Elsevier Science Home
Computer Physics Communications Program Library
Full text online from Science Direct
Programs in Physics & Physical Chemistry
CPC Home

[Licence| Download | New Version Template] ackj_v1_0.gz(16 Kbytes)
Manuscript Title: Normal coordinate analysis of crystals.
Authors: J.Th.M. de Hosson
Catalogue identifier: ACKJ_v1_0
Distribution format: gz
Journal reference: Comput. Phys. Commun. 10(1975)104
Programming language: Fortran.
Computer: CYBER 74.
Operating system: SCOPE 3.4.1, LEVEL 373.
RAM: 155K words
Word size: 60
Keywords: Solid state physics, Dynamical matrix, Normal coordinate, Lattice dynamics, Infrared spectra, Eigenvalue problem, Threshold jacobi method, Householder procedure.
Classification: 7.8.

Nature of problem:
A program has been written for solving the vibrational secular equation in cartesian coordinates and for adjusting a set of force constants to give a fit of calculated and observed frequencies, which are measured by means of infrared transmission spectroscopy.

Solution method:
Two methods for calculating the eigenvalues and eigenvectors are included in the program deck. These methods are: the threshold serial Jacobi method and the Householder procedure. Although the latter is faster than the Jacobi method, the eigenvectors are more accurate in the case of the threshold serial Jacobi procedure.

The program is dimensioned for 20 atoms per unit cell. The number of degrees of freedom may be increased to 99, depending on the available storage capacity.

Unusual features:
Using the symmetry coordinates, it is possible to diagonalize the dynamical matrix in blocks. The symmetry coordinates can be calculated in some cases by the method and the program given by Warren and Worlton, using a reducible multiplier representation of the point group of the wavevector (k=0). However, the input of the symmetry coordinates is not necessary for our program. Block-diagonalization of the dynamical matrix set up in cartesian coordinates is often not very useful for solving the eigenvalue problem.

Running time:
For the eigenvalue problem solved by threshold serial Jacobi method for KNIF 3 (without iteration procedure) : 12.6 s.