Human eyes undertake the majority of information assimilation for learning and memory. Transduction of the color and intensity of the incident light into neural signals is the main process for visual perception. Besides light‐sensitive elements that function as rods and cones, artificial retinal systems require neuromorphic devices to transform light stimuli into post‐synaptic signals. In terms of plasticity timescale, synapses with short‐term plasticity (STP) and long‐term potentiation (LTP) represent the neural foundation for experience acquisition and memory formation. Currently, electrochemical transistors are being researched as STP–LTP devices. However, their LTP timescale is confined to a second‐to‐minute level to give unreliable non‐volatile memory. This issue limits multiple‐plasticity synapses with tunable temporal characteristics and efficient sensory‐memory systems. Herein, a ferroelectric/electrochemical modulated organic synapse is proposed, attaining three prototypes of plasticity: STP/LTP by electrochemical doping/de‐doping and ferroelectric‐LTP from dipole switching. The device supplements conventional electrochemical transistors with 10000‐second‐persistent non‐volatile plasticity and unique threshold switching properties. As a proof‐of‐concept for an artificial visual‐perception system, an ultraflexible, light‐triggered organic neuromorphic device (LOND) is constructed by this synapse. The LOND transduces incident light signals with different frequency, intensity, and wavelength into synaptic signals, both volatile and non‐volatile.

Human eyes use retina photoreceptor cells to absorb and distinguish photons from different wavelengths to construct an image. Mimicry of such a process and extension of its spectral response into the near-infrared (NIR) is indispensable for night surveillance, retinal prosthetics, and medical imaging applications. Currently, NIR organic photosensors demand optical filters to reduce visible interference, thus making filter-free and anti-visible NIR imaging a challenging task. To solve this limitation, a filter-free and conformal, retina-inspired NIR organic photosensor is presented. Featuring an integration of photosensing and floating-gate memory modules, the device possesses an acute color distinguishing capability. In general, the retina-like photosensor transduces NIR (850 nm) into nonvolatile memory and acts as a dynamic photoswitch under green light (550 nm). In doing this, a filter-free but color-distinguishing photosensor is demonstrated that selectively converts NIR optical signals into nonvolatile memory.

No abstract