Benchmark Accuracy in Thermochemistry, Kinetics, and Noncovalent Interactions

Composite ab initio methods are amongst the most accurate quantum chemical theories used for exploring energetic and spectroscopic properties at the molecular and atomic levels. This field of quantum chemistry began with the development of the Gaussian-n methods by John Pople, Larry Curtiss, Krishnan Raghavachari, and co-workers in the late 1980s. Due to major advances in quantum chemical theory, composite methods have advanced to the level where they can predict challenging thermochemical and kinetic properties with confident sub-kcal/mol accuracy. In addition, due to advances in high-performance supercomputer technology, these theories are routinely used for examining chemical processes involving molecules with dozens of atoms. Consequently, composite ab initio methods are being widely used in many branches of chemistry. Here we will review the top end of these procedures, namely composite ab initio methods that are capable of achieving benchmark accuracy (i.e., confident sub-kJ/mol predictions) in thermochemistry, kinetics, and noncovalent interactions. This chapter reviews the accuracy and applicability of high-level composite ab initio methods. General recommendations are given for selecting suitable methods for a given property and level of accuracy.