Ultrawide Charge-Trap Memory Window and Photoinduced Synaptic Behavior in p-Channel Amorphous Oxide Semiconductors

Area of Science:

• Materials Science
• Semiconductor Physics
• Artificial Intelligence

Background:

• High-performance p-type amorphous oxide semiconductors (AOSs) are crucial for complementary circuits and neuromorphic computing.
• The lack of such materials has hindered progress in charge trapping phenomena research.

Purpose of the Study:

• To develop a novel p-channel amorphous oxide semiconductor.
• To demonstrate its potential for integrated neuromorphic computing applications.

Main Methods:

• Synthesized amorphous tellurium trioxide (a-TeO3) via ultraviolet-ozone oxidation of crystalline tellurium.
• Fabricated devices integrating transistor, memory, photodetection, and synaptic functions.
• Investigated charge trapping dynamics and light-induced effects.

Main Results:

• Achieved a p-channel AOS with an ultrawide memory window (>58 V) driven by oxygen vacancy-adsorbate interactions.
• Demonstrated light-programmable synaptic plasticity and associative learning via persistent photocurrent.
• Integrated with n-mode charge-trap memory for autonomous control and homeostasis.
• Achieved competitive MNIST accuracy in hardware-constrained neuromorphic networks.

Conclusions:

• The developed a-TeO3 platform overcomes critical materials limitations for neuromorphic computing.
• Enables scalable, energy-efficient neuromorphic architectures for edge AI.
• Paves the way for advanced functional integration in semiconductor devices.