A series of molecules related to malonaldehyde, containing an intramolecular H-bond, are used as the testbed for a variety of levels of ab initio calculation. Of particular interest are the excitation energies of the first set of valence excited states, nπ* and ππ*, both singlet and triplet, as well as the energetics of proton transfer in each state. Taking coupled cluster results as a point of reference, configuration interaction-singles–second-order Møller–Plesset (CIS–MP2) excitation energies are too large, as are CIS to a lesser extent, although these approaches successfully reproduce the order of the various states. The same may be said of complete active space self-consistent-field (CASSCF), which is surprisingly sensitive to the particular choice of orbitals included in the active space. Complete active space–second-order perturbation theory (CASPT2) excitation energies are rather close to coupled cluster singles and doubles (CCSD), as are density functional theory (DFT) values. CASSCF proton transfer barriers are large overestimates; the same is true of CIS to a lesser extent. MP2, CASPT2, and DFT barriers are closer to coupled cluster results, although yielding slight underestimates.
No takes yet. Share an insight, caveat, or question.
Kar et al. (1999) studied this question.