# Overview
Our recent work has focused on the electronic structure and dielectric responses of semiconductors and oxides and development and application of a Gaussian orbital code for calculation of spectral functions in materials called **EXCITON**.
We mainly use the *CRYSTAL* and *CASTEP* codes for Gaussian and plane wave density functional theory (DFT) calculations. Conventional density functionals generally underestimate band gaps of gapped materials by up to 50%. Since we are interested in calculating optical spectra and other dielectric properties of materials, we require an approach where band gaps are predicted to be close to experimental values. We mainly study these properties in systems with large unit cells, including slabs representing surfaces of semiconductors or oxides or supercells containing defects. We have found that hybrid DFT approaches, in which the exchange-correlation density functional is partly substituted by the Fock exchange functional, can be adjusted to give optical spectra in reasonable agreement with experiment, except for excitonic lines around the absorption threshold.
A Gaussian orbital basis can be chosen to describe valence and low-lying conduction bands well and therefore can calculate the part of the dielectric function measured by spectroscopic ellipsometry quite well. Optical matrix elements vary rapidly in the Brillouin zone and hence it is necessary to use dense k-meshes in order to obtain converged optical spectra. The efficient Gaussian basis approach allows bulk cells to be sampled up to 24x24x24 or higher on a Monkhorst-Pack grid.
The goal of the *EXCITON* project is to develop a many-body code capable of performing GW and Bethe-Salpeter equation calculations for crystalline solids in a Gaussian orbital basis. This is an extension of the work that we do using the *CRYSTAL* code which is a Gaussian orbital code for Hartree-Fock and density functional calculations on solids.
We are also studying defects in bulk oxides and at semiconductor-oxide interfaces which are relevant for MOSFET production. In particular we have focussed on classical and quantum mechanical approaches to generating representative interfaces and methods for passivating defects at those interfaces.
Finally, we are also interested in strongly correlated electron systems and their charge and orbital order. A recent calculation on charge and orbital order in magnetite was used to calculate ^{57}Fe NMR spectra for A and B-site ions in magnetite and showed good agreement with experiment.
For more details on these projects, click on the research link on the left. |