Effects of wave function modifications on calculated H-C and C C stretching frequencies

摘要

Two-level factorial design (FD) and principal component (PC) models are used to determine the effects of wave function modifications on calculated H-C and C C harmonic stretching frequencies for the H-C CH, H-C CF, H-C CCl, H-C CCCH, H-C CCN and H-C CCH3 molecules. Our result have shown that valence (val), diffuse (dif), polarization (pol), and electron correlation (corr) main effects as well as some second-order interaction effects are significant to build factorial models for these vibrational frequencies. For instance, when valence and diffuse function are introduced in the basis set, the calculated H-C and C C stretching frequencies decrease. However, this reduction is much more accentuated when the electron correlation effect is introduced in the Hartree-Fock wave functions. By far, the corr main effect is the most important one on the H-C and C C frequency values. The MP2 electron correlation effect provokes an increment of -157.4 cm(-1) and -283.8 cm(-1) on the H-C and C C stretching frequencies, respectively, whereas the dif effect produces only a reduction of -9.1 cm(-1) and -12.9 cm(-1) on these frequencies, respectively. The inclusion of Moller-Plesset 2 perturbation in the Hartree-Fock wave functions also produces an important pol-corr interaction effect on the H-C and C C stretching frequencies, increasing their values by +45.2 cm(-1) and +17.3 cm (1), respectively. Algebraic models were then established to explain how calculated H-C and C C stretching frequencies depend on characteristics of the molecular orbital wave functions. These models were successful in reproducing calculated H-C and C C frequency values for the H-CN and CH3C CCH3 molecules, which were not included in the training set. The principal component analysis has revealed that the calculated H-C and C C stretching frequencies can be adequately described by a single principal component. It is capable of explaining more than 99% of the total data variance. (C) 2008 Wiley Periodicals, Inc.