Generalisations of the finite-element and R-matrix methods
- 17 January 1978
- journal article
- Published by IOP Publishing in Journal of Physics B: Atomic and Molecular Physics
- Vol. 11 (2) , 193-207
- https://doi.org/10.1088/0022-3700/11/2/005
Abstract
A generalisation of the finite-element method is used to construct a series of highly efficient finite-basis-set methods for the solution of a one-dimensional Schrodinger equation in the continuum part of the eigenvalue spectrum. These methods are far more accurate than the standard R-matrix method and effectively eliminate the need for the correction procedures commonly applied to this earlier method. The poor convergence of the R-matrix method is found to be the result of an analogue of the Gibb's phenomenon of Fourier series expansions arising due to the inability of the basis functions to resolve the boundary conditions. It is shown how this problem can be removed by adding a few local basis functions which then allow the boundary conditions to be resolved at any energy.Keywords
This publication has 31 references indexed in Scilit:
- Natural boundary condition methods for nuclear reactions (II)Nuclear Physics A, 1976
- Natural boundary condition methods for nuclear reactionsNuclear Physics A, 1976
- The R-Matrix Theory of Atomic ProcessesPublished by Elsevier ,1976
- Background phase shift inR-matrix theoryPhysical Review C, 1975
- New treatment of the one-particle continuum in nuclear reaction theoryPhysical Review C, 1974
- Approximate Solution of a Sturm-Liouville System Using Nonorthogonal Expansions: Application to α-α Nuclear ScatteringJournal of Mathematical Physics, 1971
- A new method for the solution of the Schrödinger equationJournal of Computational Physics, 1970
- New Method for Constructing Wavefunctions for Bound States and ScatteringThe Journal of Chemical Physics, 1969
- Solution of Coupled Equations by-Matrix TechniquesPhysical Review B, 1967
- Une formulation unifiée de la théorie des réactions nucléairesNuclear Physics, 1957