Assessment of the Handy-Cohen optimized exchange density functional for organic reactions

We have investigated the performance of the new optimized exchange functional (OPTX) developed by Handy and Cohen [Mol. Phys. 99, 403 (2001)] for predicting geometries, heats of reaction, and barrier heights for twelve organic reactions (six closed-shell and six radical). OPTX has been used in conjunction with, among others, the well-known Lee-Yang-Parr (LYP) correlational functional to form two new functionals, OLYP and O3LYP. These are similar to the well-established BLYP and B3LYP functionals, respectively, with OPTX replacing the standard Becke exchange functional, B88. Our results strongly support claims made by their developers that OLYP is superior to BLYP, and essentially renders it obsolete. The computed OLYP heats of reaction, barrier heights, and even molecular geometries (with larger basis sets), are comparable with, if not better than, the corresponding B3LYP values. The O3LYP functional is overall better than B3LYP, albeit not by much. Both OLYP and O3LYP are among the best functionals currently available; the performance of OLYP in particular is noteworthy given that this functional includes no exact exchange. (C) 2002 American Institute of Physics.