experiment Video
Updated: Jul 25, 2025

10:31
Plasma-assisted Molecular Beam Epitaxy of N-polar InAlN-barrier High-electron-mobility Transistors
Published on: November 24, 2016
8.6K
In-Grain Ferroelectric Switching in Sub-5 nm Thin Al0.74 Sc0.26 N Films at 1 V
Georg Schönweger1,2, Niklas Wolff3,4, Md Redwanul Islam3
1Department of Electrical and Information Engineering, Kiel University, Kaiserstrasse 2, D-24143, Kiel, Germany.
Advanced Science (Weinheim, Baden-Wurttemberg, Germany)
|June 29, 2023
Summary
Ferroelectric switching in sub-5 nm aluminum scandium nitride films achieves record low 1V switching voltages. This breakthrough in wurtzite-type ferroelectrics paves the way for highly scalable neuromorphic computing devices.
Area of Science:
- Materials Science
- Solid State Physics
- Nanotechnology
Background:
- Ferroelectric devices offer high energy efficiency for neuromorphic computing.
- Device scalability is a key challenge for realizing efficient neuromorphic applications.
- Wurtzite-type ferroelectrics are promising candidates for advanced electronic devices.
Purpose of the Study:
- To investigate the ferroelectric switching characteristics of sub-5 nm AlScN films.
- To overcome scalability limitations in ferroelectric devices for neuromorphic computing.
- To demonstrate the potential of AlScN for energy-efficient analog switching.
Main Methods:
- Sputter-deposition of sub-5 nm Al0.74Sc0.26N films on Pt/Ti/SiO2/Si and epitaxial Pt/GaN/sapphire templates.
- Characterization of ferroelectric switching voltages and coercive field.
- Atomic-scale imaging using scanning transmission electron microscopy (STEM).
Main Results:
- Achieved record low switching voltages down to 1V, compatible with on-chip sources.
- Observed a significantly larger coercive field to breakdown field ratio in Al0.74Sc0.26N films on silicon substrates.
- Demonstrated the formation of ferroelectric domains and inversion domain boundaries (IDBs) at the atomic scale in sub-5 nm films.
Conclusions:
- Sub-5 nm AlScN films exhibit promising ferroelectric properties for neuromorphic computing.
- The observed switching characteristics and domain structures support a gradual domain-wall driven switching mechanism.
- These findings enable analog switching in highly scaled ferroelectric devices for neuromorphic applications.

