Labs 1/2: Getting Started.: (3.0) - Version 5.0 Has Difficulties With Visualizing Orbitals in The Friezen Lab. The Older
Labs 1/2: Getting Started.: (3.0) - Version 5.0 Has Difficulties With Visualizing Orbitals in The Friezen Lab. The Older
In this set of labs we will quickly go through the most widely used features of
calculations using gaussview/gaussian09. I will not give you very detailed instructions.
You will have to find your way using the graphical interface which is quite
straightforward to use. I encourage you to draw some molecules of your own, and see if
you can manipulate the various tools in Gaussview. Below I will give you some
instructive examples that illustrate many of the capabilities of Gaussview/Gaussian09. In
lab don't hesitate to ask questions.
In this lab we will be visualizing orbitals. We should use the old version of Gaussview,
(3.0). Version 5.0 has difficulties with visualizing orbitals in the Friezen lab. The older
3.0 version of Gaussview pops up a window when the calculation is done, which might
be hidden under other pallets. Make sure you keep track of this window, and hit the OK
button. Otherwise you will sit there for ever, waiting for things to happen.
Let us run through a very simple example first. You can create the water molecule using
the builder (pull down the menu), selecting Oxygen from the ‘atoms’ menu. Let us first
do a calculation on the guessed structure (use the broom ‘clean’ to finalize the guess).
Measure the bond angles and distance first (using inquire, ?). Let us do a Hartree-Fock /
6-31G(d,p) calculation. For this click Calculate and select Jobtype=Energy,
Method=Hartree Fock, Basis is 6-31G, and on the right hand options select from the
between brackets options (d,p), by pulling down the appropriate menus. This specifies the
basis set, of which more later. Then Submit the calculation (see lower-left button). You
will have to save your file, for example as h2o. Follow the directions (and have some
patience). In a little while you get back the result, and will be asked if you want to see it.
Look at results (which are in the file h2o.chk or h2o.log) and look under summary. Here
the main specifics of the calculation are summarized and you can find the total energy.
Next we will optimize the geometry, selecting jobtype=Optimization. Measure the
geometries again. You should find the results as listed in the Table below.
1
At the optimized geometry we can now calculate vibrational frequencies. Start from the
geometry in the h2o.chk file and select frequency under calculate->Jobtype.
After the result returns you can view the vibrational frequencies (select under results).
You can view normal modes and infrared and raman spectra. The frequencies for the
normal modes in water are as follows.
This same calculation can be done in one shot. Start anew by rebuilding H2O from
scratch, and select Opt+Freq under the Energy menu. It first does the optimization, and
then calculates the vibrational frequencies. Vibrational frequencies are only meaningful
at the optimized geometry!
Again it is useful to look in the output file. The vibrational frequencies are all listed and
so is their symmetry (open in vi and go to the end, then walk back in the file). Also the
vibrational motions have a definite symmetry with respect to planes of symmetry in the
molecule. Normal modes of A1 symmetry preserve the symmetry of the molecule.
Normal modes of other symmetry-type break the symmetry of the molecule.
I will run you through one more example doing similar things. Part of the information is
similar as for water.
Use the gv builder to construct trans butadiene. You only need to build the carbon chain
using the appropriate type of carbon atom. The hydrogens are added automatically. Go to
View to examine different ways to see the structure. In particular switch on Symbols to
see all atom types. Rotate and translate the molecule, using the mouse. If you press the
clean button in the builder window, you will get a first rough guess for the structure. It is
also usefule to click the symmetrize button. Your molecule should have C2h symmetry.
Save your input file, e.g. t_c4h6.
H H
H c3 c4
H
c1 c2
H
H
2
Next optimize the structure at the AM1 level of theory. To this end open the calculate
menu. Specify Optimize in the jobtype section, and Semi-empirical under Method (pull
down the Hartree-Fock menu to find this option). You may have a look at the other
options in the calculation menu, such as charge, spin, and so forth. We will use them
shortly. Also specify a descriptive title for your job. Organization is a major part of doing
(lots of) calculations. Therefore you should create a good structure of
directories/filenames and titles that give further specifics. Input files are always indicated
filename.com. The corresponding output file automatically gets the name filename.log,
while additional data is located on filename.chk. The calculation will be run if you click
the submit button. Simply answer to the queries. Let us analyse the results of this
calculation.
Click on the inquire button in the builder menu (question mark). This will give you
geometrical information. By clicking two atoms you get the bond distance. By selecting
three atoms you find the angle made by these atoms. Finally by clicking 4 atoms in
sequence you get the dihedral angle (in "4321" format). This is the torsional angle around
atoms 2 and 3. The dihedral angle is measured rotating the fourth atom onto the first atom
along the 3-2 bond axis. Clockwise angles are positive. Now find the C1-C2, C2-C3 and
C3-C4 bond lengths. You clearly see the double bonds in the geometrical structure.
However also C2-C3 is substantially shorter than typical C-C single bonds (quickly
construct and optimize ethane if you wish to confirm this). Compare terminal and center
CH bond lengths. Also measure various bond angles and comment on your findings.
What is the C1C2C3C4 torsional angle around the C2-C3 bond? Now measure it by
clicking the appropriate atoms. It is of interest to look at the geometry in the original
guessed structure, before optimization. Here you see there is no information on
conjugation. Instead the program guesses standard single and double bonds.
Under the Results menu you will find a number of options that allow you to analyse the
output of your calculation in an easy way. First open the output file by clicking on the
appropriate filename.chk. The so-called checkpoint file contains the most extensive set of
output data in a compressed format.
- Changing geometries.
Let us transform trans-butadiene into cis-butadiene. To this end click on dihedral in the
builder menu, and select four atoms that specify the desired dihedral angle (in '4321'
format). This opens the dihedral smartslide. Change the dihedral from 180˚ to 0˚. Save
this new structure as cis-c4h6.com and perform an optimization of the geometry. Now
you can analyse the structure (and energetic and orbitals) of cis-butadiene. You can
similarly change angles and bond distances by clicking the respective buttons in
Gaussview. What is the symmetry of Cis-butadiene?
3
- Calculating forces (do it later, or in your own time)
We can get a good idea of the strength of a bond or the stiffness w.r.t. rotation by making
a small change in one of the geometrical parameters and doing a single point energy
calculation. Do not select OPT therefore, but use the energy option (default) in the
calculation menu. Performing an optimization would bring the structure back to the
original. Please note that the starting geometry is a stationary point so the first derivatives
of the energy are equal to zero. We will be looking at the second derivatives therefore by
∂2 E
making a small distortion. You should verify that converges to a definite value
∂X 2
∂E
while depends strongly on the precise small distortion. Reopen your trans-butadiene
∂X
file and fill in the following table by performing the appropriate geometrical distortions
(always restarting from the original optimized structure). This will further characterize
the single vs. double bond nature of the C1-C2 and C2-C3 bonds. In the table below I
will use X for a general distortion, R for bondlength (in Å), D for dihedral (in ˚) and A for
angle (not used here, but for future reference). At the minimum energy geometry we can
represent the energy due to a distortion as a parabola
1 E ( x) − E (0)
E ( x) = E (0) + kx 2 → k = 2 .
2 x2
A more accurate formula would be to use so-called double sided numerical derivatives
E ( x) + E (− x) − 2 E (0)
x2
Use the default rotate-group such that you are only changing the dihedral of interest,
nothing else.
Discuss your results and confirm the character of the C1-C2 and C2-C3 bonds in trans-
butadiene. You might make a similar table for cis-butadiene, and it would also be of
interest to compare to the clearly single bond in ethane (Optimize the eclipsed geometry
of ethane at the same level of calculation, and then evaluate the analogous force
constants).
4
- Vibrational frequencies and viewing normal modes.
It is very easy to obtain vibrational frequencies in gaussview. Under the calculation menu
specify OPT+FREQ. Infrared intensities are calculated always, Raman intensities
sometimes by default. Calculate the vibrational spectrum of trans-butadiene. Upon
completion of the calculation you can display by going to Results/Vibrations. This shows
a list of the calculated vibrational frequencies. By clicking any one of the frequencies and
clicking start you can view an animation of the vibrations. Optionally you can also
display the displacement vectors that characterize the normal mode. Set the Frames/cycle
and Displacement scales to your preference. By clicking spectrum you see the complete
vibrational spectrum displayed. Characterize the various vibrations in trans-butadiene.
Where are the CH stretches and how many are there? What is the difference between
them? Where are the various C-C stretches? Can you clearly distinguish C1-C2 stretch
and C2-C3 stretches? Characterize the various bending modes, or rotations around the
various bonds.
- Charges.
For any calculation you can view the charges in the molecule. Trans-butadiene is not
particularly interesting of course. You might create an alcohol or a carbonyl compound to
see something more interesting. Play around with the options to obtain a rendering of the
charges that you prefer.
The above gives you an overview of some of the most widely applicable options in
Gaussview/Gaussian03. Let me provide you with some more illustrative examples and
exercises with this set of labs.
C. Investigating thermochemistry.
Let us build ethylene and optimize geometry and calculate frequencies at the Hartree-
Fock 3-21G level. Now we also have all the information to extract thermochemical data.
If we open the output file (view file under results), and scroll to the end and then move
upwards, we find a thermo chemistry section. It lists rotational constants (from the
optimized geometry), and then has all the information to calculate the partition function
(statistical mechanics). You can see info on entropy, Gibbs free energy (G), enthalpy (H),
heat capacity (Cv).
Note: Hartree-Fock 3-21 G is useless for thermochemistry. One needs to do a much more
accurate calculation, for example B3LYP (DFT) with a cc-PVTZ basis set. This takes a
longer time to calculate. But in this way you could calculate the enthalpy of reaction of
for example the reaction
C2 H 2 ( g ) + H 2 ( g ) → C 2 H 4 ( g )
5
All gas phase reactions can in principle be studied (given enough computer resources).
We have discussed all of the underlying theoretical principles in chem254
(thermodynamics), chem350 (kinetics and statmech) and chem356 (Quantum
Mechanics): particle in the box: translational energy; Rigid Rotor model, (using spherical
harmonics): rotational energy contributions; harmonic oscillator models for polyatomics:
vibrational part of the energy. The major component of the reaction enthalpy is the
difference in the electronic energy at the minimum of the potential (the optimized
geometry), for each of the molecules involved in the reaction. Here accurate calculations
are critical, and in prior classes we only scratched the surface regarding electronic energy
calculations. I will give some more background on this in this class on computational
chemistry (chem440), but it is a highly specialized subject.
To view orbitals with Gaussian in the Friezen lab we need to use an older version of
Gaussian. Using Gaussview version 3.0 (For now. I’ll try to fix it). Let us build the trans-
butadiene molecule, and optimize the geometry at the Hartree-Fock 3-21G level.. Once
this is done, click on the orbital icon. This lists all the orbitals and their symmetries. We
will be looking at the π -orbitals in particular, and they are around the HOMO and
LUMO levels (if we use the 3-21G basis set): orbitals of a’’ symmetry, #12-17 or
thereabouts. We need to first generate the MO’s: click on New in the orbital window, and
then click generate where it ask for checkpoint file. Once this calculation is done (click
OK on the (hidden) Gaussian pallet that says it is done), we click on visualize, and we
can select the orbitals we wish to look at. Type for example 13-17 in the window where
you can specify which orbitals to look at. Then click update. If instead you click the very
tempting OK button, you can start over again! I like to look at orbitals using the mesh
view (right click in the orbital window, and select surfaces). Then you can see both the
orbital and the molecular framework.
Sometimes, I encountered a problem with this. In that case after generate you have to
upload the .chk file yourself, before you visualize. Also hit the OK button that is hidden
behind all your windows (and scream!). I hope it works …
Excersise: Identify the π -orbitals in trans-butadiene. You can increase the isosurface
value a bit (to 0.1 for example), to get a clearer view of the orbital coefficients
(regenerate the orbitals, using update). These coefficients are not all equal. See for
example McQuarrie for the precise values of the coefficients in Hückel theory (Chapter
11).
Excersise: Calculate the molecular orbitals for water. What do you anticipate the orbitals
to look like? Two lone pairs on the oxygen atom, responsible for the hydrogen bonds in
liquid water?
Answer: Look and see for yourself. The truth is that molecular orbitals are not unique.
The only thing that matters is the Slater determinant they create. One can make linear
combinations of the canonical Hartree-Fock orbitals and make more localized orbitals, or
6
the familiar hybrid orbitals from VSEPR. You can play with the NBO option in Gaussian
to see this in action. I didn’t try it yet myself.
E. Singlet and triplet states, Jahn-Teller distortions for a Hückel type problem.
A case study.
The following problem is a bit extensive. Let’s see if we can get through it in class.
I want to create (CH2)3C, which is planar and has the maximum D3h symmetry. The first
hurdle is to create this structure. Pay attention! The central atom is like an aromatic
carbon. I started by making the structure (CH3)C(CH2)2, (using a center aromatic C, two
sp2 carbons and one sp3 carbon) then delete a H-atom (delete atom button). Next clean up
the structure (use the broom icon). Then select the dihedral angle tool to rotate the two
out of plane hydrogens in the plane of the molecule (select HCCC dihedral angle). This
should look more or less right. Then go under edit and select point group. If you use the
loose or very loose criterion, you will see the molecule is recognized as having D3h
symmetry. Click the symmetrize button on the right. Finally, you can click the reconnect
button, which draws aromatic bonds for visual pleasures. Pooh! We have our molecule.
There might be easier ways to get there!
Actually there is: If you get stuck creating this structure (or even before that for the sake
of time) go to the chem356 website and copy the input file I created for this molecule.
Right click on the Gaussian input file C_CH2_3_triplet.gjf and use save target as to save
this file as a Gaussian input file in your directory. Do not open the file directly (it will
launch the newer bad version of gaussview), but just save it. The open the file from the
Gaussview that is already open. You can visualize the input file, and do the rest of the
calculations starting from this starting point. It has the desired D3h symmetry.
Now: use the symmetric D3h structure and do a geometry and frequency calculation for
the triplet state. This has to be selected under the method button in calculate. You will see
that the triplet has D3h symmetry, as predicted by Hückel theory. Let us look at the π -
orbitals for the optimized structure, using the orbital viewer. You can see that they nicely
match the orbitals predicted by Hückel. Look at both the occupied and virtual orbitals.
Notice that the HOMO is degenerate, and both are occupied by a single alpha spin
electron. This is the hallmark of the triplet state. Also notice that the alpha and beta
orbitals have different energies (and occupation patterns).
Now we will calculate the singlet state. The molecule will distort. By enlarging one of the
CCC bond angles to say 130 degrees (use the angle button to do this), we will create a
structure in which the ‘anti-bonding’ π -orbital is doubly occupied. Optimize this
structure, and investigate the changes in structure (use the measuring tool). Next, starting
again from the triplet D3h geometry, we can make one of the CCC angles sharper (e.g.
105 degrees). This disfavors the ‘anti-bonding’ π -orbital, which will now become
empty. Instead the other orbital, which was degenerate in the triplet will become doubly
occupied. Again we can optimize the structure and find a distorted geometry, now with
one sharp (acute) and two obtuse angles (> 120 degrees). These are stable structures for
the singlet state. There are really two (or even three) singlet states involved, which are all
7
very close together in energy. This kind of molecule is called a biradical. While we get
qualitatively correct structures, the energetics at the Hartree-Fock level is probably very
poor. These states are hard to calculate as there is strong mixing of configurations. You
can anticipate this by looking at the orbital energies. Also, by slight geometrical changes
the occupation pattern of the orbitals changes completely. This is a complicated system!
How complicated? Imagine that for the molecule you can distort either one of the three
equivalent angles to become less than or greater than 120 degrees. Each of these would
yield a different minimum (observable experimentally if you would replace some of the
hydrogens by deuteriums, such that the triplet structure is no longer D3h symmetry,
although the electrons are oblivious to this. In summary the potential energy surface for
the singlet state has a bunch of local minima (6), and the occupied HF orbitals rapidly
change with geometry.
In general the same is true for low spin transition metal compounds that show
Jahn-Teller distortion. There are always very low lying excited states, and many
structures related by symmetry. The low spin surfaces are complicated.
Exercises:
2. Charge and spin state. To illustrate the options first consider BH3. Optimize the
neutral species (HF/6-31G(d,p) ) and characterize the LUMO by examining the
appropriate orbital. How would you establish that BH3 is planar? Now put an extra
electron in the system by changing the charge to -1 and the spin to be a doublet in the
calculation menu. Specify the calculation to be unrestricted (different orbitals for
different spins). Now optimize the geometry of the anion. If you view the orbitals you
find there is one more α than β orbital. Why do we refer to the present molecular state as
8
a doublet? View the HOMO of α spin. How does this relate to the LUMO of BH3.
Discuss the result. We can do some more comparions. Also construct NH3, and the
corresponding cation NH3+. Perform optimizations of both species (specify the proper
charge and spin!) and analyse the relevant orbitals. Fill in the following table and discuss
your results.
BH3
BH3-
NH3+
NH3
3. More vibrations. Consider the following molecule and find its most stable structure,
using the semi-empirical AM1 level of theory, and paying particular attention to the
H3C
CH3
H3C
O
rotation of the CH3 groups. This is primarily an exercise in creating a complicated
molecule. Start from the appropriate ring structure, and then modify the molecule until
you have the proper structure. If you have found the minimum energy structure calculate
the vibrational spectrum and analyse the normal modes. Identitify the carbonyl stretch
and discuss the various CH3 rotational motions. Also identify the double bond CC stretch
mode. List the most characteristic frequencies and the total energy in a table. This is the
easiest way for me to check if you are on the right track!