Photoelectric sensing, a photoelectric sensing device that converts optical signals of different wavelength bands into electrical signals, plays an irreplaceable role in image sensing, optical communication, environmental monitoring, and biological monitoring, and is crucial to people's lives and national defense construction. In recent years, Huang Hui's research group from the University of Chinese Academy of Sciences has developed a series of molecular design methods and device preparation strategies in this field. . Funct. Mater. 2016, 26, 6306–6315); polymers prepared by random copolymerization method achieved responsivity up to 19.1 A/W (ACS Appl. Mater. Interfaces 2018, 10, 2, 1917–1924) ; Molecular engineering through conformational adjustment and side chain modification increased the responsivity by 3 times (J. Mater. Chem. C 2019, 7, 5739--5747); Expanded the photoelectric response of organic matter to the short-wave infrared region, the preparation is flexible The device was used for 16 × 16 array imaging (Macromolecules 2020, 53, 10636-10643). Recently, self-driven optoelectronic synaptic devices on flexible substrates have been further realized and applied in array memory imaging and anti-counterfeiting models.
The traditional von Neumann computing system will face a series of problems due to the physical separation of the memory module and the processor, such as extra energy consumption during data transfer, limited computing speed, and unstructured real-time information processing. Neuromorphic computing is regarded as one of the most promising approaches to solve the von Neumann bottleneck due to its advantages of adaptive learning, high parallel computing, and low power consumption. An important prerequisite for realizing neuromorphic computing is to develop synaptic devices that can simulate the behavior of biological synapses. Electrical synaptic devices are the first developed synaptic devices, but they face great challenges in overall optimization under the consideration of bandwidth, connection, density and other factors. In recent years, artificial synapses that use light as a stimulus response have gradually developed. Compared with electrical synapses, light has the characteristics of high bandwidth, low crosstalk, low energy consumption and no delay, and can also directly simulate the vital functions such as vision. neurobehavior.
First of all, the energy consumption problem is the challenge faced by synaptic devices. The primary synapse behavior of the human body is tens to hundreds of femtojoules, and most of the current photoelectric synapse devices are three-terminal phototransistor structures, which often need to serve under higher driving voltages. Such a large energy consumption is microscopic. In the field of electronics, it is necessary to avoid even high voltage, which will inevitably lead to the accumulation of heat, which will affect the performance of the device. Therefore, many people have made efforts to reduce the energy consumption of synaptic devices. Secondly, most of the way humans obtain external information comes from vision. In order to simulate human vision, many optoelectronic synaptic devices have shown excellent performance for visible light stimulation. However, in the field of artificial intelligence, researchers hope to widen the range of visual response. , which can perceive and calculate more infrared light and ultraviolet light that cannot be seen by humans. Especially infrared light has a wide range of applications in the fields of military, remote control and optical communication.
Figure 1. Device structure and photosynthetic behavior simulation
在该工作中研究者采用一种短波红外响应的聚合物PBTT，将其旋涂在Si/SiO2基底和柔性聚对苯二甲酸乙二酯PET基底上，实现了一系列光电突触行为的模拟，例如兴奋性突触电流/后电位（EPSC），短程可塑性（STP）、长程可塑性（LTP），STP-LTP转化，双脉冲易化（PPF）等等。特别的是，在PET基底上即使不施加外加偏压，器件也能够运行，这就意味着我们可以只通过短波红外光刺激对器件性能进行调制而无需电压，大大的减少了能量消耗。
Figure 2. Exploring the self-driving mechanism of the device
In order to further study the self-driving mechanism of the flexible substrate, the researchers studied the surface morphology and charge distribution of the thin film. First, the roughness of the polymer on the PET substrate is much higher than that on the Si/SiO2 substrate, which leads to a larger localized electric field on the PET surface. Compared with the surface charge distribution of the Si/SiO2 substrate film, which has almost no change in the dark state and the light, the PET substrate shows a more obvious difference in the charge distribution under the light, which means that there is a photo-induced charge accumulation phenomenon on the PET substrate. This facilitates the dissociation of excitons under the local electric field. In order to further quantify this phenomenon, Kelvin force microscopy was used to measure the surface potential of the film under different light intensities. 105 was enhanced to 244, 313, 481 mV, which is consistent with the light intensity-dependent photoresponse current values. At the same time, the relative permittivity of PBTT reaches 5.8. This higher permittivity value reduces the exciton binding energy, so that the excitons can dissociate into carriers without the need for an external electric field. As a comparative experiment, another polymer PBTB with short-wave infrared light response cannot work at 0 V with an external bias voltage, and can only exhibit photoelectric synaptic performance when there is a driving voltage. Its dielectric constant is only 3.5, indicating that A large dielectric constant is beneficial to produce device self-driving properties.
Figure 3. Device array memory imaging and anti-counterfeiting model
Finally, the researchers applied the flexible self-driven synaptic device to array memory imaging and anti-counterfeiting models, and achieved good results, reflecting the advantages of simple device structure and no voltage energy loss. This achievement was recently published in Cell Reports Physical Science under the title "Self-powered Flexible Artificial Synapse for Near-infrared Light Detection". Chen Hao, a doctoral student at the University of Chinese Academy of Sciences, is the first author of the article, and Professor Huang Hui is the corresponding author.