Reliable Memristive Synapses Based on Parylene-MoO Nanocomposites for Neuromorphic Applications.

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Memristive devices exhibit nonvolatile resistive switching, crucial for neuromorphic computing.
• Artificial synapses require analogue tunability and low energy consumption, properties inherent in memristors.
• Improving memristor performance and reliability is essential for advancing brain-inspired computing.

Purpose of the Study:

• To develop and characterize novel memristive devices using a parylene-molybdenum oxide (PPX-Mo) nanocomposite.
• To evaluate the performance enhancements of PPX-Mo memristors compared to traditional parylene-based devices.
• To demonstrate the potential of these devices as functional artificial synapses in neuromorphic systems.

Main Methods:

• Fabrication and comprehensive characterization of PPX-Mo nanocomposite memristive devices.
• Statistical analysis to identify optimal synthesis parameters for device performance.
• Integration of memristors into a neuromorphic computing model and assessment of scalability in crossbar arrays.

Main Results:

• PPX-Mo memristors demonstrated lower switching voltage and energy consumption compared to parylene-based devices.
• Devices exhibited reduced data dispersion and enhanced resistive plasticity.
• Optimal synthesis parameters were identified through robust statistical analysis.
• The most probable resistive switching mechanism was elucidated.

Conclusions:

• The developed PPX-Mo memristive devices offer superior characteristics for artificial synapses.
• These memristors show promise for scalable, low-power neuromorphic computing applications.
• The findings contribute to the advancement of hardware-based brain-inspired computational systems.