Past Research
Electron Energy Loss Spectroscopy
(EELS) My thesis work involved the development of a new
scheme for the calculation of Energy Loss Near Edge Structure (ELNES) within
the framework of Density Functional Theory (DFT).[19,22,23,26]
Excellent agreement was found with experiment for the K-edges of graphite,
diamond and boron nitride. The use of the plane wave pseudopotential approach
offers the possibility of very large scale calculations, especially in
combination with parallel supercomputers.
Linear optical properties
To complement the calculation of ELNES, linear optical properties
were calculated through the complex dielectric function and it was implemented
within the same plane wave pseudopotential single particle approach.
Population analysis
Projecting electronic wavefunctions, expressed in terms of plane waves,
onto a localised atomic basis set allows a direct interpretation of the
electronic structure in terms of atomic charges, bond population or bond
orders and partial densities of states.[24,25]
Our method is integrated within the MSI Cerius2 interface to
CASTEP and is now widely used, allowing an interpretation more based on
chemical intuition.
k.p perturbation theory This
well known technique for the calculation of electronic band gradients and
curvatures has been extended to the case of both norm conserving and ultrasoft
non-local pseudopotentials allowing its use in modern electronic structure
calculations.[11]
Brillouin zone integration
An extrapolative approach to efficient Brillouin zone integration was developed
which avoids problems associated with band crossings.[20]
Gradients and curvatures calculated using k.p perturbation theory
were used to produce a piecewise quadratic representation of the electronic
band structure.
Ultrasoft pseudopotentials
I took part in a project to integrate the highly efficient and accurate
Vanderbilt ultrasoft pseudopotentials into the CASTEP code. I extended
the range of the pseudopotentials to cover the f-electron elements,
and implemented the calculation of stress tensors within the Vanderbilt
formalism, essential if complex, low symmetry, structures are to be relaxed.
At the same time, we developed a database of pseudopotentials to cover
the entire periodic table.[12]
Cleavage of Diamond
We calculated the theoretical strength of diamond in the <100>, <110>
and <111> directions from first principles. We were then able to explain
the remarkable dominance of the {111} cleavage plane when diamond is fractured.[15]
Complex inorganic solids
We have studied structure-property relationships in complex, low symmetry,
inorganic solids, recently investigating domain walls in NH4Cl,[14]
the structure and compressibility of garnets,[5,6,9,16]
the structure of Cu6PbO8[13]
and the structure and properties of CsHSO3.[21]
Systematic prediction of crystal
structures A generally applicable and systematic prediction
of crystal structures and their properties is an important goal of crystallography
and materials science. We have developed such a general and systematic
approach. This approach is based on a combination of graph theory with
quantum mechanics. As an application, structures, properties and relative
stabilities of small hypothetical carbon polymorphs have been investigated.[8,17]
New CASTEP I am
part of the seven strong UK based CASTEP Developers Group. We are currently
engaged in a complete rewrite of the CASTEP code -- the original being
unmaintainable. We have completed the design stage (which resulted in a
specification document for the project), and are now part way through the
coding. I am personally responsible for the ultrasoft pseudopotentials,
the I/O module, and the calculation of forces and stresses.
``Hard'' Carbon Nitride compounds
We have investigated the possible structures of CNx compounds,
and their expected properties.[10] The
collaboration has now extended to the calculation of the electronic properties
of carbon nanotubes[1]. My extrapolative
Brillouin zone integration scheme is well suited to this, and the densities
of states so calculated exhibit the clear van Hove singularities expected
for such two dimensional electronic systems.
Structural properties of lanthanides
and actinides We have shown that the structural properties
of f-electron compounds can be calculated on the same footing as
the rest of the periodic table within the ultrasoft pseudopotential approximation.[7]