Lian Zhentao, Wei Jianyong, Liu Yuzhuo, Liu Zuheng, Liu Yumeng, Xie Maosong, Dan Yaping, Tu Chang-Ching, Yang Rui
University of Michigan - Shanghai Jiao Tong University Joint Institute, Shanghai Jiao Tong University, Shanghai 200240, China.
Hon Hai Research Institute, Foxconn Technology Group, Shenzhen 518109, China.
ACS Nano. 2025 Jul 22;19(28):26041-26054. doi: 10.1021/acsnano.5c06692. Epub 2025 Jul 11.
Two-dimensional (2D) semiconductors have been of great interest for phototransistors and neuromorphic devices in recent years because of their unique optical and electronic properties. However, the detectable spectral range and light absorption efficiency are limited for 2D-semiconductor-based phototransistors. Herein, we report a high-performance deep-ultraviolet (DUV) sensitive phototransistor by integrating molybdenum disulfide (MoS) with silicon carbide nanoparticles (SiC NPs) to form a van der Waals heterostructure (vdWH), which shows ultrahigh responsivity and detectivity, especially in the DUV spectral range. The SiC NPs/few-layer MoS vdWH phototransistor shows a 20-fold enhancement in responsivity (from 9.4 × 10 to 1.9 × 10 A/W) and 11-fold enhancement in detectivity (from 7.9 × 10 to 8.4 × 10 cm × Hz/W) at 254 nm wavelength, compared to the phototransistor based on few-layer MoS alone. Moreover, the SiC NPs/few-layer MoS vdWH phototransistor also shows higher excitation postsynaptic current (EPSC) and longer retention time of postsynaptic current (PSC) compared to the phototransistor based on few-layer MoS alone. This enables vdWH devices to successfully mimic various biological synaptic functions, including paired-pulse facilitation (PPF), spike-duration-dependent plasticity, spike-number-dependent plasticity, spike-frequency-dependent plasticity, the transition from short-term plasticity (STP) to long-term plasticity (LTP), and long-term depression (LTD) capabilities. The simulation of a deep neural network (DNN) shows that the image inference accuracy based on these SiC NPs/few-layer MoS vdWH neuromorphic phototransistors reaches up to 98.99% even after considering the photoresponsivity variations. The high-performance dual-function neuromorphic optoelectronics based on SiC NPs/MoS vdWH hold great promise for ultrasensitive DUV photodetection, neuromorphic DUV visual sensing, and in-sensor computing applications in a single device.
近年来，二维（2D）半导体因其独特的光学和电子特性，在光电晶体管和神经形态器件领域备受关注。然而，基于二维半导体的光电晶体管的可检测光谱范围和光吸收效率有限。在此，我们报道了一种高性能的深紫外（DUV）敏感光电晶体管，它通过将二硫化钼（MoS）与碳化硅纳米颗粒（SiC NPs）集成形成范德华异质结构（vdWH），该结构表现出超高的响应度和探测率，特别是在深紫外光谱范围内。与仅基于几层MoS的光电晶体管相比，SiC NPs/几层MoS vdWH光电晶体管在254 nm波长下的响应度提高了20倍（从9.4×10提升至1.9×10 A/W），探测率提高了11倍（从7.9×10提升至8.4×10 cm×Hz/W）。此外，与仅基于几层MoS的光电晶体管相比，SiC NPs/几层MoS vdWH光电晶体管还表现出更高的激发突触后电流（EPSC）和更长的突触后电流（PSC）保留时间。这使得vdWH器件能够成功模拟各种生物突触功能，包括双脉冲易化（PPF）、脉冲持续时间依赖性可塑性、脉冲数量依赖性可塑性、脉冲频率依赖性可塑性、从短期可塑性（STP）到长期可塑性（LTP）的转变以及长期抑制（LTD）能力。深度神经网络（DNN）的模拟表明，即使考虑光响应度变化，基于这些SiC NPs/几层MoS vdWH神经形态光电晶体管的图像推理准确率仍高达98.99%。基于SiC NPs/MoS vdWH的高性能双功能神经形态光电器件在超灵敏深紫外光电探测、神经形态深紫外视觉传感以及单器件内的传感器计算应用方面具有巨大潜力。
