Reliability of ab initio (HF, post HF and DFT) methods and basis set dependencies for accurate prediction of equilibrium re distances of CO bond lengths

Abstract

Optimized bond lengths for molecules containing CO bonds have been calculated by HF, post HF (MP2, LMP2, MP4(SDQ), CCSD(T) and DFT (BLYP, B3LYP, B3PW91, BHandH) methods using different basis sets. Linear regressions of calculated and experimentally available re CO bond lengths of organic molecules have been performed. The best basis set out of the set of up to 10 standard basis sets within a theoretical method was derived from a ranking procedure based on statistical parameters. This ranking scheme allows also the determination of the best method/basis set combination for all selected methods and basis sets.We studied the application of our recently developed prediction schemes for estimations of higher order basis set CO distances from lower order calculations and estimations of CO re bond lengths from linear regression equations. These procedures allow predictions of near re CO distances of high accuracy even form modest level of theory. Therefore this leads to a chance for treatment of large or difficult to handle molecules for which experimental determinations of re structures are inaccessible.