Ke Chen1,Hang Hu1,Inho Song1,Ashkan Abtahi1,Jianguo Mei1
Purdue University1
Ke Chen1,Hang Hu1,Inho Song1,Ashkan Abtahi1,Jianguo Mei1
Purdue University1
Retina is an important component in the human eye that help human beings get around 80% of knowledge from surroundings. As a type of bio-inspired optoelectronics, artificial retinas have been developed to mimic the functionalities of the human visual system. They are designed to function as bionic eyes for visually impaired persons or offer vision for advanced humanoid robots. To date, most of artificial retinas rely on silicon electronics, which are often rigid and brittle. This makes non-invasive implantation and seamless interfacing to soft biological tissues challenging. On the other hand, organic semiconducting materials have demonstrated biocompatibility, mechanical comfortability, and responsivity to the analytes in biological media, making them a promising candidate for next-generation artificial retina. Here, we report an organic optoelectronic device as an artificial retina. In the device, light can manipulate ion insertion into the bulk photoactive layer, enabling the modulation of multilevel nonvolatile conductance states at low operating voltages (~1 V) and the imitation of ion flux-regulated synaptic activity inherent in living systems. The nonvolatile properties are regulated by both light intensity and wavelength across the entire visible spectrum, demonstrating the capabilities of perceiving and memorizing various visual information of the device. Taking advantage of integrated function of light perception, processing and memorization of the device, we designed and simulated a single-layer synapse array as artificial retina that enables face recognition and a hardware convolution neural network for large-scale object recognition.