In this thesis, I present the development and benchmarking of several theoretical methods designed to enable the rigorous modeling of magnetic properties of molecules containing one or a few heavy atoms, particularly single-molecule magnets. The new methods use a full four-component treatment of relativity, allowing spin–orbit effects to be taken into account from first principles, without the limitations that affect methods based on perturbation theory or coupling of a finite manifold of states. Extensions to include external magnetic fields and dynamical electron correlation in a multi-state framework are introduced. Performance benchmarks illustrate the range of molecules our implementation can handle and at what computational cost, showing that large molecules of practical significance are finally within reach of the four-component treatment. Computational results for magnetizabilities and zero-field splittings are presented for different series of molecules and compared to existing methods and experimental data, which provides invaluable information to guide the choice of parameters when using these methods. The potential of the methods and desirable future extensions are discussed.

Creator

• Reynolds, Ryan Dave

DOI

• https://doi.org/10.21985/n2-hdry-4v11

Subject

• Physical chemistry
• Chemistry
• Computational chemistry

Language

• en

Alternate Identifier

• etdadmin_upload_623099
• http://dissertations.umi.com/northwestern:14425

Keyword

• Single-Molecule Magnets
• Electronic Structure
• Relativity
• Theoretical Chemistry
• Zero-Field Splitting
• Magnetizability

Date created

• 2018-01-01

Resource type

• Dissertation

Rights statement

• In Copyright