Andrew Gassner and Nick Cirricione

The geometries, highest occupied molecular orbital, lowest unoccupied molecular orbital, electrostatic potentials, dipole moments, partial atomic charges and vibrational energies were calculated for Nitric Oxide, Difluoromethane, and Bromobenzene using different levels of molecular orbital theory. They included 3-21G, 6-21G, 6-31G, and DZV. The best level of theory to determine these characteristics for the molecules was not always the same as geometry was best found by using 6-31G theory, vibrational frequency DZV and dipole moments by AM1 theory.

There are different ways to calculate the expectation value of the energy depending on the amount of integrals a level of theory uses. Some use empirical data to provide estimates of the values for two electron overlap integrals needed for calculating the expectation value of the Hamiltonian. These methods include AM1 and PM3 and are some of the quickest because they just use these two. The best level of theory is Ab initio in which all integrals are calculated. The difference between these is the size of the basis sets, number of trial wavefunctions, used to determine the energy. These methods in order of increasing basis set size are 3-21G, 6-21G, 6-31G, and DZV.

In this experiment the program Gamess was used to calculate the molecular orbitals of nitric oxide, bromobenzene, and difluoromethane. The first step used in performing the calculations was to find the geometries of the molecules. The initial guess of the geometries for Nitric Oxide, Difluoromethane, and Bromobenzene was generated using the software program Avogadro. Using the initial geometries generated by Avogadro an AM1.inp file (to optimize geometry) was generated using the software program macmolplt, which further refined the geometries. The .inp file was then run using GamessQ, the interface for Gamess at the lowest level of molecular orbital calculations, AM1. The AM1.log file obtained from GamessQ was then used to generate a 6-21G.inp file, which in turn was used to generate a 6-31G.inp file followed by a DZV.inp file. The exception to this process was for bromobenzene where a 3-21G file was used in place of the 6-21G. Jmol, modeling software, was then used to visualize various aspects of these molecule for each level of theory as well as to display some physical constants for each molecule that were calculated by Gamess.

Use the following hyperlinks to see the calculated values:

Conclusion

The calculated dipole moments for NO and difluoromethane were useful and had percentage errors of 10.6 and 3.6% respectively. The interesting part is that the best value for NO used DZV and difluoromethane used AM1. For bromobenze an error of 97.0% was obtained which calls into question the usefulness of using these theories to calculate dipoles.

For vibrational energies the DZV theory was used to calculate at which wavelengths the vibrational levels would absorb energy. When compared to a IR spectrum for the molecules the vibrational wavelengths found matched up well with the peaks in the spectrum. This can be seen on the individual pages.

The different levels of theory were sometimes useful and sometimes not depending on the value being calculated. DZV would have been expected to be the most accurate level as it contained the largest basis set but sometimes was not. Because of the inconsistency of the data sets it is beneficial to generate data from several levels of theory however it may not be wise to base additional calculations off any of these levels of theory.

References

2.

3.

5. Benzene

6. Dimitriu, Mihaela; Ivan, Liliana-Mihaela; Dorohoi, Dana Ortansa.

7. McClellan, A. L.