Talk:STO-nG basis sets

Unfortunately this article is actually completely wrong. First, STO-nG is not just any kind of minimum basis set, but is one in which n GTOs are fitted to a Slater-type orbital (STO) by a least square fit. Thus the STO-1G orbital is not exp(-αr²) with α chosen to minimize the energy but this function fitted to exp(-βr) for H with exponent β = 1.0. The result is α = 0.2709498091 not 0.2829421200. Second, the values for STO-nG with n > 1 are not the values you get doing a least square fit to exp(-βr), with β = 1.0, but actually those fitted to the STO with β ~ 1.24. This value is used in programs such as GAUSSIAN. β ~ 1.24 is appropriate for H atoms bonded to C atoms in organic molecules and was selected by Pople to give the best energy for methane. It is not surprising that the energy you get by fitting the GTOs to the actual STO for H, are much closer to -0.5. For example with STO-6G, the energy is 0.499826837.

However, this is all original research. I have just dug out the orbital exponents from a program and done some calculations. But the article itself is all original research as there are no references to the content. The one reference given no longer works. I need to get some references for all this, but I am too busy right now. When I have time, I will rewrite the whole article. --Bduke (Discussion) 00:21, 23 May 2009 (UTC)