Molecular Engineering of Interlayer Spacings in 2D Dion-Jacobson Perovskites for High-Fidelity Neuromorphic Computing

Area of Science:

• Materials Science
• Condensed Matter Physics
• Nanotechnology

Background:

• Memristive devices are crucial for energy-efficient neuromorphic computing.
• Two-dimensional (2D) organic-inorganic hybrid perovskites offer tunable structures and stability.
• The relationship between molecular-level interlayer architecture and resistive switching (RS) kinetics in these materials needs further investigation.

Purpose of the Study:

• To investigate the impact of inorganic layer spacing on the resistive switching behavior of Dion-Jacobson (DJ) phase 2D perovskites.
• To correlate molecular-level structural modifications with device performance for neuromorphic applications.

Main Methods:

• Synthesized DJ phase 2D perovskites (BDAPbI4, HDAPbI4, ODAPbI4) by varying diammonium ligand alkyl chain length.
• Investigated the effect of modulated interlayer spacing on ion migration and RS kinetics.
• Fabricated memristive devices and characterized their electrical performance, including ON/OFF ratio and cycling endurance.
• Simulated a three-layer neural network using device characteristics to assess performance on the MNIST dataset.

Main Results:

• Expanded interlayer spacing in HDAPbI4 and ODAPbI4 devices improved RS behavior.
• Achieved ON/OFF ratios exceeding 10^3 and over 100 cycles of endurance.
• Demonstrated emulation of synaptic functions, including the transition from short-term to long-term plasticity.
• Achieved 96.6% recognition accuracy in a three-layer neural network simulation for MNIST dataset.

Conclusions:

• Interlayer spacing is a critical structural parameter regulating ion migration and RS behavior in 2D DJ perovskite memristors.
• Optimized 2D perovskite memristors show potential for energy-efficient neuromorphic computing.
• The study highlights a pathway for designing advanced memristive devices with tunable analog conductance modulation.