Can DFT methods correctly and efficiently predict the coordination number of copper(I) complexes? A case study

The coordination number of various experimentally known Cu(I) compounds is studied using density functional theory. Various basis sets are tested, aiming to establish a reliable level for prediction of the coordination number of these and other Cu(I) complexes. It is found that most levels exhibit correct trends, namely, the bulkier ligands demonstrate larger preference for coordination of two ligands. Proper absolute values are obtained when dispersion corrections are also included in the calculations. It is concluded that the fairly small modified 6-31+G* basis set due to Pulay represents a good compromise between accuracy and efficiency, followed by Balabanov and Peterson's all-electron aug-cc-pVDZ basis set. The overall energy is decomposed into various components whose relative contribution to the overall tendency of forming a complex with a particular coordination is examined. It is shown that two opposing contributions play a major role: the interaction energy of the ligand being added and the deformation energy of the copper's coordination sphere prior to the ligand addition. The former being a stabilizing contribution, leads to higher coordination numbers while the later, a destabilizing contribution, is shown to favor lower coordination numbers.