Improving the accuracy of solid-state nuclear magnetic resonance chemical shift prediction with a simple molecular correction

Abstract

A simple molecular correction improves significantly the accuracy of predictions of solid-state NMR chemical shifts.

Towards Accurate Predictions of Proton NMR Spectroscopic Parameters in Molecular Solids

“…Surprisingly, the best improvement of the calculated shieldings is obtained when the very simple and fast PBE0 correction is employed (Δδ PBE0 values ranging from −1.4 to 7.6 ppm); the maximal error drops to 1.86 ppm. These calculations thus confirm the previous finding that these corrections calculated with a hybrid density functional improve the accuracy of carbon shieldings significantly . The SO corrections are less significant; the largest correction of 2.5 ppm is found for the l ‐cysteine carbon atom directly attached to the sulfur atom, the correction is close to 1.7 ppm for all COO carbons and close to 1 ppm or smaller for all other atoms.…”

“…Experimental carbon chemical shifts, the calculated shieldings and shielding corrections are shown in Table and the parameters of the linear shielding‐shift correlations are shown in Table . The carbon shieldings calculated for the geometry optimised structures without any correction correlate reasonably well with the experimental chemical shifts, the MAE is 1.09 ppm and maximal error is 3.21 ppm, which are typical values that can be expected for molecular solids . Surprisingly, the best improvement of the calculated shieldings is obtained when the very simple and fast PBE0 correction is employed (Δδ PBE0 values ranging from −1.4 to 7.6 ppm); the maximal error drops to 1.86 ppm.…”

“…It has been shown recently that a simple correction of GIPAW shieldings calculated with the hybrid functional PBE0 on an isolated molecule improves the shielding‐shift correlations for carbons, nitrogens and oxygens significantly . The corrections to the amino‐acid proton shieldings calculated with the hybrid PBE0 functional (Δσ PBE0 ) are generally not very large (−0.2 to +0.3 ppm, shown in Table ); they are positive for all CH, CH 2 and CH 3 hydrogens except for one of the alpha hydrogens in glycine (−0.08 ppm), and negative for all NH 3 protons.…”

“…We used the 6‐311+G(2d,p) basis set for the calculations of molecular corrections because excellent results were obtained previously with this basis set for carbon chemical shifts. Furthermore, it was observed that the basis‐set choice has a relatively small effect on the calculated molecular corrections of 13 C, 15 N and 17 O shifts . In order to check whether the choice of the basis set plays an important role in the calculations of proton shieldings, we recalculated the MP2 corrections for glycine, l ‐alanine and l ‐serine with aug‐ccPVDZ and aug‐ccPVTZ basis sets.…”

“…However, it has been shown in several studies using computations on molecular clusters or fragments of infinite crystals that the use of hybrid density functional or high‐level ab initio methods, such as coupled cluster singles and doubles (CCSD), often provides more accurate predictions of NMR parameters of solids . We have recently demonstrated that a simple correction to the GIPAW‐GGA result calculated on an isolated molecule at a higher level of theory significantly improves the correlations between experimental and calculated carbon, nitrogen and oxygen chemical shifts …”