27 March 2023

Computational Chemistry II
Dr Krishna K. Govender
JOB 4205
[email protected]
Semi-empirical
• Also based on the Schrödinger equation, but like molecular mechanics methods it
makes use of parameters that are derived from experimental data (empirical = experimental)

• If data is not available then they are derived from theoretical methods

• The combination of theory and experimental data makes the method semi-empirical

• More approximations are made in solving the Schrödinger equation and very complicated
integrals (that are usually calculated for ab initio methods) are not evaluated

• The method draws on a library of integrals that is compiled by finding the best fit of some
calculated entity (e.g. energy) to an experimental value

2
Hartree-Fock

HF/6-31G

3
Semi-empirical

PM6

4
Semi-empirical
• A number of methods currently exist:
• Hückle Molecular Orbital (HMO)
• Pariser-Parr-Pople (PPP)
• Extended Hückle Theory (EHT)
• Complete Neglect of Different Overlap (CNDO)
• Intermediate Neglect of Differential Overlap (INDO)
• Modified Intermediate Neglect of Differential Overlap (MINDO)
• Neglect of Diatomic Differential Overlap (NDDO)
• Modified Neglect of Differential Overlap (MNDO)

5
MNDO
• Evaluates one-center two-electron integrals based on spectroscopic data and evaluates
other two-electron integrals using the idea of multipole-multipole interactions from
classical electrostatics
• Rather than determine various integrals analytically, numerical parameters are adjusted to
fit experimental data
• Cannot model intermolecular systems containing hydrogen bonds accurately
• Molecules possessing hypervalency are considerably unstable
• Four membered rings are too stable
• Rotational barriers are often underestimated
• Other MNDO methods were developed to overcome some of the problems. These include
Austin Model 1 (AM1), Parameteric Method Number 3 (PM3) and Parameteric Method
Number 6 (PM6)
6
Austin Model 1 (AM1)
• Re-parameterization of the general theoretical model found in MNDO

• The major difference is the addition of attractive and repulsive Gaussian Core Functions
(GCFs) to the description of the nuclear repulsion term to overcome MNDO’s hydrogen bond
problem

• Instead of fixing overlap and Slater orbital exponents, as was the case in MNDO, the terms
are parameterized separately during the development of the method

• Resulted in number of parameters being increased (7 per atom for MNDO to 13 – 19 per atom
for AM1)

7
Austin Model 1 (AM1)
• Predicts rotational barriers to be one third the actual barrier

• Predicting five-membered rings to be too stable

• Predicting hydrogen bonds with correct strength, but often the wrong orientation

• Geometries of compounds possessing hypervalent atoms are predicted poorly

• Despite the disadvantages the method is widely used due to its performance and
robustness

8
Parametric Method Number 3 (PM3)
• Unlike AM1 where parameterization was done by hand PM3 was a completely
automated process

• This was done by deriving and implementing formulas for the derivative of a suitable
error function with respect to the parameters

• All parameters could be optimized simultaneously and a significantly larger training

set with several hundred data could be employed

• The optimization does require some human intervention in selecting the experimental
data and assigning appropriate weight factors to each set of data

9
Parametric Method Number 6 (PM6)
• Is a method that has been parameterized for 70 elements of the periodic table.

• This was done by making several changes to the NDDO core-core interaction terms,
utilizing a different parameter optimization methodology and inclusion of d-orbitals to
the basis set

• For a set of 1373 compounds PM6 produced a mean unsigned error of 4.4 kcal/mol
compared to 6.3 and 10.0 kcal/mol obtained for PM3 and AM1, respectively

10
Parametric Method Number 6 (PM6)
• Missing repulsion between Na-Na, Br-N, Br-O, Br-Br, S-N, S-S, S-O, S-Cl, I-N, I-O and I-I pairs

• Production of an infinite error when the method was applied to crystal structures

• The incorrect prediction that a Si-O-H system is linear

• Procedural faults not detected during the development of the method

• Poor description of dispersion and hydrogen bond interactions

• Low accuracy in reproducing barrier heights for reactions

11
Density Functional Theory
• Requires computation of the total electron density and technically does not require a wave
function as in the case of HF

• In practice, however, DFT uses a wave function to compute some parts of the energy and the
electron density to compute other parts of the energy

• An understanding of mathematical functions and functionals is required to understand the

basic language of DFT

• Functions take a number as input and yield a number as output

𝑦 = 𝑓 𝑥 = 2𝑥
• Take x = 2 as input would yield f(x) = 4

[1] Kyle A. Baseden and Jesse W. Tye, Introduction to Density Functional Theory: Calculations by Hand on the Helium Atom, J. Chem. Educ., 2014, 91, 2116-2123 12
Density Functional Theory
• Functionals and functions of functions

• Functionals take a function as input and yield a number as output

• The input for a function is enclosed in parentheses (e.g. f(x)), whereas the input for a functional is
enclosed in square brackets (e.g. F[y])
3

𝐹 𝑦 = න 𝑓 𝑥 𝑑𝑥
0
• The above functional yields the area under any provided function y

• The functional above could take the function given earlier as input and yield the output F[2x] = 9

[1] Kyle A. Baseden and Jesse W. Tye, Introduction to Density Functional Theory: Calculations by Hand on the Helium Atom, J. Chem. Educ., 2014, 91, 2116-2123 13
Density Functional Theory
Etot = ET + EV + EJ + EX + EC

• Etot is the total energy of an atom, molecule or ion

• ET is the total kinetic energy of the electrons
• EV is the total potential energy of the electrons
• EJ is the total potential energy due to the average Coulomb repulsion between pairs of electrons
• EX is the total quantum mechanical exchange energy of the electrons
• EC is the total correlation energy of the electrons

• Ex and Ec correct for the overly simplistic EJ term, which is based on average repulsion between
electrons and does not consider the correlated motion of electrons
• ET and EJ are always positive, whereas EV, EX, and EC are always negetive

[1] Kyle A. Baseden and Jesse W. Tye, Introduction to Density Functional Theory: Calculations by Hand on the Helium Atom, J. Chem. Educ., 2014, 91, 2116-2123 14
Density Functional Theory
• Llewellyn Thomas and Enrico Fermi independently published atomic calculations that did not
require a wave function, but instead used the electron density as the fundamental variable in
Thomas-Fermi (TF) theory in 1927

• They made a lead of faith in assuming from the outset that the total electronic energy of an atom or
molecule is a functional of the electron density

• One attractive feature of the electron density is that it is experimentally observable, unlike
Schrödinger’s wave function

• TF theory can be considered a forerunner to modern DFT

• Over time several improvements were made to the original TF functional

[2] Thomas, L. H. The calculation of atomic fields. Math. Proc. Cambridge Philos. Soc. 1927, 23, 542−548.
[3] Fermi, E. Un metodo statistico per la determinazione di alcune proprietà dell’atome. Rend. Accad. Naz. Lincei 1927, 6, 602. 15
[4] Fermi, E. Eine statistische Methode zur Bestimmung einiger Eigenschaften des Atoms und ihre Anwendung auf die Theorie des periodischen Systems der Elemente. Z. Phys. 1928, 48, 73.
Hartree-Fock

HF/6-31G

16
DFT

B3LYP/6-31G

17
THE END

18