Noncovalent interactions are ubiquitous in molecular and condensed-phase environments, and hence a reliable theoretical description of these fundamental interactions could pave the way toward a more complete understanding of the microscopic underpinnings for a diverse set of systems in chemistry and biology. In this work, we demonstrate that recent algorithmic advances coupled to the availability of large-scale computational resources make the stochastic quantum Monte Carlo approach to solving the Schrödinger equation an optimal contender for attaining "chemical accuracy" (1 kcal/mol) in the binding energies of supramolecular complexes of chemical relevance. To illustrate this point, we considered a select set of seven host-guest complexes, representing the spectrum of noncovalent interactions, including dispersion or van der Waals forces, π-π stacking, hydrogen bonding, hydrophobic interactions, and electrostatic (ion-dipole) attraction. A detailed analysis of the interaction energies reveals that a complete theoretical description necessitates treatment of terms well beyond the standard London and Axilrod-Teller contributions to the van der Waals dispersion energy.
Keywords: Monte Carlo; Schrödinger equation; electrostatic attraction; hydrogen bonding; hydrophobic interaction; noncovalent; van der Waals; π−π stacking.