University of Sheffield
Home
Research
People
Publications
PhD Projects
Vacancies
Links
Department of Engineering Materials Department of Engineering Materials logo

PhD Projects

J H Beynon

J Harding

T Schrefl

K P Travis

J H Beynon

J Harding

Simulations of filled carbon nano-tubes

Most materials are three-dimensional at the atomic scale. The standard crystal structures in textbooks assume that almost all the atoms in the crystal are far away from the edge. What would happen if this was not true? It is now possible to make one-dimensional materials by using carbon nanotubes to enclose a column of crystal. Now the crystal is only a few atoms across (length and width), although it is still many hundreds or thousands of atoms long? Do the traditional ideas about crystal structure and properties still hold? Work on electron microscopy (see the picture1) suggests that sometimes they do and sometimes they don’t. However the micrographs can only give projections of the structure and simulation is needed both to understand what the experiments are showing us and also to suggest new systems to look at. There are also some basic questions that need answering. Does the carbon tube simply act as an “elastic sock” to keep the crystal in place or are interactions between the tube and the crystal determining what the structure is? Can you get electron transfer between the crystal and the tube? If you do, how does this affect the conductivity? If the tube changes diameter, do you get dislocations? This project will use a variety of simulation methods to try and answer these questions. It is also part of a collaboration between the Universities of Cambridge (where Prof. Johnston make new kinds of tube and Drs Pyper and Bichoutskaia perform simulations) and Oxford (where Dr Sloan make tubes and Dr Kirkland does the electron microscopy). The project offers training in a wide range of simulation techniques for an area which is full of questions but with (as yet) few answers.
1 R. Meyer, J. Sloan, R.E. Dunin-Borkowski, A.I. Kirkland et al. Science 289 (2000) 1324

The structures and properties of perovskites

The perovskites structure is one of the commonest crystal structures in ceramics. The simple text-book structure is cubic. In real materials it is almost always distorted; the amount and type of distortion depending on the material in question. Often, the distortion can be controlled by adding various dopants in small quantities, or by making solid solutions of different materials. Perovskite materials have a wide range of structures. Some, such as barium and strontium titanate, are electroceramics. Some are metals or superconductors. The applications are as varied: from gas sensors to matrices for nuclear waste disposal. The project will investigate the structure and properties of some insulating perovskites that are of interest to the experimental groups in this Department, and will be in close contact with It will use a variety of techniques: static lattice minimisation using classical potential models, Monte Carlo methods and ab initio electronic structure calculations to investigate why these materials have the structures and properties that they do and how these may be controlled. The project will provide training in a range of both classical and quantum simulation methods.

Hyperdynamics and the growth of films

Molecular dynamics allows you to follow the trajectories of atoms by solving the equations of motion. Provided that you have a model for the forces within your system, you can predict its behaviour over time. Many general-purpose programs have been written to do this and we use one of the most popular, DL_POLY. However, this method is limited to processes that occur reasonably fast, so that they are likely to occur on the timescale of the simulation, normally a few nanoseconds. Many processes of interest have high activation barriers, so that they are very unlikely to happen. A set of techniques, called “hyperdynamics” have been developed to make a calculation run much faster, while remaining a faithful simulation of what is going on. The most obvious idea is to raise the temperature. This will certainly make the simulation run faster. However, a high-temperature calculation is not necessarily a faithful simulation of a low-temperature one because the relative probabilities of events change with temperature. This can be allowed for, and we have written a code to do this. This project would investigate the use of hyperdynamics in simulating problems in diffusion and growth. How do you control the way films grow to ensure that they have the properties that you want? If you were interested, it could involve a fair amount of computer programming but does not have to. If you were interested in finding new ways of visualising the simulations this could certainly be tried. On the theoretical side, the project is part of a collaboration with Dr Art Voter (Los Alamos) and Dr Bill Smith (Daresbury Laboratories). On the experimental side, there is a collaboration with Dr Rodney McKee at Oak Ridge National Laboratory.
1. Picture from RA McKee, FJ Walker and MF Chisholm, Science 293 (2001) 468

Modelling of biominearlisation and biominerals

Materials usually crystallise in fairly simple shapes. If they manage to produce a complex structure like the coccolith shown in the picture, then something unusual is going on. Somehow, the crystallisation is being organised. This can happen in a number of ways. One way is that large organic molecules form a template which tells the mineral crystal how to grow. Another way is that molecules attach themselves to the steps and kinks which are a vital part of the growth process. The coccolith on the left (courtesy Susan Stipp, Karen Henrikson) almost certainly uses both methods. If we could understand how it manages this, perhaps we could imitate it and manufacture new kinds of materials. This is the field of biomimetics . This idea of using molecules to control crystal growth is also important in more basic applications. The stuff that people put into their heating systems in hard-water areas is there to inhibit crystal growth and it (presumably – but no-one is entirely sure) works by trapping steps and kinks and stopping them growing. This project will investigate how molecules can control the growth of crystals. We will start with simple organic acid molecules and see how they stick onto the surface and whether they can prevent the mineral crystal growing. We will then move onto more complex systems. This is part of a collaboration with Susan Stipp in Denmark and also is connected to a major EPSRC modelling initiative on the modelling of bio-inorganic materials. The project will give training in a range of simulation techniques, in particular molecular dynamics methods.

Modelling of boundaries and interfaces

Real materials stop somewhere. This may be a surface, a grain boundary or some more complicated interface. In all cases, the structure of the solid changes; the atoms near the interface see a different environment to those in the bulk. For grain boundaries (and boundaries where different materials meet) the two crystalline arrangements must somehow join up. In simple cases, such as twin boundaries (shown on the left for strontium titanate) a perfect match is possible. In other cases the boundary contains defects of one kind or another. The properties of a solid at or near a boundary are often very different from those of the bulk. In particular, impurities often end up at boundaries. Atoms that are the wrong size or charge for the material can often find a congenial site at a boundary. The diffusion rates at boundaries are often much greater at boundaries than in the bulk and in ceramics the boundaries themselves are often charged. This fast diffusion at boundaries is often important in corrosion studies. The grain boundaries in oxide coatings that protect metals from further attack are usually the places where corrosion occurs. This project will simulate the structures of boundaries in crystals using both static and dynamic simulation methods. It will compare the simulated structures with high-quality TEM and diffraction work (some of which is done in the Department) and calculate the diffusion of ions at the boundaries. It will also look at the ways of preventing grain boundary diffusion (and hence corrosion) by stuffing the boundaries with atoms that inhibit boundary diffusion – the so-called reactive element effect. This project will give training in a number of simulation methods; in particular static and dynamic simulation using classical models.
1. Picture is work by S. Hutt, O. Kienzle, F. Ernst, S. Köstlmeier, C. Elsässer and M. Rühle ( web site hrem.mpi-stuttgart.mpg.de/English/projects/grainboundaries.html)

Modelling of trace elements in mantle materials

Most of what we know about the origin and subsequent differentiation of the Earth and terrestrial planets comes from the chemical analyses of trace elements and their isotopes in rocks. Interpretation of the chemical data relies on knowledge of the mechanisms of trace element incorporation into minerals at high temperatures and pressures. This project is part of a collaboration between the University of Bristol, three Italian universities (Messina, Bologna and Pavia), the University of Kiel (Germany) and the University of Oslo (Norway) to combine experiment and theory to understand why trace elements appear where they do: in bulk, solid rock, in the melt or at the boundaries between grains in the material. The idea that rocks can hold significant amounts of trace elements at the grain boundaries is held to explain why basalts can contain the concentrations of trace elements that they do. The collaboration will combine high pressure and temperature synthesis experiments, microbeam and X-ray characterization of run products with atomistic simulations and electronic structure calculations to a wide range of substituents in important mantle minerals (including perovskite-structured silicates, olivine, pyroxenes, garnet and amphiboles). The project at Sheffield will concentrate on calculating the segregration isotherms (i.e. the extent to which an impurity element wants to go to a boundary rather than stay in the bulk) and how these vary depending on which boundary is involved and the external conditions. This project will give training in a number of simulation methods; in particular static and dynamic simulation using classical models.
1. The picture on the left comes from T, Hlraga, I.M. Anderson and D.L. Kohlstedt; Nature 427 (2004) 699.

T Schrefl

Magnetic recording simulations on ultra high density media

Future magnetic recording media will store data at an areal densit of 1 Tbit per square inch and beyond. Computer simulations wich use state-of-the-art numerical algoritms are used to optimize both, the write head and the magnetic data layer.

Thermal fluctuations in magnetic memory elements

The basic structural unit that store information magnetically extends only a few nano-meters. Thermal fluctuations may cause information loss. Computer simulations of rare events are used to characterize the thermal decay in hard disk media and magnetic random access memories (MRAM).

K P Travis

Simulations of filled carbon nano-tubes

Most materials are three-dimensional at the atomic scale. The standard crystal structures in textbooks assume that almost all the atoms in the crystal are far away from the edge. What would happen if this was not true? It is now possible to make one-dimensional materials by using carbon nanotubes to enclose a column of crystal. Now the crystal is only a few atoms across (length and width), although it is still many hundreds or thousands of atoms long? Do the traditional ideas about crystal structure and properties still hold? Work on electron microscopy (see the picture1) suggests that sometimes they do and sometimes they don’t. However the micrographs can only give projections of the structure and simulation is needed both to understand what the experiments are showing us and also to suggest new systems to look at. There are also some basic questions that need answering. Does the carbon tube simply act as an “elastic sock” to keep the crystal in place or are interactions between the tube and the crystal determining what the structure is? Can you get electron transfer between the crystal and the tube? If you do, how does this affect the conductivity? If the tube changes diameter, do you get dislocations? This project will use a variety of simulation methods to try and answer these questions. It is also part of a collaboration between the Universities of Cambridge (where Prof. Johnston make new kinds of tube and Drs Pyper and Bichoutskaia perform simulations) and Oxford (where Dr Sloan make tubes and Dr Kirkland does the electron microscopy). The project offers training in a wide range of simulation techniques for an area which is full of questions but with (as yet) few answers.
1 R. Meyer, J. Sloan, R.E. Dunin-Borkowski, A.I. Kirkland et al. Science 289 (2000) 1324



| Modelling Activities in Engineering Materials | Department of Engineering Materials |
University Home Help and Assistance Search Send Feedback
Copyright © 2008, The University of Sheffield.