Superconductivity allows electric conductance with no energy losses when the ambient temperature drops below a critical value (T c ). Currently, the machine learning (ML)-based prediction of potential superconductors has been limited to chemical formulas without explicit treatment of material structures. Herein, we implement an efficient structural descriptor, the smooth overlap of atomic position (SOAP), into the ML models to predict the T c values with explicit atomic structural information. Using a data set containing 5713 compounds, our ML models with the SOAP descriptor achieved a 92.9% prediction accuracy of coefficient of determination (R 2 ) score via rigorous multialgorithm cross-verification procedures, exceeding the 86.3% accuracy record without atomic structure information. Several new high-temperature superconductors with T c values over 90 K were predicted using the SOAP-assisted ML model. This study provides insights into the structure−property relationship of high-temperature superconductors.

In this work, Ge2Sb2Te5 (GST) thin films are irradiated by a 1064 nm pulsed laser heat treatment system with different beam profiles. The surface effects induced by different laser conditions are studied systematically by atomic force microscope, spectroscopic ellipsometry, and Raman spectroscopy. It is found that a top-hat beam profile with uniform intensity distribution demonstrates the advantages of a non-destructive and homogeneous surface, which is critical for large-scale processing uniformity. The threshold laser fluence for the amorphization process is predicted by simulation and further proved by the laser irradiation experiment to be 27.9 mJ/cm2 at 1 ns pulse width. We further show that modulation of complex refractive indices of GST thin films can be achieved with different duty ratios (spatial ratio of amorphization part) from 0% to 100%. Our approach paves the way for the precise control of the optical properties of PCMs in emerging optical applications such as photonic switches, optical memories, and all-optical neural networks.

No abstract