Sampling the Materials Space for Conventional Superconducting Compounds

Abstract

It performs a large scale study of conventional superconducting materials using a machine‐learning accelerated high‐throughput workflow. It starts by creating a comprehensive dataset of around 7000 electron–phonon calculations performed with reasonable convergence parameters. This dataset is then used to train a robust machine learning model capable of predicting the electron–phonon and superconducting properties based on structural, compositional, and electronic ground‐state properties. Using this machine, it evaluates the transition temperature (Tc) of approximately 200 000 metallic compounds, all of which are on the convex hull of thermodynamic stability (or close to it) to maximize the probability of synthesizability. Compounds predicted to have Tc values exceeding 5 K are further validated using density‐functional perturbation theory. As a result, it identifies 541 compounds with Tc values surpassing 10 K, encompassing a variety of crystal structures and chemical compositions. This work is complemented with a detailed examination of several interesting materials, including nitrides, hydrides, and intermetallic compounds. Particularly noteworthy is LiMoN2, which it predicts to be superconducting in the stoichiometric trigonal phase, with a Tc exceeding 38 K. LiMoN2 is previously synthesized in this phase, further heightening its potential for practical applications.

Explorations of boron-based materials through theoretical simulations

“…DG Pettifor designed phenomenological structure maps to elucidate the importance of atomic size, electron per atom ratio, and s, p, d bonding characteristics of the valence orbitals for binary A m B n compounds [15] . In addition, numerous HT searches have been applied to certain structures in hopes of discovering novel materials [16][17][18][19][20] . So, what is crucial in the design of materials with high hardness?…”

High-Throughput Screening for Boride Superconductors

“…In recent years, it has been shown that computational prediction and design can greatly facilitate the discovery of new superconducting materials. , The computational algorithms based on the density-functional perturbation theory (DFPT) can provide a satisfactory description of the electron–phonon coupling (EPC) and T c for conventional superconductors. Due to the significant cost of the DFPT calculations, T c calculations are now combined with information technologies such as data mining, machine learning, and high-throughput screening to guide the theoretical search of conventional superconductors. − …”

Data-driven design of high pressure hydride superconductors using DFT and deep learning

“…With the advancements in computing power, DFT databases, and robust and efficient computational workflows, high-throughput calculations (on the order of hundreds to thousands) of superconducting properties are now in the realm of possibility [28][29][30][31][32][33][34][35]. In addition, various machine learning and data-driven approaches have been used to aid in the discovery of new and novel superconductors [36][37][38][39][40][41][42][43][44][45][46][47][48]. In our previous works, we expanded the existing JARVIS (Joint Automated Repository for Various Integrated Simulations) [33,35] DFT database to include calculations for over 1000 bulk [36] and 150 two-dimensional (2D) superconductors [37] (all at zero pressure).…”

“…In addition to our own work, there has been a recent highthroughput DFT and machine learning study which performed EPC calculations for over 7000 superconducting candidates [48] (all at zero pressure), which makes it one of the largest high-throughput DFT study of conventional superconductors to date. In contrast to the DFT and machine learning works that span several material classes, there have been few highthroughput computational efforts that have focused on high pressure hydrides [46].…”

“…After running these DFT calculations, we trained a deep learning model (using the relaxed high-pressure structures and computed T c values) to predict T c given a crystal structure, specifically tailored to high pressure hydrides. Such deep learning models have shown remarkable potential for superconductor design [36,45,48,[51][52][53]. After testing and benchmarking our newly trained deep learning property prediction model, we used a unified machine learning-based force-field to relax the hydride structures under pressure, further reducing the number of DFT calculations needed to obtain a complete map of superconducting properties under pressure.…”