2. Definition of Computational Chemistry
• Computational Chemistry: Use mathematical
approximations and computer programs to obtain results
relative to chemical problems.
• Computational Quantum Chemistry: Focuses specifically
on equations and approximations derived from the
postulates of quantum mechanics. Solve the Schrödinger
equation for molecular systems.
• Ab Initio Quantum Chemistry: Uses methods that do not
include any empirical parameters or experimental data.
3. What’s it Good For?
• Computational chemistry is a rapidly growing field in chemistry.
– Computers are getting faster.
– Algorithims and programs are maturing.
• Some of the almost limitless properties that can be calculated
with computational chemistry are:
– Equilibrium and transition-state structures
– dipole and quadrapole moments and polarizabilities
– Vibrational frequencies, IR and Raman Spectra
– NMR spectra
– Electronic excitations and UV spectra
– Reaction rates and cross sections
– thermochemical data
4. Motivation
• Schrödinger Equation can only be solved exactly for simple
systems.
– Rigid Rotor, Harmonic Oscillator, Particle in a Box, Hydrogen Atom
• For more complex systems (i.e. many electron atoms/molecules)
we need to make some simplifying assumptions/approximations
and solve it numerically.
• However, it is still possible to get very accurate results (and also
get very crummy results).
– In general, the “cost” of the calculation increases with the accuracy of the
calculation and the size of the system.
5. Getting into the theory...
• Three parts to solving the 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
7. Time Independent Schrödinger Equation
• We’ll be solving the Time-Independent Schrödinger Equation
ˆ
H E
ˆ
H ˆ
T ˆ
V ˆ
H ˆ
Ekinetic ˆ
Epotential
Your book writes this as:
ˆ
H
2
2
1
m
2
2
2me
i
2
i
Z Z
e 2
r
Z
e 2
ri
i
e 2
rij
i j
j
For Many electron atoms/molecules:
Nuclei
kinetic
energy
Electron
kinetic
energy
Nuclear-
Nuclear
repulsion
Nuclear-
electron
attraction
Electron-
electron
repulsion
ˆ
T ˆ
V
i
2
2
x2 2
y2 2
z2
where:
2
2m
d
2
dx2
8. The Born-Oppenheimer Approximation
• 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
• And consequently nuclei move much more slowly than do
the electrons (E=1/2mv2
). To the electrons the nuclei
appear fixed.
• Born-Oppenheimer Approximation: to a high degree of
accuracy we can separate electron and nuclear motion:
(R,r)= el(r;R) N(R)
9. -20
-16
-12
-8
-4
0
4
8
12
3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
Electronic Schrödinger Equation
• Now we can solve the electronic part of the Schrödinger
equation separately.
• BO approximation leads
to the idea of a potential
energy surface.
Diatomic Potential Energy Surface (HgBr)
U(R)
(kcal/mol)
R (a0)
Atomic unit of length
1 bohr = 1 a0 = 0.529177 Å
U(R)
De
Re
ˆ
Helel (r;R) Eelel (r;R)
ˆ
Hel
2
2me
i
2
i
Z
e 2
ri
i
e 2
rij
i j
j
U(R) Eel VNN
VNN
Z Z
e 2
r
10. Nuclear Schrödinger Equation
• Once we have the Potential Energy Surface (PES) we can
solve the nuclear Schrödinger equation.
• Solution of the nuclear SE
allow us to determine a large
variety of molecular properties.
An example are vibrational
energy levels.
ˆ
HNN (R) ENN (R)
ˆ
HN
2
2
1
m
2
U(R)
0
10000
20000
30000
40000
50000
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Vibrational Energy Levels of HF
U(R)
(cm
-1
)
R (a0)
v=0
v=1
v=2
v=3
v=17
11. Polyatomic Potential Energy Surfaces
120
100
80
60
40
20
0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
0.030
0.060
Theta (HOCl),
degrees
r (OH),
bohr
r(OCl)=4.99 a0
10-17
10-16
10-15
10-14
10-13
10-12
0.0005 0.0015 0.0025 0.0035
Thermal Rate Coefficients
PES1
Data Points from experiment
PES2
1/T (K-1
)
k
(cm
-3
molecules
-1
)
• We can only look at cuts/slices
• 3n-6 degrees of freedom
• Minima and Transition states
• Minimum energy path
• Like following a stream-bed
O + HCl OH + Cl
13. So how do we solve Electronic S.E.?
• For systems involving more than 1 electron, still isn’t
possible to solve it exactly.
– The electron-electron interaction is the culprit
ˆ
Hel
2
2me
i
2
i
Z
e 2
ri
i
e 2
rij
i j
j
14. Approximating : The Method
• After the B-O approximation, the next important
approximation is the expansion of in a basis of Slater
determinants:
• Slater Determinant:
/ are spin-functions (spin-up/spin-down)
i are spatial functions (molecular orbitals
i and i are called spin-orbitals
– Slater determinant gives proper anti-symmetry (Pauli Principle)
el dii d00
i
d11 d22 ...
0
1
N!
1(1) 1(1) 2(1) M(1)
1(2) 1(2) 2(2) M(2)
1(N ) 1(N) 2(N ) M (N)
15. Hartree-Fock Approximation
• Think of Slater determinants as configurations.
– Ex: Neon:
• Ground-state electron configuration 1s2
2s2
2p6
this would be 0
1 might be 1s2
2s2
2p5
3s1
• If we had a complete set of i’s the expansion would be
exact (not feasible).
• Hartree-Fock (HF) Approximation: Use 1 determinant,
0.
– A variational method (energy for approximate will always be
higher than energy of the true )
– Uses self-consistent field (SCF) procedure
– Finds the optimal set of molecular orbitals for 0
– Each electron only sees average repulsion of the remaining
electrons (no instantaneous interactions).
16. Accuracy of Hartree-Fock Calculations
• Hartree-Fock wavefunctions typically recover ~99% of the total electronic energy.
– total energy of O-atom ≈ -75.00 Eh (1 Hartree= 1 Eh = 2625.5 kJ/mol).
– 1 % of total energy is 0.7500 Eh or ~1969 kJ/mol
– With more electrons this gets worse. Total energy of S atom ≈ -472.88 Eh (1%
of energy is 12415 kJ/mol)
• Fortunately for the Hartree-Fock method (and all Quantum Chemists) chemistry is
primarily interested in energy differences, not total energies. Hartree-Fock
calculations usually provide at least qualitative accuracy in this respect.
– Bond lengths, bond angles, vibrational force constants, thermochemistry, ... can
generally be predicted qualitatively with HF theory.
Re (Å) e (cm-1
) De (KJ/mol)
HF/cc-pV6Z 1.10 2427 185
Experiment 1.13 2170 260
% Error 2.7% 11.8% 28.8%
Spectroscopic Constants of CO (Total Ee≈-300,000 kJ/mol)
17. Electron Correlation
• Electron Correlation: Difference between energy calculated with exact
wave-function and energy from using Hartree-Fock wavefunction.
Ecorr = Eexact - EHF
• Accounts for the neglect of instantaneous electron-electron interactions of
Hartree-Fock method.
• In general, we get correlation energy by adding additional Slater
determinants to our expansion of .
• Hartree-Fock wavefunction is often used as our starting point.
• Additional Slater determinants are often called “excited.”
– Mental picture of orbitals and electron configurations must be abandoned.
• Different correlation methods differ in how they choose which i to include
and in how they calculate the coefficients, di.
el d0HF dii
i
1
19. Configuration Interaction
• Write as a linear combination of Slater Determinants and calculate the
expansion coeficients such that the energy is minimized.
• Makes us of the linear variational principle: no matter what wave function
is used, the energy is always equal to or greater than the true energy.
• If we include all excited i we will have a full-CI, and an exact solution for
the given basis set we are using.
• Full-CI calculations are generally not computationally feasible, so we must
truncate the number of i in some way.
• CISD: Configuration interaction with single- and double-excitations.
– Include all determinants of S- and D- type.
• MRCI: Multireference configuration interaction
• CI methods can be very accurate, but require long (and therefore
expensive) expansions.
– hundreds of thousands, millions, or more
el d0HF dii
i
1
20. Møller-Plesset Perturbation Theory
• Perturbation methods, like Møller-Plesset (MP)
perturbation theory, assume that the problem we’d like to
solve (correlated and E) differ only slightly from a
problem we’ve already solved (HF and E).
• The energy is calculated to various orders of
approximation.
– Second order MP2; Third order MP3; Fourth order MP4...
– Computational cost increases strongly with each succesive order.
– At infinite order the energy should be equal to the exact solution of
the S.E. (for the given basis set). However, there is no guarantee
the series is actually convergent.
– In general only MP2 is recommended
• MP2 ~ including all single and double excitations
21. Coupled Cluster (CC) Theory
• An exponential operator is used in constructing the
expansion of determinants.
• Leads to accurate and compact wave function expansions
yielding accurate electronic energies.
• Common Variants:
– CCSD: singles and doubles CC
– CCSD(T): CCSD with approximate treatment of triple excitations.
This method, when used with large basis sets, can generally
provide highly accurate results. With this method, it is often
possible to get thermochemistry within chemical accuracy, 1
kcal/mol (4.184 kJ/mol)
22. Frozen Core Approximation
• In general, only the valence orbitals are involved in
chemical bonding.
• The core orbitals don’t change much when atoms are
involved in molecules than when the atoms are free.
• So, most electronic structure calculations only correlate
the valence electrons. The core orbitals are kept frozen.
– i.e., 2s and 2p electrons of Oxygen would be correlated, and the 1s
electrons would not be correlated.
23. Density Functional Theory
• The methods we’ve been discussing can be grouped
together under the heading “Wavefunction methods.”
– They all calculate energies/properties by calculating/improving
upon the wavefunction.
• Density Functional Theory (DFT) instead solves for the
electron density.
– Generally computational cost is similar to the cost of HF
calculations.
– Most DFT methods involve some empirical parameterization.
– Generally lacks the systematics that characterize wavefunction
methods.
– Often the best choice when dealing with very large molecules
(proteins, large organic molecules...)
25. Basis Set Approximation: LCAO-MO
• Slater determinants are built from molecular orbitals, but how do
we define these orbitals?
• We do another expansion: Linear Combination of Atomic
Orbitals-Molecular Orbitals (LCAO-MO)
– Molecular orbital coefs, cki, determined in SCF procedure
– The basis functions, i, are atom-centered functions that mimic solutions of
the H-atom (s orbitals, p orbitals,...)
• The larger the expansion the more accurate and expensive the
calculations become.
0
1
N!
1(1) 1(1) 2(1) M(1)
1(2) 1(2) 2(2) M(2)
1(N ) 1(N) 2(N ) M (N)
i cki k
k
M
26. Gaussian Type Orbitals
• The radial dependence of the H-atom solutions are Slater type
functions:
• Most electronic structure theory calculations (what we’ve been talking
about) use Gaussian type functions because they are computationally
much more efficient.
• 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).
,n,l,m(r,,) NYl,m (,)rn 1
er
-3 -2 -1 1 2 3
0.2
0.4
0.6
0.8
1
Gaussian type function
Slater type function
,l x ,ly ,lz
(x,y,z) Nxlx
y
ly
zlz
e r 2
27. Pople-style basis sets
• Named for Prof. John Pople who won the Nobel Prize in Chemistry
for his work in quantum chemistry (1998).
• Notation: 6-31G
Use 6 primitives
contracted to a single
contracted-Gaussian
to describe inner (core)
electrons (1s in C)
Use 2 functions to
describe valence orbitals (2s, 2p in C).
One is a contracted-Gaussian
composed of 3 primitives,
the second is a single primitive.
6-311G Use 3 functions to describe valence orbitals...
6-31G* Add functions of ang. momentum type 1 greater than
occupied in bonding atoms (For N2 we’d add a d)
6-31G(d) Same as 6-31G* for 2nd and 3rd row atoms
28. Correlation-Consistent Basis Sets
• Designed such that they have the unique property of forming a
systematically convergent set.
• Calculations with a series of correlation consistent (cc) basis sets can
lead to accurate estimates of the Complete Basis Set (CBS) limit.
• Notation: cc-pVnZ
– correlation consistent polarized valence n-zeta
• n = D, T, Q, 5,... (double, triple, quadruple, quintuple, ...)
– double zeta-use 2 Gaussians to describe valence orbitals; triple zeta-use 3
Gaussians...
– aug-cc-pVnZ: add an extra diffuse function of each angular momentum
type
• Relation between Pople and cc basis sets
– cc-pVDZ ≈ 6-31G(d,p)
– cc-pVTZ ≈ 6-311G(2df,2pd)
30. -875.50
-875.45
-875.40
-875.35
-875.30
-875.25
1 2 3 4 5 6
Basis set convergence for the BrCl total energy:
CCSD(T)/aug-cc-pVnZ
Total
Energy
(Eh)
n (basis set index)
calculated
CBS (mixed)
En=ECBS + Ae-(n-1)
+ Be-(n-1)
2
31. -875.50
-875.45
-875.40
-875.35
-875.30
-875.25
1 2 3 4 5 6
Basis set convergence for the BrCl total energy:
CCSD(T)/aug-cc-pVnZ
Total
Energy
(Eh)
n (basis set index)
calculated
CBS (mixed)
CBS (1/n3
)
En=ECBS+A/n3
37. 400
410
420
430
440
450
460
1 2 3 4 5 6
Basis set convergence for the BrCl e:
CCSD(T)/aug-cc-pVnZ
e
(cm-1
)
n (basis set index)
calculated
experiment
CBS (avg)
38. Exact Solution
HF MP2 CCSD(T)
DZ
TZ
QZ
Wave Function Expansion
Basis
Set
Expansion
Full CI
HF
Limit
Complete Basis Set Limit
Typical Calculations
Basis Set
All possible
configurations
Interaction between basis set
and correlation method
require proper treatment of
both for accurate calculations.
Need to specify method and
basis set when describing a
calculation
39. Computational Cost
• Why not use best available correlation method with the
largest available basis set?
– A MP2 calculation would be 100x more expensive than HF
calculation with same basis set.
– A CCSD(T) calculation would be 104
x more expensive than HF
calculation with same basis set.
– Tripling basis set size would increase MP2 calculation 243x (35
).
– Increasing the molecule size 2x (say ethanebutane) would
increase a CCSD(T) calculation 128x (27
).
Method Scaling of Cost
HF M2
– M3
MP2 M5
CCSD M6
CCSD(T) M7
40. High accuracy possible
Molecule CBS De EZPVE ECV ESR Atomic
SO
Theoretical
D0 (0K)
Experimental
D0 (0K)
NH2 (2
B1) 181.8 -12.0 0.4 -0.2 0.0 170.0 170.0
0.3
H3CSH (1
A) 473.5 -28.6 1.5 -0.6 -0.6 445.2 445.1
SO2 (1
A1) 257.6 -3.9 1.0 -0.9 -1.0 253.7 254.0
0.2
Na2 (1
g
+
) 16.8 -0.2 0.3 0.0 0.0 16.9 16.8
0.3
BrCl (1
+
) 56.6 -0.6 0.4 -0.4 -4.4 51.6 51.5
0.3
CH3I (1
+
) 372.3 -22.4 3.7 -0.9 -7.3 345.4 344.8
0.3
Si2H6 (1
A1g) 536.1 -30.5 0.0 -1.1 -0.9 503.6 500.1
• Despite all these approximations highly accurate results
are still possible.
CCSD(T) Atomization Energies for Various Molecules
Atomization energies are notoriously difficult to calculate.
41. 0 200 400 600 800
0
10
20
30
40
Intensity
(km/mol)
Frequency (cm-1
)
Predicted IR Spectrum
HgBrO
HgBrCl
HgBr2
0.0
2.0
4.0
6.0
8.0
10.0
12.0
4.0
5.5
7.0
8.5
10.0
11.5
-60
-20
20
60
Dynamics and Spectroscopy of the reactions of Hg and Halogens
r, bohr
R, bohr
kcal/mol
r
R
44. Biochemistry applications
Laboratory of Computational Chemistry and
Biochemistry Institute of Chemical Sciences and
Engineering
Swiss Federal Institute of Technology EPF Lausanne
Group Röthlisberger
45. Get your paper and pencil ready...
• There exist a large number of software packages capable
of performing electronic structure calculations.
– MOLPRO, GAMESS, COLUMBUS, NWCHEM, MOLFDIR,
ACESII, GAUSSIAN, ...
– The different programs have various advantages and capabilities.
• In this class we will be using the Gaussian program
package.
– Broad capabilities
– Relatively easy for non-experts to get started with
– Probably most widely used
• We also have available to us Gaussview which is a GUI
that interfaces with Gaussian for aiding in building
molecules and viewing output.
46. Caution!
• Different choices of methods and basis sets can yield a
large variation in results.
• It is important to know the errors associated with and
limitations of different computational approaches.
• This is important when doing your own calculations, and
when evaluating the calculations of others.
• Don’t just accept the numbers the computer spits out at
face value!
47. Conclusion
• Born-Oppenheimer Approximation
– Separate electronic motion from nuclear motion and solve the
electronic and nuclear S.E. separately.
• Expansion of the many electron wave function: “The
Method”
– Represent wave function as linear combination of Slater
determinants.
– More Slater determinants (in principle) yield more accurate
results, but more expensive calculations.
• Expansion of molecular orbitals: “The Basis Set”
