A dataset of 175k stable and metastable materials calculated with the PBEsol and SCAN functionals

Abstract

In the past decade we have witnessed the appearance of large databases of calculated material properties. These are most often obtained with the Perdew-Burke-Ernzerhof (PBE) functional of density-functional theory, a well established and reliable technique that is by now the standard in materials science. However, there have been recent theoretical developments that allow for increased accuracy in the calculations. Here, we present a dataset of calculations for 175k crystalline materials obtained with two functionals: geometry optimizations are performed with PBE for solids (PBEsol) that yields consistently better geometries than the PBE functional, and energies are obtained from PBEsol and from SCAN single-point calculations at the PBEsol geometry. Our results provide an accurate overview of the landscape of stable (and nearly stable) materials, and as such can be used for reliable predictions of novel compounds. They can also be used for training machine learning models, or even for the comparison and benchmark of PBE, PBEsol, and SCAN.

Transfer learning on large datasets for the accurate prediction of material properties

“…We started our transfer learning experiments by training crystal graph-attention neural networks 35 on a PBE dataset with 1.8 M structures 18 from the DCGAT database, and on the extended PBEsol and SCAN datasets from ref. 24. The DCGAT dataset combines compatible data from AFLOW, 14 the materials project 17 and ref.…”

Transfer learning on large datasets for the accurate prediction of material properties

“…We started our transfer learning experiments by training crystal graph-attention neural networks [31] on a PBE dataset with 1.8M structures [18] from the DC-GAT database, and on the extended PBEsol and SCAN datasets from Ref. [24]. The DCGAT dataset combines compatible data from AFLOW [14], the materials project [17] and Refs.…”

“…In Ref. [24] we reoptimized the geometries of 175k materials using PBEsol followed by a final energy evaluation with the PBEsol and SCAN functional as described in [24]. By now we extended these datasets by another 50k randomly selected materials arriving at 225k entries.…”

“…An exception is the smaller JARVIS database that encompasses ∼ 55k calculations obtained using OptB88-vdW [19,20] and the modified Becke-Johnson potential [21][22][23]. The first large databases beyond the PBE functional have been published only recently [24,25]. These datasets are based on the PBE functional for solids (PBEsol) [26], the highly constrained and appropriately normalized semilocal functional (SCAN) [27], and the R2SCAN functional [28].…”

Calibrating DFT Formation Enthalpy Calculations by Multifidelity Machine Learning

“…The methodology of this work can also be applied to the updated Materials Project database (such as V2021.05.13) and other large DFT databases. It is expected that, with more accurate low-fidelity data (DFT formation enthalpy), such as the very recent datasets by Kingsbury et al 44 and by Schmidt et al 101 with thousands of materials calculated by meta-GGA functionals, the method in this work can be used to provide more accurate calibration (exp. formation enthalpy).…”