The singlet-state potential energy surface of the disilaethynyl anion (Si2H-) has been investigated using ab initio self-consistent-field (SCF), configuration interaction with single and double excitations (CISD), coupled cluster with single and double excitations (CCSD), and CCSD with perturbative triple excitations [CCSD(T)] levels of theory with large basis sets. Four stationary points [cyclic (monobridged) (1)A(1) (C-2v), linear (1)Sigma(+) (C-infinityv), bent (1)A' (C-s), and quasilinear (1)A' (C-s) structures] were located with the correlated wave functions, while only two stationary points [cyclic (monobridged) (1)A(1) (C-2v) and linear (1)Sigma(+) (C-infinityv) structures] were found with the SCF method. The cyclic structure (C-2v) is predicted to be the global minimum at all levels of theory. The linear structure (C-infinityv) is found to be a transition state between the two quasilinear structures (C-s) at the correlated levels of theory, while the SCF linear structure is predicted to be a transition state between the two cyclic structures. The quasilinear structure possesses a Si-Si-H bond angle similar to that of the monobridged Si2H2 molecule. The bent geometry is assigned to a transition state for the isomerization reaction between the cyclic and quasilinear structures. With the most reliable level of theory, augmented correlation-consistent polarized valence quadruple-zeta CCSD(T), the quasilinear structure is predicted to be 8.6 kcal/mol [7.9 kcal/mol with the zero-point vibrational energy (ZPVE) correction] above the cyclic (monobridged) structure, and the energy barrier for the cyclic-->quasilinear isomerization reaction is determined to be 12.1 kcal/mol (11.0 kcal/mol with the ZPVE correction). The inversion reaction between the quasilinear and linear structures is found to have a very small energy barrier. With the estimated aug-cc-pCVQZ CCSD(T) method the electron affinity of Si2H is predicted to be 2.31 eV, which is in excellent agreement with the experimental value 2.31+/-0.01 eV. (C) 2003 American Institute of Physics.