An Introduction to Computational Chemistry
- 2 weeks
by Flick Coleman, Wellesley College,
modified by James Hamilton, UWP - Spring 2000
the rudiments of computational chemistry, an area of chemistry undergoing
a hands-on introduction to a modern computational chemistry program --
PC Spartan .
bond angles predicted by VSEPR to those obtained from calculations and
the utility of electron density models.
PC Spartan, Molecules 3D and Hyperchem and casually compare them.
either PC Spartan or Hyperchem (whichever does what you need), prepare
a report comparing your results to available spectra of some compounds
directly to tutorial
2.1 What is Computational Chemistry?8
Computational chemistry (a.k.a. molecular modeling) is the application
of computer-based models to the simulation of chemical processes and the
computation of chemical properties. It accounts for roughly a third of
the supercomputer usage worldwide. According to a recent text book, "Today,
the situation has been reached where, in many cases, the computational
chemist can substitute the computing machine for the test tube."
In this lab we will begin to scratch the surface of this enormously important
and growing field.
2.2 Why Computational Chemistry?
Computational chemistry is a valuable tool for experimental chemists
to bypass tedious, time consuming, costly, and sometimes dangerous experiments.
In the drug industry, computer design of molecules with specified properties
is now becoming more common.
A new method for synthesizing useful antibacterial agents was recently
devised using computer calculations.
Drug design is often modeled first on computers to see what molecules "fit"
in chemical docking or receptor sites.
Furthermore, computational chemistry allows one to
investigate molecules that are too unstable to be studied experimentally,
analyze quantities (such as atomic charges) that are not experimentally
observable, and rectify incorrect experimental assignments.
Based on spectroscopic experimental results, Gerhard Herzberg, a Nobel
prize winning scientist, concluded that the methylene radical (:CH2)
with two unpaired electrons had a linear geometry. Sophisticated calculations
by Bender and Schaefer, however, demonstrated that methylene was bent with
a bond angle of 135.1° . Further experiments of methylene confirmed
the latter assignment. The measured bond angle was 137.7°.
Also, computational chemistry allows one to both calculate
certain quantities (such as heats of formation) with more accuracy than
can be determined experimentally (special cases only) and improve one’s
general understanding of chemical phenomena.
Understanding how an expensive catalyst such as Pt works can lead to finding
cheaper alternatives. The rational design of catalysts is a long sought
after goal of chemists.
EXAMPLE: Chemical properties
such as solubility, enthalpies of solution and evaporation rates can now
be calculated accurately instead of estimating them from homologous series
or similar compounds.
2.3 What can be Calculated with Computational Chemistry?
Properties of stable molecules, properties of transition states, and
reaction rates can be calculated using computational chemistry methods.
Each of the above three type of calculation is described further below.
I. Molecular Properties
Selected examples of molecular properties that can be calculated:
Molecular Geometry: The arrangement of nuclei for which the potential energy
is a minimum.
Probability Density: Orbital (wavefunction) squared represents the probability
distribution function for an electron in a particular molecular orbital
Total Electron Density: Experimentally found by analyzing X-ray diffraction
data of crystals.
UV-Visible spectroscopy: Calculate the energy difference between ground
and first excited electronic state (the basic principles are similar to
the H-atom emission spectrum).
Vibrational analysis and IR spectroscopy: Vibrational transitions of molecules
occur in the infrared region of the electromagnetic spectrum.
Atomic charges: The total charge on an atom is the net excess of nuclear
charge over electronic charge. Atomic charges are difficult to compute.
Dipole moments: An electric dipole consists of two equal and opposite charges
separated by a distance. Molecules that have a permanent dipole are classified
Ionization energies: The minimum energy needed to remove the most loosely
bound electron from a molecule in the gas phase.
Electron affinity: The energy released when an electron attaches to a gas
Dissociation energies: To calculate the dissociation energy theoretically,
one subtracts the calculated electronic energy at the equilibrium geometry
from the calculated energies of the separated atoms that form the molecule.
II. Transition State Energy and Geometry Calculations:
The transition state (a.k.a. the activated complex) is a molecule (lasting
only a few molecular vibrations) that is formed by the reactant molecules
before they form the product. While the geometry of even complex stable
molecules can be obtained routinely via x-ray crystallography, detailed
structural characterization of transition states by experimentation is
III. Reaction Rates:
The computation of reaction rates (kinetics) is perhaps the most demanding
type of computational chemistry calculation. Reaction rates have been calculated
only for a limited number of reactions. Examples:
+ H2 --> H2 + H (calculations agree well with experimental data)
F + H2 --> FH + H (need
for further calculations)
2.4 Brief Introduction to the Five Major Methods
of Computational Chemistry:
1. Ab initio:
Means based on first principles.
Uses an electronic structure method based on quantum mechanics and uses
no experimental parameters in calculations.
The major disadvantage of ab initio quantum chemistry is the heavy
demands on computer power.
2. Density Functional Theory (DFT):
Uses electronic structure method based on quantum mechanics.
DFT is a less expensive alternative to ab initio calculations.
Method not available in PC Spartan Plus
This is also an electronic structure method based on quantum mechanics.
Calculations use experimental parameters to simplify the computation.
Examples of semiempirical methods available in PC Spartan Plus: AM1, PM3.
HyperChem has numerous semiempirical mMethods we will use later.
4. Molecular Mechanics (MM):
Newtonian mechanics (classical physics) is used to predict the structures
and properties of molecules. Since the theory is not based on quantum mechanics,
electrons are not treated explicitly in molecular mechanics.
Treats molecules as spheres (nuclei) connected by springs (bonds). Actually,
molecules are treated as though they are a collection of charged balls
connected by springs.
MM can be used for very large systems containing many thousands of atoms
because the above approximation markedly simplifies calculations.
Molecular mechanics methods cannot be used to address bond making or breaking
because electrons are not treated explicitly in these methods.
SYBIL is the only Molecular Mechanics method available in PC Spartan Plus.
HyperChem has numerous MM methods we will use later.
5. Molecular Dynamics (MD):
MD also uses Newtonian mechanics (classical physics).
MD makes it possible to study the dynamic behavior of a collection of molecules
as large as 1000 molecules.
MD is not available in PC Spartan Plus.
HyperChem has numerous MD methods we will use later.
3. Equipment and Materials
PC Spartan is loaded on the PC computers in Room OTTS301. If the
program does not work on a particular PC, check with your lab instructor
to ensure that the appropriate hardware lock is installed.
PC Spartan comprises six independent program modules: a graphical
user interface, three modules for geometry optimization calculations, a
module for the calculation of molecular properties, and a module for computing
graphical surfaces. Without leaving the graphical interface, the user can
construct a complex molecular structure, refine its geometry, specify a
task and level of calculation, designate any graphical surfaces for later
display, and then submit for calculation.
This tutorial describes the use of PC Spartan, a general-purpose molecular
modeling program sold by Wavefunction, Inc. (Irvine, California). Spartan
is an unusual molecular modeling packagein that it is very user-friendly
and can be used by students with very little training in chemistry or molecular
modeling. In addition, it is a fully functional research tool, and most
Spartan users are research chemists, although they usually use a $3k per
copy version Spartan (Version 5.0).
A Spartan Modeling Session
The Spartan software consists of a suite of computer programs. These
programs perform different functions and include a "Builder" program for
building and modifying models, several programs for calculating molecular
structure and energy, a "Properties" program for deriving molecular properties
from previously calculated structures, and a "Graphics" program which can
be used to display properties using three-dimensional graphs. All of the
programs communicate with one another automatically as needed, therefore,
you do not need to concern yourself with their operations. From your point
of view, it will always seem as though you are using a single program:
Although Spartan's large assortment of modeling tools may seem a bit
intimidating, there is a natural relationship between them, and this can
be seen by considering a typical modeling session (Fig. 1a). The starting
point for most molecular modeling is the user's need to learn something
about a molecule. Thus, the user typically 1)builds a trial model of the
molecule (Builder), 2) optimizes the molecule's structure and calculates
its energy (Setup ...), 3) examines the results to see if they are reasonable
(Display), and then repeats one of the earlier steps with a new model if
needed, or proceeds to calculate interesting properties (Properties/Graphics.)
Each tool, therefore, plays a logical role in the modeling process, and
one quickly becomes familiar with each tool after just a small amount of
Figure 1a. Modeling Session Flow-chart
4. PCSpartan Plus Tutorial
4.1.1. General Operating Features of PC Spartan
To launch PC Spartan, double click on the PC Spartan icon in the
desktop. If the icon is not present, click on Start on the taskbar
of Windows 95 located on the lower left of the screen. Highlight Programs,
then click on PC Spartan . PC Spartan’s window will appear with
its menu bar at the top of the screen. Maximize the window. Program
functions may be accessed using pull-down menus under the headings in the
menu bar. Entries under the individual menu items control all of the program’s
functions. Two or more molecules can be simultaneously displayed in the
window. However, only one molecule may be selected at a time. The selected
molecule has access to all capabilities while non-selected molecules may
only be displayed as static images.
4.1.2. Learning How to Use PC Spartan
We will use acetonitrile to illustrate the basics of PC Spartan .
Before beginning the exercise, create a new folder on the c: drive and
give it your last name. To do so, open Windows Explorer, click on
new, and finally folder. Next, type in your
last name. At the end of the exercise, you will delete this folder.
Based on your understanding of Lewis structures, draw the structure of
In steps 2,3, and 4 we will build (draw) the acetonitrile
Back in PC Spartan , select New from the File menu to enter
the build mode. The entry model kit appears at the right of the screen.
Among the simple building blocks incorporated
into PC Spartan’s entry model kit are "atomic fragments." These fragments
specify the atom type (e.g., carbon) and environment (e.g., tetrahedral
– a carbon atom attached to four other atoms). Relatively few fragments
allow construction of a wide variety of simple molecules. In addition to
atomic fragments (e.g., ),
the entry model kit also contains libraries of functional groups (e.g.
amide) and rings (e.g. phenyl), and additional tools for molecule building.
Click (left mouse button) on tetrahedral carbon ()
from the library of fragments. Bring the cursor anywhere on the blue part
of the screen and click. Tetrahedral carbon with its four free valences
indicated by "yellow vectors" appears.
The entry model kit connects atomic fragments (as
well as groups and rings) through free valences. Valence electrons are
outer electrons of an atom which are involved in chemical bonding. Any
free valences remaining will automatically be converted to hydrogen atoms
upon exiting build mode and will revert to free valences upon reentering
Click on the linear carbon
from the model kit, then click on the tip of one of the free valences
of the tetrahedral carbon atom on screen. The linear carbon is automatically
connected to the tetrahedral carbon by a single bond. To obtain a better
view rotate the structure by moving the mouse while pressing the left button.
. If you make a mistake, use the Delete Atom button to remove atoms
PC Spartan’s entry model kit connects atoms only
through the same type of free valence, e.g., single to single, double to
Click on the triple bond containing nitrogen
from the model kit, then click on the triple free valence of the
linear carbon on the screen. This connects carbon and nitrogen by a triple
In this step, a
preliminary structure refinement of the initial molecular geometry will
be performed. First, rotate the structure to obtain a better view. Next,
click on Minimize in the model kit to perform a crude minimization
of potential energy to obtain a better geometry. At the equilibrium geometry
of the molecule, the slope (gradient) of the potential energy should be
zero. The iterations (number of cycles of calculations), the molecular
mechanics (one of the computational chemistry methods discussed previously)
strain energy (-0.183 kcal/mol), the gradient, as well as information about
the symmetry of the molecule appear below the menu bar. Click on
Done to remove them.
In this step you will perform several operations (e.g., rotations) on the
acetonitrile molecule. Select View from the Build menu. This
removes the model kit from the screen and replaces the ball-and-wire model
of acetonitrile with a tube model. This model can be manipulated using
the mouse in conjunction with the keyboard. Functions associated with the
two-button mouse are summarized in the following table:
left mouse button
picking, X/Y rotate
global X/Y rotate
global X/Y translate
Control + Shift
global Z rotate
The left mouse button is used for both picking
(of graphical objects and/or of menu items), and for rotation of objects,
and the right mouse button is used for translation or scaling of objects.
Rotation and translation functions may be modified by holding down specific
Control, Alt) in addition to the appropriate
mouse buttons. With no keys depressed, the left mouse button gives rise
to rotation about the X and Y (screen) axes; the right mouse button gives
rise to translation in the X and Y (screen) directions. Together with the
Shift key, the left mouse button gives rise to rotation about the
Z direction and the right mouse button gives rise to scaling. The Control
key (Ctrl) in conjunction with the left or right mouse buttons
and (optionally) the Shift key signifies that rotations or translations
are to be carried out globally, i.e., on all molecules presently displayed,
rather than only on the "selected" molecule. In build mode only, use of
the Alt key in addition to the left mouse button allows for rotation
about a "selected" bond.
To picture the acetonitrile structure in various renditions, try several
options in the Model menu.
Labeling the atoms can be useful to look at the molecule’s structure. To
do this, select Ball and Wire from the Model menu. This model,
along with the wire model, are the only models for which the atom numbers
may be displayed. Select Labels from the Model menu. Numbers
appear next to the individual atoms together with atomic symbols, and a
check mark appears to the left of Labels in the Model menu.
The numbering is somewhat arbitrary and helps distinguish between atoms
of the same type. Remove the atom labels by again selecting Labels from
the Model menu; the check mark disappears.
Next, the molecule’s geometry will be examined through bond distances,
bond angles, and dihedral angles . Select Distance from the Geometry
menu. The model displayed is replaced by a ball-and-wire model and a message
appears below the menu bar (Distance: Select 2 atoms, a bond, or a distance
constraint). Clicking on two atoms results in each being marked
in a gold sphere and the distance between the two being displayed below
the menu bar. When the two atoms are bonded, an alternative is to click
on the bond. Another distance maybe obtained by selecting another pair
of atoms (or another bond), and so forth. When you are finished, click
To obtain a more accurate structure, a geometry optimization calculation
will be performed on acetonitrile using an ab initio method (one
of the five computational chemistry methods discussed previously). Select
Calculation from the Setup menu. Perform the following operations:
Optimization from the Task pop-up menu. This specifies optimization
of equilibrium geometry.
from the Level pop-up menu. This specifies a type of ab initio
calculation for performing the computation.
that Charge is 0 (the molecule is neutral) and Multiplicity
is 1 (these are the defaults). You can modify the contents of these
boxes, if necessary, for other molecules.
Give the molecule a name in the box to the right of Title.
on OK to exit the dialog.
To perform the actual calculation, select Submit from the Setup
menu. In the dialog which appears, you need to specify your last name (the
newly created folder) next to Save in:, supply a name for your job
in the text box File Name, and then click on Save.
(You must save or else the calculation cannot be performed.) A new
dialog appears telling you that your calculation has started. Click
on OK to remove the dialog box . When the calculation is done after
a few minutes, you will be notified with a pop-up menu. Click on
OK to remove the dialog box.
Next, we will examine the results of your calculations. Select Output
Display menu. The output for the calculation will appear
at the left of the screen. Information inside the viewing area may be scrolled
in the usual manner. At the "top" of the output is the title you have provided,
the task at hand (‘geometry optimization"), the level of calculation ("3-21G(*)),
as well as other relevant information, such as the initial set of atomic
coordinates. The next series of lines that appear tell the history of the
optimization process. Each line provides results for a particular geometry;
"Energy =" gives the energy (in hartrees; 1 hartree = 627.5 kcal/mol) for
this geometry and "rmsG ="gives the root-mean-square gradient. Ideally,
the energy will approach a minimum value for an optimized geometry and
rmsG will approach zero. If the geometry was not optimized satisfactorily,
an error message, such as "Optimization has exceeded N steps - Stop," will
be displayed following the last optimization cycle. Next, the new atomic
coordinates for the geometry optimized molecule is displayed.
At the end of the output
is a listing of atomic charges and the calculated dipole moment (the measure
of the extent to which a separation exists between the centers of positive
and negative charge within a molecule). Because charge distributions provide
information about a molecule's reactivity, chemists are very interested
in obtaining charge distributions via electronic structure calculations.
Unfortunately, there are several methods of calculating atomic charges
(and dipole moments) and different methods yield different results! PC
Spartan Plus uses two methods for calculating atomic charges (and dipole
to Electrostatic Potentials
output may be removed from view by reselecting Output from the Display
The dipole moment is also available from the Display menu under
Properties. Select Dipole from the sub-menu which appears.
The dipole moment appears below the menu bar. In addition, the dipole moment
vector is displayed on the screen attached to a ball and wire model. Click
on Done to remove both. Remember that the dipole moment vector is
defined to point from the positive charge to the negative charge (physics
uses the opposite convention).
Next, we will calculate the isoelectron density surface of the molecule
which serves to locate chemical bonds, indicate overall molecular shape
and size, and illustrate the electrostatic potential (the energy of interaction
of a point positive charge with the nuclei and electrons of a molecule).
Select Surfaces from the Setup menu. Select density from
the Surface pop-up menu, and elpot from the Property
pop-up menu. This specifies calculation of an isoelectron density surface
onto which the value of electrostatic potential has been mapped. This will
be referred to as an "electrostatic potential map." Click on Add;
the line "surface=density property=elpot pending" appears in the box at
the top of the dialog. Click on OK to close the dialog.
Select Submit from the Setup menu in order to submit the
job. Handle the various message boxes as you did during the geometry optimization.
When the calculation has completed, select Surfaces from the Display
menu. Click on the text string "surface=density…" at the top of
the dialog to highlight it. Click on the button to the left of Display
Surface and then select any of the various display styles. Click
on OK to exit the dialog. Repeat the last few instructions to change
the display surface.
Two or more molecules, or two or more copies of
one molecule, may be simultaneously displayed although only one molecule
may be selected. The selected molecule has access to all program capabilities
while non-selected molecules may only be displayed as static images. Selection
of one of the molecules currently on screen occurs by clicking on
A revealing comparison follows by displaying
a transparent density surface along with a space-filling model on the same
structure. Reenter the Surfaces dialog under the Display
menu, highlight the line "surface=density…" and then select Transparent
Mesh) from among the style options. Click on OK.
Replace whatever model you are using by a space-filling model. You can
now see clearly that the two images are nearly identical.
In this step we will examine the electrostatic potential map you calculated
in steps 14 and 15. Return to the tube model and then reenter the Surfaces
dialog in the Display menu. Highlight the line, then select Solid
from the style choices, and finally click on the button to the left
of Map Property. This requests display of the electrostatic potential
map. Exit the dialog by clicking OK. Examine the graphic
which appears on the screen. The shape of the surface is the same as before,
and the color indicates values of the electrostatic potential evaluated
on this surface. The potential is coded according to the visible spectrum
(red<orange< yellow<green <blue), with red corresponding to
the most negative potentials and blue corresponding to the most positive
potentials. Nitrogen will appear negative (electron rich) and the hydrogens
will appear positive (electron poor). This is reasonable as nitrogen is
the most electronegative atom in the molecule and the hydrogens are the
most electropositive. Electronegativity is the measure of the electron-attracting
power of a bonded atom. The electrostatic
potential map suggests that the
dipole moment which you calculated earlier has the hydrogens at the positive
end and the nitrogen at the negative end. Go back and examine the dipole
moment vector again to see if this is indeed the case.
Remove acetonitrile from the screen (choose Close from the File
Close PC Spartan by clicking on the most upper right hand box with
an "x" in it in the window.
Now, for the Pchem lab, construct benzene, napthalene or anthracene
and look at the electron densities of the HOMO and LUMO's. Show the
dipole moments and print the nice color pictures of the electronic ground
state and excited states to take home and put on your refrigerator.
Calculate the vibrations and then look at the frequncies as they vibrate.
Have fun! Use Hyperchem to look at the IR absorption and UV/VIS spectrum.
At the end of the second week, write an abstract and summarize results,
comparing to known UV/VIS absorptiona dn IR Absorption spectra.
to top of tutorial