Programs in Physics & Physical Chemistry
|[Licence| Download | New Version Template] aajc_v1_0.gz(44 Kbytes)|
|Manuscript Title: EDWIN: a program for calculating inelastic molecular collision cross sections using the exponential distorted wave and related approximate methods.|
|Authors: G.G. Balint-Kurti, J.H. van Lenthe, R. Saktreger, L. Eno|
|Program title: EDWIN|
|Catalogue identifier: AAJC_v1_0|
Distribution format: gz
|Journal reference: Comput. Phys. Commun. 19(1980)359|
|Programming language: Fortran.|
|Computer: IBM 360/195.|
|Operating system: IBM SYSTEM/360 OS.|
|RAM: 118K words|
|Word size: 8|
|Keywords: Molecular physics, Rotationally inelastic Scattering, Cross sections, Schrodinger equation, Elastic, Distorted wave, Expotential distorted Wave, Magnus method, Expotential method.|
Nature of problem:
EDWIN calculates inelastic molecular cross sections for collisions between an atom and a rigid rotor molecule.
Four approximate methods are available as options, all based upon the distorted wave Born approximation. They are: the distorted wave Born method, the centrifugally decoupled distorted wave method and their expotential or unitarised couterparts. The diagonal channel wavefunctions are propagated from R to R + delta R using the Magnus propagation method. At large separations as semiclassical WKBJ form is used. The contributions to the non-zero distorted wave integrals are accumlulated for each propagation interval. By using a phase-amplitude representation of the channel wavefunctions, analytic expressions can be found for these contributions, while each propagation interval may contain many wavelengths without loss of accuracy in the integrals.
The present program treats only atom-linear molecule rotational scattering, but it is reasonably simple to adopt it to other physical problems. There are some limits on the number of channels and the potential which can be easily extended. The potential should be provided as a Lengendre polynomial expansion in the atom-molecule angle. Both the value and first derivative of the R-dependent part must be available.
The program is designed around the use of arrays in blank common which are dynamically allocated and whose storage requirements are assigned during execution. In the present version the dynamical storage allocation has been replaced by a single parameter in the main program which the user can set to suit his needs.
These depends on the various accuracy criteria and the complexity of the problem. Very roughly, the running time goes up with the square of the number of rotational levels. A typical calculation (Ar-N2, 768K, 8 rotational levels, 18 total angular momentum values) using the CDEDW method with an accuracy of about 1% reequires 3 min on the IBM 360/195.
|Disclaimer | ScienceDirect | CPC Journal | CPC | QUB|