Page 220 - Adsorbents fundamentals and applications
P. 220

MOLECULAR ORBITAL THEORY CALCULATIONS  205

            8.2.5. Basis Set
            Both ab initio and DFT methods use sets of mathematical functions to represent
            the atomic orbitals. These are called the basis set. These mathematical func-
            tions are themselves made from a combination of simpler mathematical functions
            called primitives. Increasing the number of primitive functions and including
            contributions from valence orbitals imposes less restriction on the location of
            the electron and therefore more accurately models the exact molecular orbitals,
            but correspondingly increases the computational cost. The molecular orbitals are
            approximated as linear combinations of the basis functions. In the ab initio meth-
            ods, a Gaussian-type atomic function is used as the basis function, which has the
            general form:
                                              n m l −αr
                                   g(α,  r) = cx y z e  2                  (8.5)

                       r
            where vector   is the position of the electron, which is composed of coordinates
            x, y,and z,and α is a constant that determines the size (radial extent) of the
            function. The Gaussian function e −αr 2  is multiplied by powers (possibly zero) of
            x, y,and z and is normalized by constant c,sothat:

                                               2
                                             g = 1                         (8.6)
                                        allspace
            Linear combinations of primitive Gaussian functions shown in Eq. 8.5 are used
            to form the basis functions.
              A set of standard basis sets has been devised to increase the comparability
            between researchers and to simplify the nomenclature when describing the model
            chemistry: STO-3G, 3-21G, 6-311G, 6-311G(d,p), 6-311 + G(d,p)... in the
            order of increasingly large basis sets. The smallest of these, STO-3G, is the
            abbreviated name for three Gaussian primitives (3G) to model a Slater-type orbital
            (STO). When each orbital is represented by two or more sizes of basis functions,
            we have split-valence basis set. The 3-21G uses two sizes of Gaussian primitives
            to represent each orbital, and a 6-311G uses three sizes of Gaussian primitives
            to represent each orbital. To allow orbitals to change shape, we can include
            contributions from orbitals of higher angular momentum, for instance, include
            one of more d orbitals on carbon. The notation for this involves adding a letter
            for the orbital type to the end of the abbreviation, for example, 6-31G(2d,p) adds
            two d orbitals to heavy atoms and one p orbital to hydrogen atoms. A further
            development in basis has been to add very large versions of orbitals to the basis
            set, called diffuse functions. This is denoted by adding a + to the abbreviation
            before the G, so 6-31+G adds diffuse functions to heavy atoms and 6-31++G
            adds diffuse functions to both heavy and hydrogen atoms. The best choice of
            basis set is largely dependent on the chemistry being studied.

            8.2.6. Effective Core Potentials
            Relativistic effects must be considered in the applications of ab initio molecular
            orbital calculations for the heavier elements; they have a significant influence on
   215   216   217   218   219   220   221   222   223   224   225