Quantum Chemical Calculations of 31P NMR Chemical Shifts in Nickel Complexes: Scope and Limitations

The scope and limitations of a simple approach for the P-31 NMR chemical shift calculations of phosphorus atoms directly involved in the formation of coordination bonds with Ni have been analyzed. A comparative analysis of calculated versus experimental P-31 NMR shifts for the wide range of model nickel complexes based on small/-medium-sized organophosphorus ligands was carried out. Several functional-basis set combinations were tested. In general, for neutral singlet Ni complexes based on sigma- and pi-type ligands the P-31 NMR shifts can be calculated quite well in the framework of the Kohn-Sham level of theory with hybrid functionals (PBE0, B3LYP, B97-2). In the case of charged complexes, the predictions are less accurate due to the inherent fluxionality of the systems. For complexes with triplet contamination this approach cannot be used. The most accurate results were reached with the PBE0/6-311G(2d,2p)//PBE0/6-31+G(d) combination (RMSE < 7 ppm), while the GGA type functionals showed the most unreliable results, particularly for the pi-donating phosphorus. There are only two examples where calculated values disagreed with experiment. In the first case of a three-coordinate nickel phosphinidene complex, although calculations reproduce the exceptional low-field shift, the qualitative agreement is worse; this may be due to the effects of high spin states and medium effects. In the second case, a dramatic disagreement between calculations and experiment is due to the incorrect establishment of the structure. On the basis of these calculations, the structure should be revised. Thus, we concluded that in Ni complexes the Kohn-Sham level calculations can be safely used to predict P-31 NMR shifts of directly coordinated phosphorus. Moreover, the approach allows for the assignment of challenging structures with several coordination types.