In situ tailoring of two-dimensional materials' phases under external stimulus facilitates the manipulation of their properties for electronic, quantum and energy applications. However, current methods are mainly limited to the transitions among phases with unchanged chemical stoichiometry. Here we propose on-device phase engineering that allows us to realize various lattice phases with distinct chemical stoichiometries. Using palladium and selenide as a model system, we show that a PdSe2 channel with prepatterned Pd electrodes can be transformed into Pd17Se15 and Pd4Se by thermally tailoring the chemical composition ratio of the channel. Different phase configurations can be obtained by precisely controlling the thickness and spacing of the electrodes. The device can be thus engineered to implement versatile functions in situ, such as exhibiting superconducting behaviour and achieving ultralow-contact resistance, as well as customizing the synthesis of electrocatalysts. The proposed on-device phase engineering approach exhibits a universal mechanism and can be expanded to 29 element combinations between a metal and chalcogen. Our work highlights on-device phase engineering as a promising research approach through which to exploit fundamental properties as well as their applications.
The prediction of crystal properties has always been limited by huge computational costs. In recent years, the rise of machine learning methods has gradually made it possible to study crystal properties on a large scale. We propose an attention mechanism-based crystal graph convolutional neural network, which builds a machine learning model by inputting crystallographic information files and target properties. In our research, the attention mechanism is introduced in the crystal graph convolutional neural network (CGCNN) to learn the local chemical environment, and node normalization is added to reduce the risk of overfitting. We collect structural information and calculation data of about 36 000 crystals and examine the prediction performance of the models for the formation energy, total energy, bandgap, and Fermi energy of crystals in our research. Compared with the CGCNN, it is found that the accuracy (ACCU) of the predicted properties can be further improved to varying degrees by the introduction of the attention mechanism. Moreover, the total magnetization and bandgap can be classified under the same neural network framework. The classification ACCU of wide bandgap semiconductor crystals with a bandgap threshold of 2.3 eV reaches 93.2%, and the classification ACCU of crystals with a total magnetization threshold of 0.5 µBreaches 88.8%. The work is helpful to realize large-scale prediction and classification of crystal properties, accelerating the discovery of new functional crystal materials.
