Abstract
The problem of evaluating the molecular force field, that is a potential surface, from experimental observations is examined in some detail using carbonyl sulphide as the example and basing the calculation of vibrational levels and rotational constants on matrix diagonalization. The influence of matrix size, of the expansion parameter Δr/r and of other features are discussed, as are dodges for hastening the least squares convergence. A method for numerical comparison of different fields is suggested.