Early processing of visual information takes place in the human retina. Mimicking neurobiological structures and functionalities of the retina provides a promising pathway to achieving vision sensor with highly efficient image processing. Here, we demonstrate a prototype vision sensor that operates via the gate-tunable positive and negative photoresponses of the van der Waals (vdW) vertical heterostructures. The sensor emulates not only the neurobiological functionalities of bipolar cells and photoreceptors but also the unique connectivity between bipolar cells and photoreceptors. By tuning gate voltage for each pixel, we achieve reconfigurable vision sensor for simultaneous image sensing and processing. Furthermore, our prototype vision sensor itself can be trained to classify the input images by updating the gate voltages applied individually to each pixel in the sensor. Our work indicates that vdW vertical heterostructures offer a promising platform for the development of neural network vision sensor.

Inorganic CsPbX (X = Cl, Br, I, or hybrid among them) perovskite quantum dots (IPQDs) are promising building blocks for exploring high performance optoelectronic applications. In this work, the authors report a new hybrid structure that marries CsPbX IPQDs to silicon nanowires (SiNWs) radial junction structures to achieve ultrafast and highly sensitive ultraviolet (UV) detection in solar-blind spectrum. A compact and uniform deployment of CsPbX IPQDs upon the sidewall of low-reflective 3D radial junctions enables a strong light field excitation and efficient down-conversion of the ultraviolet incidences, which are directly tailored into emission bands optimized for a rapid photodetection in surrounding ultrathin radial p-i-n junctions. A fast solar-blind UV detection has been demonstrated in this hybrid IPQD-NW detectors, with rise/fall response time scales of 0.48/1.03 ms and a high responsivity of 54 mA W @200 nm (or 32 mA W @270 nm), without the need of any external power supply. These results pave the way toward large area manufacturing of high performance Si-based perovskite UV detectors in a scalable and low-cost procedure.

The unipolar resistive switches are investigated in silicon highly rich SiOx (x < 0.75) films. The as-deposited SiO0.73 films contain high concentration (1.0 × 1019 cm−3) of silicon dangling bonds (Si-DBs) and are rich in SiO2≡Si–Si and O3≡Si–Si configurations. Unlike the currently reported normal silicon-rich SiOx (x > 1.8) based devices, our Pt/SiO0.73/Pt devices operate at lower voltage regime (<2.0 V) and exhibit much lower resistance (∼30 Ω). The reset voltage (∼0.7 V) is lower than set voltage (∼1.7 V) and the performance is reduced in the vacuum environment. We propose a Si-DBs percolation model to explain the above characteristics. The experimental evidences for supporting our model are presented and discussed.

The realization of ultra-low power Si-based resistive switching memory technology will be a milestone in the development of next generation non-volatile memory. Here we show that a high performance and ultra-low power resistive random access memory (RRAM) based on an Al/a-SiNx:H/p+-Si structure can be achieved by tuning the Si dangling bond conduction paths. We reveal the intrinsic relationship between the Si dangling bonds and the N/Si ratio x for the a-SiNx:H films, which ensures that the programming current can be reduced to less than 1 μA by increasing the value of x. Theoretically calculated current-voltage (I–V ) curves combined with the temperature dependence of the I–V characteristics confirm that, for the low-resistance state (LRS), the Si dangling bond conduction paths obey the trap-assisted tunneling model. In the high-resistance state (HRS), conduction is dominated by either hopping or Poole–Frenkel (P–F) processes. Our introduction of hydrogen in the a-SiNx:H layer provides a new way to control the Si dangling bond conduction paths, and thus opens up a research field for ultra-low power Si-based RRAM.