Two dimensional layered heterojunction is a new kind of material stacked layer by layer like a thousand layer cake. Due to its sensitive light response, signal plasticity and low power consumption, it is considered to be one of the most promising materials for constructing artificial neural morphological visual sensing (such as bionic eyes). Recently, the research group of Professor Liu Juqing and Associate Professor Li Yinxiang from the School of Flexible Electronics (Future Technology) of Nanjing University of Technology created a universal method to construct large two-dimensional carbon based heterostructures based on gas-liquid interface co self-assembly strategy, and obtained a series of homogeneous centimeter level heterostructure double-layer films. This work provides material and technical support for the construction of artificial neural morphological vision sensor. The related research paper was published in Nature Communication recently. "The so-called gas-liquid interface co self-assembly strategy is vividly understood as that two kinds of solutions are successively dropped on the water surface, and these two substances will self assemble into films at the two-dimensional interface of water and air - a small molecule is uniformly arranged into a two-dimensional film, and the two films form layered heterostructures through the van der Waals force between them at the same time. This strategy is universal, and the heterojunction area formed is large, up to cm in size; The heterojunction has high uniformity, and the quality of the whole film is close to each other and the thickness is uniform. " Li Yinxiang introduced that the van der Waals force is similar to an attraction, so that the two can be stacked together. "Generally speaking, it is precisely because of the superposition of light absorption ranges of each layer of material in the heterostructure that it has a wide range of light absorption. The oxygen-containing functional groups in the heterostructure make it have a defect state. The defect state will capture photogenerated electrons, and after the light is removed, the electrons will be slowly released. Because the energy bands of each material do not coincide with each other, there is a difference, and because the defect state captures and releases the charge, the heterostructure will pair Optical signals have memory plasticity. " Dong Xuemei, the co first author of the paper and a doctor from Nanjing University of Technology, explained that memory plasticity refers to giving different memory time, different levels of memory, and different times of forgetting. "The effect of seeing a picture once is different from that of watching it ten times, and the latter will remember more deeply and forget more slowly.". The team designed a brain like optoelectronic synaptic device based on this kind of heterojunction. Dong Xuemei introduced that, in contrast, the device designed by the team greatly simplifies the complexity of the device. Like the human visual system, it integrates the perception, storage and processing of optical signals, which reduces the time delay and energy consumption caused by information transmission between different devices. The relevant person in charge of the research group introduced that the energy consumption of the device is only 10-9 watts, which can sense the light stimulus in a wide range from ultraviolet light, visible light to near-infrared light, and can detect very weak light signals in dark environments. In addition to its strong light perception, the device, like the synaptic structure between two neurons in the human brain, can also remember and process the detected light signals at the same time. 214% double pulse facilitation index shows that the device has a very high memory capacity for the detected optical signals. Optical signals are used to regulate the synaptic weights of devices. Brain like optoelectronic synapses based on heterostructures can also realize neural morphological networks for visual learning and recognition,