Ab initio valence-bond wave functions are reported for the 3B1, 1A1, and 1B1 states of CH2 as a function of angle with R(C-H) = 2,00 bohrs. Up to 48 valence-bond structures for each state were used to expand the variational wave functions (each VB structure is a single term of an expansion which is composed of those appropriate linear combinations of determinants required to assure a definite spin and space symmetry). A nonorthogonal free atom basis constructed with Gaussian lobe functions was employed. Calculations with scaled hydrogen orbitals were likewise carried out. The total energy of the ground state, 3B1, near its equilibrium position is ET (140˚) = –38.9151 Hartree units. Ab initio MO-SCF wave functions were also computed as a function of angle for the 1A1 state, and MO wave functions for the 1,3B1 states were constructed from single excitations of the MO 1A1 state. The order of states is computed to be 3B1 < 1A1 < 1B1 with large equilibrium angle for 1,3B1, 108˚ for 1A1, and an lA1-1B1 energy separation of 1.52 eV. Spectroscopic data are available for these properties, and the agreement between theory and experiment is quite good. Larmor diamagnetic susceptibility terms are computed to be –19.75 × 10-6 and –19.57 × 10-6 erg/(G2 mole) for the 1A1 and 1B1 states, respectively, compared to an estimated experimental value of –12 × 10-6. The 3B1 heat of atomization is calculated as 6.36 eV (experimental value ≅ 8.5 eV). Expectation values of the following quantities (for which experiments are currently lacking) have been obtained as a function of angle: dipole moments, quadrupole tensor, diamagnetic contribution to the nuclear magnetic shielding constant of the protons, diamagnetic anisotropy, electric field gradient tensor, quadrupole coupling constants for deuterated methylene, 〈l/rH〉, 〈1/rc〉, heats of atomization of the 1Ai and 1B1 states, the 2B1-1A1 energy separation, and oscillator strength for singlet-singlet transitions. Considerable attention has been given to detailed descriptions of the charge distributions and to the implications of our results for divalent carbon chemistry. There is a long and intricate history of experimental and theoretical attempts to elucidate the electronic structure of methylene, and in order to aid overall understanding of this molecule, a rather extensive review and analysis of previous work has been included. © 1969, American Chemical Society. All rights reserved.
