1 Answers
STO-nG basis sets are minimal basis sets, where n {\displaystyle n} primitive Gaussian orbitals are fitted to a single Slater-type orbital. n {\displaystyle n} originally took the values 2 – 6. They were first proposed by John Pople. A minimum basis set is where only sufficient orbitals are used to contain all the electrons in the neutral atom. Thus for the hydrogen atom, only a single 1s orbital is needed, while for a carbon atom, 1s, 2s and three 2p orbitals are needed. The core and valence orbitals are represented by the same number of primitive Gaussian functions ϕ i {\displaystyle \mathbf {\phi } _{i}}. For example, an STO-3G basis set for the 1s, 2s and 2p orbital of the carbon atom are all linear combination of 3 primitive Gaussian functions. For example, a STO-3G s orbital is given by:
where
The values of c1, c2, c3, α1, α2 and α3 have to be determined. For the STO-nG basis sets, they are obtained by making a least squares fit of the three Gaussian orbitals to the single Slater-type orbitals. This differs from the more common procedure where the criteria often used is to choose the coefficients and exponents to give the lowest energy with some appropriate method for some appropriate molecule. A special feature of this basis set is that common exponents are used for orbitals in the same shell as this allows more efficient computation.
The fit between the Gaussian orbitals and the Slater orbital is good for all values of r, except for very small values near to the nucleus. The Slater orbital has a cusp at the nucleus, while Gaussian orbitals are flat at the nucleus.