POS5-0263
Single-Molecule-Thick Organic Junctions for Ultra-Low-Power Synaptic Devices
When and Where
Nov 30, -0001
00:00 - 00:00
Presenter(s)
Junwoo Park (Sogang University)
Co-Author(s)
Abstract
Organic and molecular electronics are increasingly moving beyond static device metrics toward adaptive, brain-inspired function. Self-assembled monolayers (SAMs) represent the ultimate thickness limit of an organic active layer—a single molecule thick—in which the operative charge-transport (CT) mechanism, set by molecular structure and applied bias, directly governs the electronic response. We exploit this principle to build energy-efficient molecular synapses and to clarify the transport physics that makes them work.
The device is a molecular junction formed from alkanethiolate SAMs terminated with 2,2′-bipyridine complexed with cobalt chloride (BIPY–CoCl₂). Driving the junction into the incoherent CT regime injects charge that lowers the barrier for C–C bond rotation, producing reversible, electric-field-driven conformational changes within the monolayer. These conformational dynamics translate into analog conductance potentiation and depression, emulating synaptic plasticity at an energy cost of only 8.0 pJ µm⁻². Implemented as artificial synapses, the device reaches 90% accuracy on MNIST handwritten-digit recognition—demonstrating that monolayer-thin, molecular-scale switching is a viable, low-power route to in-memory neuromorphic computing.
Central to this behavior is which transport regime is engaged. Beyond incoherent hopping, resonant coupling of carriers to specific intramolecular vibrational modes shapes how injected energy is dissipated and how switching proceeds; bias-dependent spectroscopic probing of this non-equilibrium vibrational excitation resolves these pathways directly. This mechanistic picture turns molecular-synapse design from empirical tuning into regime-targeted engineering, and positions molecular monolayers as a scalable, structurally tunable building block for organic neuromorphic and flexible electronics.
The device is a molecular junction formed from alkanethiolate SAMs terminated with 2,2′-bipyridine complexed with cobalt chloride (BIPY–CoCl₂). Driving the junction into the incoherent CT regime injects charge that lowers the barrier for C–C bond rotation, producing reversible, electric-field-driven conformational changes within the monolayer. These conformational dynamics translate into analog conductance potentiation and depression, emulating synaptic plasticity at an energy cost of only 8.0 pJ µm⁻². Implemented as artificial synapses, the device reaches 90% accuracy on MNIST handwritten-digit recognition—demonstrating that monolayer-thin, molecular-scale switching is a viable, low-power route to in-memory neuromorphic computing.
Central to this behavior is which transport regime is engaged. Beyond incoherent hopping, resonant coupling of carriers to specific intramolecular vibrational modes shapes how injected energy is dissipated and how switching proceeds; bias-dependent spectroscopic probing of this non-equilibrium vibrational excitation resolves these pathways directly. This mechanistic picture turns molecular-synapse design from empirical tuning into regime-targeted engineering, and positions molecular monolayers as a scalable, structurally tunable building block for organic neuromorphic and flexible electronics.
