Applications of computing to the study of large chemical systems

Abstract

Experimental data pertaining to electronic structure of purine are briefly reviewed. It is shown from thermodynamic considerations that an approximate experimental order of stability for purine tautomers may be inferred from pKa data. This order serves as a reference point for comparison with ab initio computing on the same line of structures. Comparison of computed and experimental results indicates correct computed order of stability, but some energy differences are too large. The disagreement is attributed to ‘environmental effects’. In order to gain insight into the nature of solute—solvent interactions, McWeeny's ‘Group Functions’ formalism is introduced; a single solute molecule and the surrounding solvent form the interacting groups. In subsequent development it is found that the sum of ‘Polarization’ and ‘Dispersion’ terms which arise in the theory from a perturbation treatment of group interactions can be approximately estimated by a classical electrostatic treatment. The computed solvent effect operates to improve agreement with experimental results. The nature of approximations and limitations of this procedure is also discussed.