scholarly journals Accuracy of electron densities obtained via Koopmans-compliant hybrid functionals

2018 ◽  
Vol 2 (4) ◽  
Author(s):  
A. R. Elmaslmane ◽  
J. Wetherell ◽  
M. J. P. Hodgson ◽  
K. P. McKenna ◽  
R. W. Godby
2018 ◽  
Vol 115 (10) ◽  
pp. 2287-2292 ◽  
Author(s):  
Neil Qiang Su ◽  
Zhenyu Zhu ◽  
Xin Xu

Recently, it was argued [Medvedev MG, et al. (2017) Science 355:49–52] that the development of density functional approximations (DFAs) is “straying from the path toward the exact functional.” The exact functional should yield both exact energy and density for a system of interest; nevertheless, they found that many heavily fitted functionals for molecular energies actually lead to poor electron densities of atoms. They also observed a trend that, for the nonempirical and few-parameter functionals, densities can be improved as one climbs up the first four rungs of the Jacob’s ladder of DFAs. The XYG3 type of doubly hybrid functionals (xDHs) represents a less-empirical and fewer-parameter functional on the top fifth rung, in which both the Hartree–Fock-like exchange and the second-order perturbative (MP2-like) correlation are hybridized with the low rung functionals. Here, we show that xDHs can well describe both density and energy for the same atomic set of Medvedev et al., showing that the latter trend can well be extended to the top fifth rung.


2019 ◽  
Author(s):  
Mark Iron ◽  
Trevor Janes

A new database of transition metal reaction barrier heights – MOBH35 – is presented. Benchmark energies (forward and reverse barriers and reaction energy) are calculated using DLPNO-CCSD(T) extrapolated to the complete basis set limit using a Weizmann1-like scheme. Using these benchmark energies, the performance of a wide selection of density functional theory (DFT) exchange–correlation functionals, including the latest from the Truhlar and Head-Gordon groups, is evaluated. It was found, using the def2-TZVPP basis set, that the ωB97M-V (MAD 1.8 kcal/mol), ωB97X-V (MAD 2.1 kcal/mol) and SCAN0 (MAD 2.1 kcal/mol) hybrid functionals are recommended. The double-hybrid functionals PWPB95 (MAD 1.6 kcal/mol) and B2K-PLYP (MAD 1.8 kcal/mol) did perform slightly better but this has to be balanced by their increased computational cost.


Author(s):  
Golokesh Santra ◽  
Nitai Sylvetsky ◽  
Gershom Martin

We present a family of minimally empirical double-hybrid DFT functionals parametrized against the very large and diverse GMTKN55 benchmark. The very recently proposed wB97M(2) empirical double hybrid (with 16 empirical parameters) has the lowest WTMAD2 (weighted mean absolute deviation over GMTKN55) ever reported at 2.19 kcal/mol. However, our xrevDSD-PBEP86-D4 functional reaches a statistically equivalent WTMAD2=2.22 kcal/mol, using just a handful of empirical parameters, and the xrevDOD-PBEP86-D4 functional reaches 2.25 kcal/mol with just opposite-spin MP2 correlation, making it amenable to reduced-scaling algorithms. In general, the D4 empirical dispersion correction is clearly superior to D3BJ. If one eschews dispersion corrections of any kind, noDispSD-SCAN offers a viable alternative. Parametrization over the entire GMTKN55 dataset yields substantial improvement over the small training set previously employed in the DSD papers.


2019 ◽  
Author(s):  
Golokesh Santra ◽  
Nitai Sylvetsky ◽  
Gershom Martin

We present a family of minimally empirical double-hybrid DFT functionals parametrized against the very large and diverse GMTKN55 benchmark. The very recently proposed wB97M(2) empirical double hybrid (with 16 empirical parameters) has the lowest WTMAD2 (weighted mean absolute deviation over GMTKN55) ever reported at 2.19 kcal/mol. However, our xrevDSD-PBEP86-D4 functional reaches a statistically equivalent WTMAD2=2.22 kcal/mol, using just a handful of empirical parameters, and the xrevDOD-PBEP86-D4 functional reaches 2.25 kcal/mol with just opposite-spin MP2 correlation, making it amenable to reduced-scaling algorithms. In general, the D4 empirical dispersion correction is clearly superior to D3BJ. If one eschews dispersion corrections of any kind, noDispSD-SCAN offers a viable alternative. Parametrization over the entire GMTKN55 dataset yields substantial improvement over the small training set previously employed in the DSD papers.


Nanomaterials ◽  
2021 ◽  
Vol 11 (6) ◽  
pp. 1581
Author(s):  
José C. Conesa

Two DFT-based methods using hybrid functionals and plane-averaged profiles of the Hartree potential (individual slabs versus vacuum and alternating slabs of both materials), which are frequently used to predict or estimate the offset between bands at interfaces between two semiconductors, are analyzed in the present work. These methods are compared using several very different semiconductor pairs, and the conclusions about the advantages of each method are discussed. Overall, the alternating slabs method is recommended in those cases where epitaxial mismatch does not represent a significant problem.


2019 ◽  
Vol 3 (12) ◽  
Author(s):  
Thomas Bischoff ◽  
Julia Wiktor ◽  
Wei Chen ◽  
Alfredo Pasquarello

2021 ◽  
Vol 140 (2) ◽  
Author(s):  
Rafael López ◽  
Frank Martínez ◽  
José Manuel García de la Vega

Sign in / Sign up

Export Citation Format

Share Document