Computational chemists are concerned about two aspects when choosing between the myriad of theoretical methodologies: the accuracy (the
"truth") and the computational cost (the tractability). Among the least expensive methods are the Hartree-Fock (HF), density functional theory (DFT), and second-order Moller-Plesset perturbation theory (MP2) methods. While each of these methods yield excellent results in many
cases, the inadequate inclusion of certain types of electron correlation (either high-orders or nondynamical) can produce erroneous results.
The compromise for the computation of noncovalent interactions often comes from empirically scaling DFT and/or MP2 methods to fit benchmark
data sets. The DFT method with an empirically fit dispersion term (DFT-D) often yields semi-quantitative results. The spin-component
scaled MP2 (SCS-MP2) method parameterizes the same- and opposite-spin correlation energies and often yields less than 20% error for prototype
noncovalent systems compared to chemically accurate CCSD(T) results. There is no simple fix for cases with a large degree of nondynamical
correlation (such as transition metal-salen complexes). While testing standard and new DFT functionals on the spin-state energy gaps of
transition metal-salen complexes, no DFT method produced reliable results compared to very robust CASPT3 results. Therefore each metal-salen
complex must be evaluated on a case-by-case basis to determine which methods are the most reliable. Utilizing a combination of DFT-D and SCS-MP2 methods, the reaction mechanism for the addition of cyanide to unsaturated imides catalyzed by the Al(Cl)-salen complex was performed. Various experimental observations are rationalized through this mechanism.
Identifer | oai:union.ndltd.org:GATECH/oai:smartech.gatech.edu:1853/34764 |
Date | 01 July 2010 |
Creators | Takatani, Tait |
Publisher | Georgia Institute of Technology |
Source Sets | Georgia Tech Electronic Thesis and Dissertation Archive |
Detected Language | English |
Type | Dissertation |
Page generated in 0.002 seconds