Programs in Physics & Physical Chemistry
|[Licence| Download | New Version Template] acct_v1_0.gz(39 Kbytes)|
|Manuscript Title: CFASYM: a program for the calculation of the asymptotic solutions of the coupled equations of electron collision theory.|
|Authors: C.J. Noble, R.K. Nesbet|
|Program title: CFASYM|
|Catalogue identifier: ACCT_v1_0|
Distribution format: gz
|Journal reference: Comput. Phys. Commun. 33(1984)399|
|Programming language: Fortran.|
|Computer: NAS 7000.|
|Operating system: MVT.|
|RAM: 400K words|
|Word size: 8|
|Keywords: Atomic physics, Asymptotic solutions, Electron-ion, Electron-atom, Electron-molecule, Scattering, Expansion methods, Continued fraction, Pade approximants.|
Nature of problem:
The linearly independent asymptotic solutions of the second-order differential equations encountered in close-coupling formulations of electron-atom, electron-molecule and electron-ion scattering are calculated by analytically continuing the appropriate asymptotic expansions. These solutions provide the boundary conditions necessary to compute the K- or S-matrices for the complete scattering problem.
The asymptotic expansions of the scattering wavefunction for an electron interacting with a target system at large separations via simple multipole potentials suggested by Burke and Schey, Burke et al. and by Gailitis are analytically continued and calculated in the form of a continued fraction using an algorithm devised by Nesbet. The method is equivalent to the use of Pade approximants and allows the radial point at which the asymptotic boundary conditions are used to determine the K- matrix to be considerably reduced compared to previous methods. This results in significant reductions in the computational expense of the overall calculation in the asymptotic region.
At energies close to thresholds the required accuracy can only be obtained at larger radial distances or by increasing the number of terms used in the asymptotic expansion. The second alternative is eventually limited by numerical round-off or by exponent overflow. Additional core storage is necessary if the number of channels or the number of terms in the asymptotic series is increased. Other limitations can arise from from the Coulomb function routines employed (these could easily be replaced).
The running time depends on the number of channels included. The total time for the test run on the NAS 7000 was 24 s (1.4 s for the execution step) and 1.96 s (0.48 s for the execution step) on the CRAY-1S.
|Disclaimer | ScienceDirect | CPC Journal | CPC | QUB|