2022
DOI: 10.26434/chemrxiv-2022-3ptwx
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

A flexible and scalable scheme for mixing computed formation energies from different levels of theory

Abstract: Phase stability predictions are central to computational materials discovery efforts and have been made possible by large databases of computed properties from high-throughput density functional theory (DFT) calculations. Such databases now contain millions of calculations at the generalized gradient approximation (GGA) level of theory, representing an enormous investment of computational resources. Although it is now feasible to carry out large numbers of calculations using more accurate methods, such as meta… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
5
0

Year Published

2022
2022
2022
2022

Publication Types

Select...
1
1

Relationship

2
0

Authors

Journals

citations
Cited by 2 publications
(5 citation statements)
references
References 54 publications
0
5
0
Order By: Relevance
“…Figure a shows a selected Pourbaix region for ZnTiN 2 calculated using data from the Materials Project , (complete diagrams with multiple ionic concentrations are given in Figure S5). These diagrams are built from a combination of r 2 SCAN meta generalized gradient approximation (GGA) and PBE GGA DFT calculations using the computational Pourbaix formalism of Persson et al and the DFT mixing scheme of Kingsbury et al (see 4. Experimental Section).…”
Section: Resultsmentioning
confidence: 99%
See 3 more Smart Citations
“…Figure a shows a selected Pourbaix region for ZnTiN 2 calculated using data from the Materials Project , (complete diagrams with multiple ionic concentrations are given in Figure S5). These diagrams are built from a combination of r 2 SCAN meta generalized gradient approximation (GGA) and PBE GGA DFT calculations using the computational Pourbaix formalism of Persson et al and the DFT mixing scheme of Kingsbury et al (see 4. Experimental Section).…”
Section: Resultsmentioning
confidence: 99%
“…S3). These diagrams are built from a combination of r 2 SCAN metaGGA and PBE GGA DFT calculations using the computational Pourbaix formalism of Persson et al 61 and the DFT mixing scheme of Kingsbury et al 62 (see Methods). Including metaGGA calculations is beneficial because SCAN (on which r 2 SCAN is based) was shown to predict ternary nitride formation enthalpies in the nitrogen-rich region of the phase diagram more accurately than GGA.…”
Section: Electrochemical and Surface Propertiesmentioning
confidence: 99%
See 2 more Smart Citations
“…Our electrochemical stability calculations (Pourbaix diagrams) are built from a combination of PBE GGA DFT calculations retrieved from the Materials Project database and regulated-restored strongly-constrained and appropriately normed (r 2 SCAN) metaGGA calculations calculated using the workflow detailed in Kingsbury et al 90 We used the Materials Project DFT mixing scheme 62 to combine these two sets of calculations and create a solid phase diagram of the Zn-N-Ti-O-H chemical system, from which we constructed Pourbaix diagrams using the computational formalism of Persson et al 61 The mixing scheme allowed us to build the convex energy hull from higher-level metaGGA calculations by recomputing only the stable phases and phases close to the hull with r 2 SCAN (170 calculations total) instead of the entire Zn-N-Ti-O-H chemical system (more than 500 total phases according to the Materials Project Database). For two stable phases where the large number of sites made r 2 SCAN structure optimizations impractical (Ti3Zn22 and Ti20H2N17, with 100 and 39 sites, respectively) we employed single point calculations, as suggested in Kingsbury et al 62 The stability predictions of this mixed Pourbaix diagram were qualitatively similar to those obtained from a pure GGA phase diagram constructed without any r 2 SCAN calculations, but predicted a slightly larger region of decomposition to Ti3Zn2O8 and/or ZnO (see Figure S3).…”
Section: Calculationsmentioning
confidence: 99%