2026 | 2025 | 2024 | 2023 | 2022 | 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 | 2011 | 2010 | 2009 | 2008

Machine Learning Screening of Superconducting Thin Films

Authors: Thalis Henrique Bispo da Silva

Supervisors: Tiago Cerqueira

MSc thesis, Master in Physics (2025)

Abstract: In this thesis, a large-scale computational investigation aimed at identifying two-dimensional (2D) superconducting materials is presented. The approach combines electron-phonon interaction calculations, performed using density-functional perturbation theory (DFPT), with machine learning models to enable efficient screening. Specifically, over 140,000 candidate materials from the Alexandria database were examined, employing a high-throughput methodology to assess their superconducting potential.The results reveal a wide variety of 2D superconductors with diverse chemical compositions and crystal structures. Interestingly, it is observed that 2D materials tend to exhibit stronger electron-phonon coupling compared to their three-dimensional (3D) counterparts. However, due to their generally lower average phonon frequencies, the resulting superconducting transition temperatures (Tc) tend to be modest. Despite this, several outlier materials with relatively high Tc values were identified.In total, 105 2D systems were found to exhibit Tc > 5 K. Among these, compounds such as CuH2, NbN, and V2NS2 emerged as particularly promising due to their favorable combination of high Tc, mechanical robustness, and excellent thermodynamic stability. These findings not only contribute to the ongoing global search for practical 2D superconductors but also strongly underscore the importance of combining large computational databases with advanced machine learning techniques to accelerate the discovery of novel and technologically relevant quantum materials.