Self-Rectifying MoS₂ Memtransistor via Asymmetry Contact Metal Engineering for Neuromorphic Computing

Area of Science:

• Materials Science
• Electronics Engineering
• Computer Engineering

Background:

• Memtransistors combine memristor and transistor functionalities for advanced computing.
• Scalability in memtransistor crossbar arrays is hindered by sneak leakage currents.

Purpose of the Study:

• To develop a self-rectifying memtransistor to overcome scalability limitations.
• To enhance performance metrics for neuromorphic and in-memory computing applications.

Main Methods:

• Fabrication of Molybdenum Disulfide (MoS2) memtransistors with asymmetric Platinum (Pt) and Bismuth (Bi) contacts.
• Characterization of electrical properties, including rectification and switching ratios.
• Validation through temperature-dependent studies, energy band analysis, and TCAD simulations.
• Evaluation in an Echo State Network (ESN) for neuromorphic computing tasks.

Main Results:

• Achieved a high rectification ratio of 10^4 and switching ratio of 10^4.
• Demonstrated a retention time exceeding 10^5 seconds.
• Reduced crossbar array leakage currents by five orders of magnitude.
• Enhanced power efficiency by 61 times compared to non-rectifying devices.

Conclusions:

• Asymmetric metal contacts enable self-rectifying MoS2 memtransistors.
• The developed memtransistors offer a scalable and energy-efficient solution for neuromorphic computing.
• The devices show robustness in low-precision and noisy environments, crucial for practical applications.