SiC2 revisited in a DFT light

For the first time various pure and hybrid density functionals with three larger basis sets have been used to study the troublesome and the most forensic molecule, SiC2. It is concluded that BLYP and B3LYP functionals have failed to predict the cyclic form of SiC2 as the most stable structure over its linear counterpart even with the most flexible basis set, 6-311+G(3df). On the contrary, the other two pure density functionals of Perdew, Perdew and Wang with larger basis sets correctly identify the minimum energy structure. Finally, the two hybrid functionals B3P86 and B3PW91 predict the cyclic isomer of SiC2 to be more stable than the linear and the increase in the size of the basis set increases the stability of the former. The energy difference between the cyclic and the linear structures, 2.89 kcal/mol, obtained in B3PW91/6-311+G(3df) is better than that of MBPT(2) using the largest 120 CGTO basis set. This difference in energy is better than that of most of the highly sophisticated ab initio levels. Besides, it just deviates 1.03 and 0.62 kcal/mol from the two accurate levels in ab initio, CBS-Q and G2, respectively. The linear structure is shown to be a transition state in all the functionals with the two larger basis sets. The geometries obtained at B3P86 and B3PW91 levels using 6-311+G(3df) basis set excellently agree with the experimental values. The controversial v(3) vibrational mode obtained in the abovesaid hybrid functionals with the same large basis set coincides better with the experimental value than any of those obtained using ab initio methodologies. Some LYP-containing functionals result in a more stable bent structure on the SiC2 pinwheel surface lying between the linear and cyclic structures, but in some other functionals this bent structure is shown to be a transition state. Finally, no such bentlike structure of the SiC2 molecule has been found in B3P86 and B3PW91 levels with the 6-311+G(3df) basis set.