Norm-conserving pseudopotentials with chemical accuracy compared to all-electron calculations

Abstract

60439By adding a non-linear core correction to the well established Dual Space Gaussian type pseudopotentials for the chemical elements up to the third period, we construct improved pseudopotentials for the Perdew Burke Ernzerhof (PBE) [J. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996)] functional and demonstrate that they exhibit excellent accuracy. Our benchmarks for the G2-1 test set show average atomization energy errors of only half a kcal/mol. The pseudopotentials also remain highly reliable for high pressure phases of crystalline solids. When supplemented by empirical dispersion corrections [S. Grimme, J. Comput. Chem. 27, 1787 (2006); S. Grimme, J. Antony, S. Ehrlich, and H. Krieg, J. Chem. Phys. 132, 154104 (2010)] the average error in the interaction energy between molecules is also about half a kcal/mol. The accuracy that can be obtained by these pseudopotentials in combination with a systematic basis set is well superior to the accuracy that can be obtained by commonly used medium size Gaussian basis sets in all-electron calculations.

Computationally efficient characterization of potential energy surfaces based on fingerprint distances

“…Fig. 8 displays disconnectivity graphs for Si 20 , (NaCl) 29 , (NaCl) 32 , LJ 19 , LJ 38 , and LJ 55 . As above, the PES of Si 20 was computed at the DFT level of theory as implemented in the BigDFT code (PBE exchange correlation functional).…”

“…5, the fitting function Eq. (4) is plotted on top of the uphill barrier distributions of Si 20 , (NaCl) 29 , (NaCl) 32 , LJ 19 , LJ 38 , and LJ 55 . From these plots it is evident that the binned average of the absolute values of the energy differences of hopping pair minima is a reasonable guess for the uphill barrier energy.…”

“…3 is seen. For this verification, transition states and their directly connected minima were computed for Si 20 and Au − 26 at the DFT level of theory as implemented in the BigDFT [36][37][38] code and for (NaCl) 32 and (NaCl) 29 using the Born-Mayer-Huggins-Tosi-Fumi [46][47][48][49][50] − 26 , the geometries and energies of the minima, as well as their connectivity, were established using the MHGPS method as implemented in the BigDFT suite. In the case of Au − 26 the data were taken from a previous study and it is referred to this study for a description of its computation.…”

“…The structures, energies and the connectivity of the stationary points were determined at the DFT level of theory (PBE exchange correlation functional) by using the MHGPS method coupled to the BigDFT code. 21,22,[36][37][38] The shown density was obtained from the corresponding scattered data by means of a Gaussian kernel density estimate as implemented in Python's scipy library. The red bold line shows the same data, but averaged within 25 bins along the RMSD axis.…”

A generalized Poisson and Poisson-Boltzmann solver for electrostatic environments

“…The Perdew-Berke-Ernzerhof (PBE) functional 37 and the extended triple-zeta 6-311+g(d,p) basis set were used for both geometry optimizations and energy calculations, in vacuum and in solution, consistently with the setup used for the parameterization of the electrostatic solvation energy in sccs. 15 In all calculations, soft norm-conserving pseudopotentials including non-linear core correction 38,39 along with PBE functional were used to describe the core electrons and exchange-correlation, respectively. For the charge-dependent cavity, values of ρ max = 5 · 10 −3 a.u.…”