Machine-Learning-Guided Prediction Models of Critical Temperature of Cuprates

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computational Chemistry

Background:

• The mechanism of superconductivity in cuprates remains a significant challenge in condensed matter physics.
• Predicting the critical temperatures (Tc) of cuprates is crucial for discovering new high-temperature superconductors.

Purpose of the Study:

• To predict the maximum superconducting transition temperature (Tc,max) of hole-doped cuprates using machine learning.
• To identify key material descriptors that influence Tc,max.
• To explore hypothetical cuprates for enhanced superconducting properties.

Main Methods:

• Utilizing machine learning algorithms combined with first-principles calculations.
• Developing a functional form to estimate Tc,max based on identified descriptors.
• Generating and evaluating hypothetical cuprate structures by substituting apical cations.

Main Results:

• A predictive model for Tc,max was established with a root-mean-square-error of 3.705 K and R² of 0.969.
• Bader charge of apical oxygen, apical atom bond strength, and the number of superconducting layers were identified as critical factors.
• Hypothetical cuprates incorporating Gallium (Ga) showed predicted Tc,max values up to 131 K.

Conclusions:

• Machine learning provides a powerful tool for predicting and guiding the design of novel superconductors.
• The identified descriptors offer insights into optimizing cuprate structures for higher critical temperatures.
• This approach accelerates the discovery of materials with potential for high-Tc superconductivity.