Jorge Escorihuela

COMPUTATIONAL ORGANIC CHEMISTRY

CONCEPTS
Summary

❑ Introduction.
❑ Quantum mechanics.
❑ Theoretical methods.
❑ Ab initio methods.
❑ Semiempirical methods.
❑ Methods based on density-functional theory.
❑ Solvation models.
❑ Analysis of the potential energy surfaces.
❑ Electronic properties.
❑ Thermodinamics.
Bibliography
Summary

❑ Introduction.
❑ Quantum mechanics.
❑ Theoretical methods.
❑ Ab initio methods.
❑ Semiempirical methods.
❑ Methods based on density-functional theory.
❑ Solvation models.
❑ Analysis of the potential energy surfaces.
❑ Electronic properties.
❑ Thermodinamics.
Introduction
Computational Chemistry (molecular modeling):
• A set of theoretical methods and computational techniques used to model or mimic the
behaviour of molecules and investigate chemical problems on a computer

The common feature of molecular modeling methods is the atomistic level of

description of the molecular systems

Why Computational Chemistry?

Because it is:

• a tool for understanding and predicting chemical behaviour

• a road map guiding experimental work (when used properly)

Introduction
Challenges in Catalysis
Catalysis is the quiet force behind the modernization of our chemical industry. It
ensures:
❑ Efficiency (Energy)
• Optimizing reaction conditions: mild conditions
❑ Prevent waste
• Increasing selectivity (avoiding side products)
• Maximizing the number of atoms of reactants incorporated into the
products: Atom economy
❑ Industry safer
• Moving to greener reaction media (Ex. water)
• Avoiding the use of hazard chemicals (other routes)
Chemists need a higher level of control
over the chemical processes

Deeper knowledge at molecular level

Introduction
The Chemists’ paradox …
BUT, we
How to obtain
Chemists deal with “don’t see” information at the
molecules: synthesize them,
molecules microscopic level?
determine their properties,
make them react…..

The microscopic world is accessible …

❑ Spectroscopy ❑ Another technique :
Use the ▪ Computational Modeling
❑ Scanning Probe Microscopy
computer as a and Simulation
▪ STM microscope to
▪ AFM look into a
▪ Etc. virtual world
Introduction
Molecular Modelling
Selection of the “Molecular Modelling Methods”
depends on:
 Chemical Problem
◼ Heterogeneous Catalysis – Surface chemistry
◼ Homogeneous Catalysis – Transition Metal Complexes
◼ Organic reactions
◼ Enzymes
◼ Etc.

 Accuracy of the results

◼ Obtaining general trends
◼ Obtaining accurate parameters
(geom., energies, NMR, …)
Introduction
The Old The New

Project design Explaining after-facts Project design

Missing out the full Initial Screening

Experimental potential of the and Mechanistic
Screening technique Studies
Guiding
Calculations must be experiments
Experimental incorporated at the Calculations
Scope Studies experimental
design stage

Experimental Experiment
Mechanistic Calculations
Studies
Adapted from Peng Liu, Univ. Pittsburgh
Introduction
The Old
Proposed mechanism

Angew. Chem., Int. Ed. 2020, 59, 11808−11813

Introduction
The Old

▪ Gaussian 09
▪ B97D functional
▪ PCM solvent DCE
▪ SDD basis set (Au,Sb,I)
▪ 6-31G(d,p) basis set

▪ B97D functional
▪ PCM solvent DCE
▪ SDD basis set (Au,Sb,I)
▪ 6-311+G(d,p)
J. Org. Chem. 2022, 87, 4078−4087
Introduction
Molecular Modelling
Two aspects that need to be considered when applying
Molecular Modelling:

Chemical Computational
model model
How to represent the Computational techniques
“real” chemical word ! to be used.

The Modelers’ dilemma: which model to choose?

Introduction
Calculations deal with Models
The Modelers’ dilemma: which model to choose?

A model of a cat

<

“A Model must be wrong, in some respects, else it would be the thing itself.
The trick is to see where it is right.”
Henry A. Bent
Introduction
Chemical Model Catalysis
Heterogeneous Homogeneous

SOLVENT

The larger the model the more expensive

the calculation

Periodic Cluster Modelled Complete

methods methods complex complex
Introduction
Computational Model

F=m·a

H=E

First-principles calculations Atomistic simulations

Introduction
Computational Model

Journal of Physics D Applied Physics 2015, 49(5):054002

Introduction
Computational Model
Applied to Transition Metal Compounds
Historical Overview

1970s Qualitative MO Theory

Computer Performance
Qualitative. Bonding, Reactivity
1990s Standard Ab Initio, DFT First Cray

Quantitative. Geometries, E, IR, NMR,....

Model Systems

2000s Hybrid Methods. QM/MM

IBM RS/6000
Steric Ligand Effects
“Real” TM Systems
2010s Molecular Dynamics, Carr-Parrinello
PC Cluster
Dynamic effects. Solvent effects
Introduction
Computational Model
The underlying physical laws necessary for the mathematical theory of
a large part of physics and the whole of chemistry are thus completely
known, and the difficulty is only that the exact application of these
laws leads to equations much too complicated to be soluble. It the
It therefore
refore becomes
becomes desirable
desirable that approximate
that approximate practical
practical methods
methods of of
applying quantum mechanics should be developed, which can lead to
an explanation of the main features of complex atomic
compatomic systems
ystems without
without too
too
much much computation.
computation.
Paul A. M. Dirac
Proceedings of the Royal Society of London. Series A, Containing
Papers of a Mathematical and Physical Character, Vol. 123, No. 792
(1929)
The Nobel Prize in Physics 1933
Erwin Schrödinger, Paul A.M. Dirac

"for the discovery of new productive

forms of atomic theory."

48 years later…..
Introduction
Computational Model

Kenichi Fukui Roald Hoffmann

The Nobel Prize in Chemistry 1981

"for their theories, developed independently, concerning the course of
chemical reactions”
17 years later…..
Introduction
Computational Model

John Pople Walter Kohn

The Nobel Prize in Chemistry 1998

The Nobel Prize in Chemistry 1998 was divided equally between Walter Kohn
"for his development of the density-functional theory"

and John A. Pople

"for his development of computational methods in quantum chemistry“.
17 years later…..
Introduction
Computational Model

Martin Karplus Michael Levitt Arieh Warshel

The Nobel Prize in Chemistry 2015

"for the development of multiscale models for complex chemical systems".

QM MM Quantum Mechanics vs. Molecular Mechanics

Introduction
Advantages from Computation
▪ Calculations are easy to perform, whereas experiments are often difficult.

▪ Calculations are becoming less costly, whereas experiments are becoming

more expensive.

▪ Calculations can be performed on any system, even those that don’t exist,
whereas many experiments are limited to relatively stable molecules.

▪ Calculations are safe, whereas many experiments can have an intrinsic

danger associated.
Introduction
Desadvantages from Computation
▪ Calculations are too easy to perform, whereas experiments are often
difficult.

▪ Calculations can be very expensive in terms of the amount of time

required.

▪ Calculations can be performed on any system, even those that don’t exist.

Computational chemistry does not replace experimental

studies, but plays an important role in:
▪ Rationalising and explaining known chemistry
▪ Exploring new or unknown chemistry
Summary

❑ Introduction.
❑ Quantum mechanics. Esta parte es la base de lo de despues pero esto no se pregunta en el examen
❑ Theoretical methods.
❑ Ab initio methods.
❑ Semiempirical methods.
❑ Methods based on density-functional theory.
❑ Solvation models.
❑ Analysis of the potential energy surfaces.
❑ Electronic properties.
❑ Thermodinamics.
Computational Models
Methodology:
(Listed in order from most to least accurate)
– Ab initio - DFT
• uses Schrödinger's equation, or Kohn-Sham equations but with approximations
• Ab initio means “from the beginning” or “from first principles”.
– Semi Empirical
• uses experimental parameters and extensive simplifications of Schrödinger's
equation
– Molecular Mechanics
• does not use Schrödinger's equation
• Based on classical laws (Newton, Coulomb, Hook)
Computational Model
Theoretical basis

Hartree-Fock based Density Functional Theory

(HF) (DFT)

 Hartree-Fock:  LDA:
 D (Slater determinant)  GGA:
 Meta-GGA:
 Post-Hartree-Fock:  Hyper-GGA
 Double hybrid
Solving the Schrödinger Equation

Hˆ = Tˆ + Vˆ Hˆ = Eˆ kinetic + Eˆ potential
For many electron atoms/molecules:

Nuclei Electron Nuclear- Nuclear- Electron-

kinetic kinetic Nuclear electron electron
energy energy repulsion attraction repulsion

Tˆ
æ - 2 d2 ö
Vˆ
ç ÷ where:
è 2m dx 2 ø 27
Solving the Schrödinger Equation

 Three parts to solve Schrödinger equation for molecules:

 Born-Oppenheimer Approximation
◼ Leads to the idea of a potential energy surface

 The expansion of the many-electron wave function in

terms of Slater determinants.
◼ Often called the “Method”

 Representation of Slater determinants by molecular

orbitals, which are linear combinations of atomic-like-
orbital functions.
◼ The basis set
Born-Oppenheimer aproximation
 The wave-function of the many-electron molecule is a
function of electron and nuclear coordinates: (R,r)
(R=nuclear coords, r=electron coords).
 The motions of the electrons and nuclei are coupled.
 However, the nuclei are much heavier than the electrons
mp ≈ 2000 me
 Consequently, nuclei move much more slowly than do the
electrons (E=1/2mv2). Nuclei appear fixed to e−
 Born-Oppenheimer (BO) Approximation: to a high degree
of accuracy we can separate electron and nuclear motion:

(R,r)= el(r;R) N(R)

Approximating ψ: The Method

 elec Hˆ elec elec dr
*

Eelec = 0

Eelec =  elec Hˆ elec  elec
 elec  elec dr = 1 (normalization)
*
0 Dirac “braket” notation for
integrals

The Schrödinger equation can be solved analytically (‘exactly’) only for the simplest
systems (H, He+, Li2+, etc).

Variational Principle:

E'(F’')

Approximate energies, E'' (F'')

trial wavefunctions

E'''(F''')

Exact energy,
real wavefunction Eelec (elec)
Approximating ψ: The Method
Hartree-Fock approximation:
Expansion of orbitals in terms of some basis functions centered Unoccupied
on the nuclei: (virt)

c
coefficients
b

 i =  ci  basis functions

 k
j
Linear Combination of Atomic Orbitals (LCAO)
i

Occupied
(occ)
Approximating ψ: The Method
Hartree-Fock approximation:
Total wave function as Slater Determinant

LCAO: Linear Combination of

Atomic Orbitals

SCF

→
Variación Principle: method
Hartree-Fock (HF) energy
The lowest the best
 Approximating ψ: The Method
Hartree-Fock approximation:
Restricted (RHF,
(RHF) ROHF) and unrestricted (UHF) solutions:

Closed shell Open shell

/b /b  b  b
E

RHF ROHF UHF UHF

singlet doublet singlet doublet
 Computational methods
Current status of HF methods

☺ Strong points:  Weak points:

 HF limit result is upper bound of true energy for system  Bond interactions tend to be overestimated
 Acceptable results for minimum energy structures  Bond lengths are usually too short
 ! Using double-zeta polarized basis sets...  Failure to account for electron correlation
 Errors in heavy atom bond length: ~0.03 A  Should not be used for reaction energies
 Errors in heavy atom-hydrogen bond length: ~0.015 A  Should not be used for structures with dispersion
 Errors in bond angles: ~1.5o
 Dihedral angles are generally good
 Heavier atom predictions lack inclusion of relativistic effects
 H-bonds are generally good  Unoccupied orbitals are not optimized
 Approximating ψ: The Method
Electron Correlation
Hartree-Fock
E H · + H· H–H
calculations recover ~99%
HF underestimated
of total energy
“exact” binding energy
at HF level
HF energy
Exact (correlated) energy

Why is the correlation energy so important?

Due to the absence of correlation energy, HF calculations usually lead to:

- too large stretching bond energies too large activation energies for bond
formation reactions.
- too short bonds
- too large vibrational frequencies
- wavefunctions with a too ionic character.
Approximating ψ: The Method
Electron Correlation
 Electron Correlation: Difference between energy calculated with exact
wave-function and energy with Hartree-Fock wavefunction.
Ecorr = Eexact - EHF
 Accounts for the neglect of instantaneous electron-electron interactions of
Hartree-Fock method.
 In general, correlation energy is considered by adding additional Slater
determinants to expansion of .
y el = d0 F HF + å di F i
i =1
 Hartree-Fock wavefunction is often used as starting point.
 Additional Slater determinants are often called “excited”.
 Different correlation methods differ in how they choose which Fi to include
and how to calculate the di coefficients.
Approximating ψ: The Method
Electron Correlation
Orbital Energy →

HF S-type S-type D-type D-type T-type Q-type

Excited Slater Determinants

Approximating ψ: The Method
Post-Hartree-Fock Methods
Electron Correlation
Methods

Configuration Interaction Coupled Cluster Many Body Perturbation Theory

(CI) (CC) (MBPT)

CISD CCSD MP2

CISD(T) CCSD(T) MP3
CISDT CCSDT MP4
CISDTQ QCISD ……
…… QCIST(T)
……
In practice, none of these methods are exact, but they all provide a well-defined route to exactitude.
Approximating ψ: The Method
Post-Hartree-Fock Methods
❑ Moller-Plesset perturbation theory (MPx):
Use of perturbation to account for electron correlation
❑ Multi-Configurations Self Consistent Field (MCSCF):
Orbitals are optimized for a combination of configurations : molecular spin states and
occupation numbers of orbitals
❑ Configuration Interaction (CI):
Use of CASSCF (Complete Active Space SCF) with inclusion of specified excited states e.g.
CISD includes single and double excited states
❑ Coupled Cluster theory (CC):
Includes excitation operator to generate specified excited states CCSD(T) is the "gold
standard" of quantum chemistry
Computational methods
Current status of post-HF methods
☺ Strong points:
 Well defined hierarchy of methods
HF < MP2 < CASPT2 < CCSD(T) < Full CI
 Necessary for high accuracy or tricky systems

 Weak points:
 Poorer quality / price ratio than DFT
◼ MP2 quality similar to GGA quality
◼ MP2 computational price higher than GGA
Post-HF methods
Moller-Plesset perturbation theory (MPx)
Correlation can be added as a perturbation from the Hartree-
Fock wave function using the Moller-Plesset perturbation
theory. In mapping the HF wave function onto a perturbation
theory formulation, HF becomes a first-order perturbation.
Thus, a minimal amount of correlation is added by using the
second-order MP2 method. Third-order (MP3) and fourth-
order (MP4) calculations are also common. The accuracy of
an MP4 calculation is roughly equivalent to the accuracy of a
CISD calculation. MP5 and higher calculations are seldom
done due to the high computational cost (N10 time
complexity or worse)
Post-HF methods
Configuration Interaction (CI)

A configuration interaction wave function is a multiple-determinant wave function. This is

constructed by starting with the HF wave function and making new determinants by promoting
electrons from the occupied to unoccupied orbitals. Configuration interaction calculations can be
very accurate, but the cost in CPU time is very high (N8 time).
Configuration interaction calculations are classified by the number of excitations used to make
each determinant. If only one electron has been moved for each determinant, it is called a
configuration interaction single-excitation (CIS) calculation. Single-and double excitation (CISD)
calculations yield a ground-state energy that has been corrected for correlation. Triple-excitation
(CISDT) and quadruple-excitation (CISDTQ) calculations are done only when very-high-accuracy
results are desired.
Post-HF methods
Multi-Configurations Self Consistent Field (MCSCF)
MCSCF calculations also use multiple determinants. However, in an MCSCF calculation the
orbitals are optimized for use with the multiple-determinant wave function. These calculations can
often give the most accurate results for a given amount of CPU time. Compared to a CI
calculation, an MCSCF gives more of the correlation energy with fewer configurations. However,
CI calculations usually give more correlation energy in total because so many more configurations
are included.

Coupled Cluster theory (CC)

Coupled cluster calculations are similar to configuration interaction calculations in that the wave
function is a linear combination of many determinants. However, the means for choosing the
determinants in a coupled cluster calculation is more complex than the choice of determinants in
a CI. Like CI, there are various orders of the CC expansion, called CCSD, CCSDT, and so on. A
calculation denoted CCSD(T) is one in which the triple excitations are included perturbatively
rather than exactly.
Solving the Schrödinger Equation

 Three parts to solve Schrödinger equation for molecules:

 Born-Oppenheimer Approximation
◼ Leads to the idea of a potential energy surface

 The expansion of the many-electron wave function in

terms of Slater determinants.
◼ Often called the “Method”

 Representation of Slater determinants by molecular

orbitals, which are linear combinations of atomic-like-
orbital functions.
◼ The basis set
 Basis set "HF/6-31G(d)"
basis functions "method/basis set"

molecular orbitals Slater type orbitals (STO) Gaussian type orbitals (GTO)

John Clarke Slater John Anthony Pople

1900-1976 1925-2004
 Basis set n-1 -ar
 c
The radial dependence of the H-atom solutions are Slater type functions: a ,n,l ,m (r,q ,j) = NYl,m (q ,j )r e
 Most electronic structure theory use Gaussian type functions because they are
computationally much more efficient. l z -ar 2
ca ,l (x,y,z) = Nx y z e
lx ly
x ,l y ,l z

 lx + ly + lz = l and determines type of orbitals (l=1 is a p...)

 ’s can be single Gaussian functions (primitives) or themselves be linear combinations of
Gaussian functions (contracted).

STO-3G basis function Slater-type vs. Gaussian-type

Computational methods #opt HF/6-311G
(que optimice)(met hartre)
Esta diapo es una pregunta del examen fijisimo
concretamente lo de los números de la esquina Basis set
6 -> cuantas ecuaciones primitivas se utilizan para describir el orbital

❑ Pople’s basis set

Minimal basis sets
Core orbitals are represented by a single orbital
Split-valence basis sets
with N primitives.
Valence orbitals are represented by 2 orbitals:
one orbital with X primitives and one diffuse
❑ Correlation consistent basis set (Dunning) orbital.

basis core valence valence

❑ Other basis set 3-21G 3 2 1

4-31G 4 3 1
6-31G 6 3 1
Computational methods
Basis set
❑ Pople basis set
Designation inner valence
STO-3G 3G 3G
3-21G 3G 2G + 1G
4-31G 4G 3G + 1G
6-31G* 6G 3G + 1G + d on heavy
6-31G** 6G 3G + 1G + d on heavy
p on H

Gaussian09
(program)
4-31G*
Computational methods
Basis set
❑ Pople basis set
Designation inner valence
STO-3G 3G 3G
3-21G 3G 2G + 1G
4-31G 4G 3G + 1G
6-31G* 6G 3G + 1G + d on heavy
6-31G** 6G 3G + 1G + d on heavy
p on H

1 + = fx difusa para los que no son H

el otro + = para los que no son C, N, O

Gaussian09
(program)
 Computational methods Basis set
Polarization functions for double zeta basis sets
To further increase the flexibility of the orbital description, double zeta basis sets are frequently
augmented with basis functions of higher angular momentum. A typical first step consists of the addition
of a set of d-type functions to the basis sets of those atoms, which have occupied s- and p-shells in
their electronic ground states. For hydrogen, this corresponds to the addition of a set of p-type
functions.
Two different notations exist to specify the addition of polarization functions:
(1) The first notation adds one asterisk to the basis set to specify addition of polarization functions to
nonhydrogen atoms, while two asterisks symbolize the addition of polarization functions to all
atoms (including hydrogen): 6-31G**
(2) In the second (preferable) notation the polarization functions are specified through their angular
quantum number explicitly. The 6-31G** basis set would then be termed "6-31G(d,p)". This latter
notation is much more flexible as multiple sets of polarization functions can be specified much more
easily.
 Computational methods Basis set
Diffuse basis functions for split valence basis sets
The theoretical description of negatively charged species is particularly challenging for ab initio MO
theory. This is due to the fact that the excess negative charge spreads outward to a much larger degree
than is typically the case for uncharged or positively charged molecules. The description of such a
diffuse charge distribution is not very well possible with the typical split valence basis sets discussed
before. Addition of very diffuse basis functions (with correspondingly small orbital exponents) cures this
problem to a certain extend as it allows the description of electron density relatively far from the
nucleus. Diffuse basis functions are typically added as an additional set of uncontracted Gaussian
functions of the same angular momentum as the valence electrons. To reflect the addition of diffuse
basis functions on all nonhydrogen atoms, a +-sign is added to the standard basis set notation. If
diffuse s-type functions are also added to the basis set of hydrogen atoms, a second +-sign is
appended. Using carbon as an example, the combination of the 3-21G basis set with one set of diffuse
sp-functions yields the improved "3-21+G" basis set:.
Computational methods
Basis set
❑ Correlation consistent basis set (Dunning)
- cc-pVNZ cc = correlation consistent
p = polarization functions added
V = Valence
N = D (doble), T (triple), …
- aug-cc-pVNZ augmented with diffuse functions
❑ Other basis set
Other split-valence basis sets often have rather generic names:
−SV(P)
−SVP
−DZV
−TZV
−TZVPP - Valence triple-zeta plus polarization
−QZVPP - Valence quadruple-zeta plus polarization

https://www.basissetexchange.org/
Computational methods

Osea difusa para H y para no H

Computational methods
Basis set

https://www.basissetexchange.org/
Computational methods Basis set Gaussian09 (program)

Stoichiometry CH4O
 Computational methods Basis set
HF/3-21G Stoichiometry CH4O
3 basis functions for core electrons C: 1s2 2s2 2p4
2 basis functions for valence electrons
1 basis functions for valence electrons
O: 1s2 2s2 2p6
H: 1s1
26 basis functions
 Computational methods Basis set
HF/6-31G(d,p)
6 basis functions for core electrons
3 basis functions for valence electrons 50 basis functions
1 basis functions for valence electrons
 Computational methods
Basis Set Superposition Error (BSSE)
The interacting energy in a complex or cluster is computed as the difference between the energy of the
complex and the total energy of the (noninteracting) monomers, which form the complex

A+B->AB ΔE=EAB-EA-EB

Such calculations are known to be sensitive to the basis set superposition error (BSSE). The error is due
to the fact that the wave function of a molecular complex is expanded in a set of basis functions that are
composed from basis functions centered on nuclear positions of the interacting molecules.
Hence, the space spanned by the basis functions depends on the actual geometry of the studied
complex, which obviously varies when its potential energy surface is scanned (inside the complex, the
basis functions of a fragment cover also the other fragment). Boys and Bernardi suggested an elegant
method, which they named the counterpoise (CP) correction, to cope with this problem. According to this
method, the individual monomers are calculated using the basis set of the complex. Since the energies of
the individual molecules usually are lower when computed within the composite basis of the interacting
molecules rather than in the monomer’s own basis, it follows that the CP corrected interaction energies
are smaller than the uncorrected ones.
Computational methods
Effective Core Potential (ECP)

In 1935 Hellmann proposed to replace the core

electrons with analytical functions that would
reasonably accurately, (and much more efficiently !)
represent the combined nuclear-electronic core to
the remaining electrons.

Advantages:
• Reduction of number of electrons Common ECPs:
• Reduction of basis set size • LANL2DZ
• SDD
• Inclusion of relativistic and other effects
Computational methods
❑ Method and Basis set should be “balanced”
❑ Small basis FCI no more accurate than large basis HF!

HF MP2 CCSD(T) CCSDT FCI

Small basis
Bad!

Good
Bad!
Complete
Basis

❑ Small basis + CCSD(T), often yields poor results

Computational methods Jacob’s ladder for DFT functionals
Step calculations
HF/6-311+G(d,p)//HF/3-21G(d)

To save calculation efforts, on can use different accuracy for optimization of

geometry and calculation of properties on the optimized geometry.
HF/3-21G(d)
means that the geometry was optimized using HF/3-21G(d) and the final
result (single point) was calculated using HF/6-311+G(d,p)
Método
HF/STO-3G + 3 2
HF/3-21G
HF/6-31G
HF/6-31G(d)
DEexp -37 kcal/mol
HF/6-31G(d,p)
HF/6-31+G(d,p)
HF/6-311G
HF/6-311G(d,p)
HF/6-311+G(d,p)
HF/6-311++G(d,p)
Método E N2 (hartree) E H2 (hartree) E NH3 (hartree) DE (kcal/mol)
HF/STO-3G
HF/3-21G
HF/6-31G
HF/6-31G(d)
HF/6-31G(d,p)
HF/6-31+G(d,p)
HF/6-311G
HF/6-311G(d,p)
HF/6-311+G(d,p)
HF/6-311++G(d,p)
MP2/6-311++G(d,p)
CCSD/6-311++G(d,p)