Hydrogen-Assisted Asymmetric and Nonlinear Memristor Array for Reconfigurable Olfactory Graph Networks

Area of Science:

• Materials Science
• Neuroscience
• Electrical Engineering

Background:

• Memristor-based systems are promising for olfactory applications.
• Current limitations include lack of multifunctionality (sensing, memory, computation).

Purpose of the Study:

• Develop a multifunctional memristor array for olfactory sensing.
• Achieve selector-free operation with integrated sensing, memory, and computation.

Main Methods:

• Fabrication of Pt/HfO2 nanorods/TiN chemo-memristive devices in a crossbar array (CBA).
• Investigation of current-voltage (I-V) characteristics under hydrogen (H2) atmosphere.
• Utilizing oxygen vacancies (VO) and VO-H complexes for trap-assisted conduction.
• Demonstration of selector-free CBA operation and pattern recognition simulations.

Main Results:

• Achieved asymmetric I-V characteristics in response to H2 exposure.
• Enabled analog resistance tuning and precise conductance modulation.
• Demonstrated high inference accuracy (>94%) in digit pattern recognition.
• Validated selector-free operation of a 3x3 CBA for H2 sensing and computation.

Conclusions:

• The developed chemo-memristive CBA integrates selective sensing and in-memory computation.
• Adjustable I-V properties allow for reconfigurable olfactory networks.
• The device enables combined neuromorphic and electronic-olfaction functionality.