| Summary: | Two new methods to aid in the calculation of ab initio energies are presented. The first method sets out to change the way that systems that have multiple elements that would benefit from a multireference treatment are handled. The method proposes splitting the system into multiple small active spaces in order to avoid the computational issues present with a single large active space. The method is developed using localised orbitals and tested on Cr2 and the molecule N2 both at long bond lengths. The second method proposes and develops a method for the production of starting orbitals for CASSCF and CASPT2 calculations. This method requires multiple subunits of the system which are then optimised to produce subsets of orbitals. These orbitals are then combined using a custom script (detailed within) to form a coherent orbital set for the entire system. This method is then tested on two transition metal complexes, one of which could not be successfully treated using traditional methods. Along with the two new methods, two studies of individual transition metal systems are presented. The first covers a comparison between a published DFT/OLYP study of complexes of the form (nacnac)MIII (NPh) with a DFT/B3-LYP, CASSCF, CASPT2 and CCSD(T) study of similar systems presented here. The second studies the complex [CrII (CN)5 ] 3– . It produces a possible transition state, and discussed its role in a Berry pseudorotation-like mechanism for this system.
|
|---|
