Efficient calculation of heats of formation

J Phys Chem A. 2009 Mar 12;113(10):2165-75. doi: 10.1021/jp810144q.

Abstract

An efficient procedure has been devised for calculating heats of formation of uncharged, closed-shell molecules comprising H, C, N, O, F, S, Cl, and Br. Known as T1, it follows the G3(MP2) recipe, by substituting an HF/6-31G* for the MP2/6-31G* geometry, eliminating both the HF/6-31G* frequency and QCISD(T)/6-31G* energy and approximating the MP2/G3MP2large energy using dual basis set RI-MP2 techniques. Taken together, these changes reduce computation time by 2-3 orders of magnitude. Atom counts, Mulliken bond orders, and HF/6-31G* and RI-MP2 energies are introduced as variables in a linear regression fit to a set of 1126 G3(MP2) heats of formation. The T1 procedure reproduces these values with mean absolute and rms errors of 1.8 and 2.5 kJ/mol, respectively. It reproduces experimental heats of formation for a set of 1805 diverse organic molecules from the NIST thermochemical database with mean absolute and rms errors of 8.5 and 11.5 kJ/mol, respectively. Heats of formation of flexible molecules have been approximated by the heats of formation of their lowest-energy conformer as given by the T1 recipe. This has been identified by examining all conformers for molecules with fewer than 100 conformers and by examining a random sample of 100 conformers for molecules with more than 100 conformers. While this approximation necessarily yields heats of formation that are too negative, the error for typical organic molecules with less than 10 degrees of conformational freedom (several thousand conformers) is <2-3 kJ/mol. T1 heats of formation have been used to calculate energy differences for a variety of structural, positional, and stereoisomers, as well as energy differences between conformers in a variety of simple acyclic and cyclic molecules for which reliable experimental data are available. In terms of both overall error and errors for individual systems, T1 provides a better account of the experimental thermochemistry than any practical quantum chemical method that we have previously examined. A database of approximately 40,000 T1 calculations for both rigid and flexible organic molecules has been produced and is available as part of the Spartan Molecular Database (SMD) in the current version of the Spartan electronic structure program (Spartan'08). (A subset of approximately 5000 molecules is provided as part of the standard release, and the full T1 database can be licensed.). This collection differs from the other components of SMD in that the lowest-energy conformation for each molecule has been assigned using a high-level quantum chemical method and not molecular mechanics. Thus, it is not only a source of "high-quality" calculated heats of formation for organic molecules but also a source of conformational preferences.