Heterointerface engineering of layered double hydroxide/MAPbBr3 heterostructures enabling tunable synapse behaviors in a two-terminal optoelectronic device.

Solution-processable semiconductor heterostructures enable scalable fabrication of high performance electronic and optoelectronic devices with tunable functions via heterointerface control. In particular, artificial optical synapses require interface manipulation for nonlinear signal processing. However, the limited combinations of materials for heterostructure construction have restricted the tunability of synaptic behaviors with simple device configurations. Herein, MAPbBr3 nanocrystals were hybridized with MgAl layered double hydroxide (LDH) nanoplates through a room temperature self-assembly process. The formation of such heterostructures, which exhibited an epitaxial relationship, enabled effective hole transfer from MAPbBr3 to LDH, and greatly reduced the defect states in MAPbBr3. Importantly, the ion-conductive nature of LDH and its ability to form a charged surface layer even under low humidity conditions allowed it to attract and trap holes from MAPbBr3. This imparted tunable synaptic behaviors and short-term plasticity (STP) to long-term plasticity (LTP) transition to a two-terminal device based on the LDH-MAPbBr3 heterostructures. The further neuromorphic computing simulation under varying humidity conditions showcased their potential in learning and recognition tasks under ambient conditions. Our work presents a new type of epitaxial heterostructure comprising metal halide perovskites and layered ion-conductive materials, and provides a new way of realizing charge-trapping induced synaptic behaviors.

Deep Learning for Additive Screening in Perovskite Light-Emitting Diodes.

Additive engineering with organic molecules is of critical importance for achieving high-performance perovskite optoelectronic devices. However, experimentally finding suitable additives is costly and time consuming, while conventional machine learning (ML) is difficult to predict accurately due to the limited experimental data available in this relatively new field. Here, we demonstrate a deep learning method that can predict the effectiveness of additives in perovskite light-emitting diodes (PeLEDs) with a high accuracy up to 96 % by using a small dataset of 132 molecules. This model can maximize the information of the molecules and significantly mitigate the duplicated problem that usually happened with previous models in ML for molecular screening. Very high efficiency PeLEDs with a peak external quantum efficiency up to 22.7 % can be achieved by using the predicated additive. Our work opens a new avenue for further boosting the performance of perovskite optoelectronic devices.