The Maxwell-Landau-Lifshitz-Gilbert System: Mathematical Theory and Numerical Approximation

J. A. Bohn
DOI: https://doi.org/10.5445/IR/1000133728
Abstract:This thesis deals with the mathematical theory and numerical approximation of the Landau--Lifshitz--Gilbert equation coupled to the Maxwell equations without artificial boundary conditions. As a starting point, the physical equations are stated on the unbounded three dimensional space and reformulated in a mathematically precise way to a coupled partial differential -- boundary integral system. We derive a weak form of the whole coupled system, state the relation to the strong form and show uniqueness of the Maxwell part of the solution. A numerical algorithm is proposed based on the tangent plane scheme for the LLG part and using a finite element and boundary element coupling as spatial discretization and the backward Euler method and Convolution Quadrature as time discretization for the interior Maxwell part and the boundary, respectively. Under minimal assumptions on the regularity of solutions, we present well-posedness and convergence of the numerical algorithm. For the pure Maxwell equations without the coupling to the LLG equation, we are able to show stronger results than in the coupled case. We derive a weak form for the Maxwell transmission problem and demonstrate existence and uniqueness of the weak solutions as well as equivalence with a strong solution. The proposed algorithm of finite-element/ boundary-element coupling via Convolution Quadrature converges with only minimal assumptions on the regularity of the input data. Again for the full Maxwell--LLG system, we show a-priori error bounds in the situation of a sufficiently regular solution. This is done by a combination of the known linearly implicit backward difference formula time discretizations with higher order non-conforming finite element space discretizations for the LLG equation and the leapfrog and Convolution Quadrature time discretization with higher order discontinuous Galerkin elements and continuous boundary elements for the boundary integral formulation of Maxwell's equations. The precise method of coupling allows us to solve the system at the cost of the individual parts, with the same convergence rates under the same regularity assumptions and the same CFL conditions as for an uncoupled examination. Numerical experiments illustrate and expand on the theoretical results and demonstrate the applicability of the methods. For the formulation of the boundary integral equations, the study of the Laplace transform is inevitable. We collect and extend the properties of the Laplace transform from literature. In the suitable functional analytic setting, we give extensive proofs in a self contained way of all the required properties.
Mathematics,Physics
What problem does this paper attempt to address?