HR: 0830h
AN: GP31B-0744    [PDF]
TI: The Effect of Grain Shape on the Magnetic Properties of Magnetite: A Finite Element Approach
AU: * Williams, W
EM: wyn.williams@ed.ac.uk
AF: Edinburgh University, School of GeoSciences, Kings Buildings, West Mains Road, Edinburgh, EH9 3JW United Kingdom
AU: Ridley, P
EM: phil@informatics.bangor.ac.uk
AF: University of Wales, Bangor, School of Informatics, Dean Street, Bangor, LL57 1UT United Kingdom
AB: Most micromagnetic models in rock magnetism use a regular grid of cells to construct the geometry of a magnetic grain. This has the advantage of allowing rapid evaluation of demagnetising fields using fast Fourier transforms, but means that significant modelling errors are introduced for non-cuboid shaped grains. Micromagnetic solutions for magnetite have usually predicted values of saturation remanence far lower than those observed experimentally. Although there are many possible reasons for this discrepancy, the inability of regular grids to model realistic grain geometries is a major drawback of this type of model. Since most naturally occurring magnetic minerals have irregular grain shapes a different approach is needed to produce a more realistic model. We will present a finite element (FE) micromagnetic model, allowing the definition of an arbitrary geometry for grain (or grains) of magnetite. The FE approach allows far more flexibility when modelling irregular grain shapes, and when modelling interacting grains of different geometries. The details of the different micromagnetic approaches will be examined, and the validity of different convergence criterion for the models will be discussed. In particular the suitability of energy minimisation versus a dynamic solution of the Landau-Lifshitz-Gilbert equation will be examined. The advantages of the finite element model will be illustrated by modelling cubic and spherical grains of magnetite. The FE model holds great promise for modelling multidomain grains, which is of particular importance to rock and palaeomagnetism. We will discuss how the current FE methods can be extended to model such grains.
DE: 1540 Rock and mineral magnetism
DE: 3210 Modeling
DE: 3230 Numerical solutions
DE: 5109 Magnetic and electrical properties
SC: Geomagnetism and Paleomagnetism [GP]
MN: 2003 Fall Meeting