Dual Light Temporal Coding Modes Enabled by Nanoparticle-Mediated Phototransistors via Gate Bias Modulation for Brain-Inspired Visual Perception

Abstract

The core integration and cooperation of the retina, neurons, and synapses in the visual systems enable humans to effectively sense and process visual information with low power consumption. To mimic the human visual system, an artificial sensory nerve, along with optical sensing�a paired-pulse ratio (PPR) of the light pulse stimulated currents�and neural coding has been developed. For performing the artificial visual perception functions, we consistently reveal the positive and negative correlations between the PPR index and light pulse time interval by applying two consecutive light stimuli with gate voltages of −10 and 5 V, respectively, to a phototransistor. This phototransistor contains a heterostructured channel layer composed of zinc-oxide nanoparticles (ZnO NPs) interconnected with a solution-processed zinc−tin oxide (ZTO) film. The oxygen adsorption and desorption on the ZnO NP surface under light illumination are responsible for the positive-sloped PPR; the electron trapping effect at the ZnO NP/SiO 2 interface is attributed to the negative-sloped PPR. The various accountable light power densities and number of surface trap states are considered to be directly realizing these spike-timing interval-dependent characteristics. The actual benefit of these characteristics is the dual temporal coding modes based on multiplicative operation using a ZTO/ZnO NP phototransistor realized via the active gate voltage modulation. The contrary tendency of the PPR index and temporal coding�a major biological neural coding�is well demonstrated by the potential of ZTO/ZnO NP phototransistors to be implemented in sensor networks for an artificial visual perception.