The many faces of halogen bonding: A review of theoretical models and methods

Halogen bonds, the formally noncovalent interactions where the halogen acts as a Lewis acid, have brought several controversies to the theoretical world regarding its nature and components, e.g., charge transfer (CT), electrostatics, dispersion, and polarization. The debate on whether all characteristics are accounted for by electrostatics is examined, highlighting the importance of the CT and repulsive interactions. A number of strongly halogen-bonded complexes are as covalent as metal-ligand coordination bonds. Different levels of computational methods are reviewed with the objective of finding the best accuracy/cost ratios. The unusual electronic anisotropy of the halogen donor and its interaction with a Lewis base demand specific calculation schemes. From the wave-function theory methods, only the ones with empirical corrections (spin-component-scaled MP2 or CCSD, and MP2.5) are suitable when CCSD(T) is unattainable. Density functional theory functionals with a high amount of exact exchange are fast and reliable methods for halogen bonds, but double hybrids are more robust if other types of interactions are involved. Molecular mechanics methods can be useful, but only when specific corrections are added to compensate for the inability of such methods to describe CT. The most common method introduces a virtual site with a partial positive charge to account for the quantum chemical effect of the halogen bond. This methodology has been successfully applied to study protein-ligand interactions for drug design.