Volume: 1.5 Mb
Download the package: here

    The package had been originally designed by Robert D.Cowan at LLNL, USA. The last versions of the original codes can be obtained by ftp from  t4.lanl.gov/pub/cowan.

    The present version for the personal computers has been designed by A.Kramida  at the Institute for Spectroscopy RAS, Troitsk, Russia.

    Executables compiled with Lahey Fortran 77/32 can run on versions of Windows that has a 32-bit support (95, 98, NT, 2000),.but cannot run under Windows XP or later versions. To run under these latter ones you should upgrade as described here for compilation with Salford Fortran 77.

The package includes the following programs:

• RCN Mod 36k - calculates single-configuration radial wavefunctions for a spherically symmetrized atom via one of the four homogeneous-differential-equation approximations to the Hartree-Fock method, including relativistic corrections.
• RCN2k accepts radial wavefunctions from RCN and calculates multiple-configuration radial integrals. Prepares the input files for RCG.
• RCG Mod 11k computes angular matrix elements, energy levels, and atomic spectra, including energy levels, transition wavelengths, radiative and autoionization transition rates, dielectronic recombination cross-sections, electron-impact excitaion cross-sections, etc.
• RCE Mod 20k makes a least-squares fitting of atomic energy levels, i.e. adjusts the radial integrals computed by RCG to obtain the best fit of experimentally known energy levels.

The new features are present only in RCG and RCE codes.

In RCG, an efficient packing of temporary files is used. This permits to use this program for calculation of very large sets of interacting configurations which, in the original code version, produced on the hard disk temporary files of several-hundred-megabytes size, impeding the use of this program for such calculations on most personal computers.

In RCE, a new concept of vector-input file is implemented. This makes the usage of the program much more convenient and efficient. A different algorithm of level naming is used in the fitting procedure which makes the convergence of iterations much more stable in case of multiple interacting configurations.

More information is contained in the README file.