This page's content is no longer actively maintained, but the material has been kept on-line for historical purposes.
The page may contain broken links or outdated information, and parts may not function in current web browsers.


MOLSCAT is a code for quantum mechanical (coupled channel) solution of the nonreactive molecular scattering problem. Code is implemented for various types of collision partners. In addition to the essentially exact close coupling method several approximate methods, including the Coupled States and Infinite Order Sudden approximations, are provided. The code is in near standard FORTRAN 77 and has been ported to a large number of platforms. The source code supplied was run at this site on an IBM RS/6000 and also on IBM and compatible mainframes. It should work on most other machines; for the Cray, however, modifications to run in single precision should be made.

MOLSCAT is an evolving computer program designed for use in a research environment. Although the authors made every attempt to ensure that it works correctly, it is the responsibility of the user to determine whether the code is suitable for a given application and to determine accuracy of any results. No warranties of any kind are made.

Bug Fixes

Changes to the code between the original release of Version 14 in August 1994 and the death of co-author Sheldon Green in December 1995 are documented on the bug-fix page. Corrected versions of subroutines can also be accessed there.


Complete documentation for MOLSCAT, version 14, is available as a plain text (ascii) file or in hypertext suitable for viewing with a web browser.

The HTML version of the documentation is also available for installation on your local machine.

New users might want to start with a tutorial on input data which walks through construction of an input deck for simple, standard cases.

Several sample input decks and output files are provided below as examples of using various options.

Source Code

Source for version 14 is available as uncompressed FORTRAN source (about 665 Kbyte) or as Unix-compressed source (about 270 Kbyte).

Matrix and vector manipulations are handled in version 14 by calls to LAPACK and BLAS library routines. (See Section 10.7 of the Documentation.) Optimized versions of these libraries are available for many platforms and should be used if at all possible. The most recent FORTRAN source codes can be obtained from a Netlib server.

FORTRAN source for the LAPACK and BLAS routines which are used by the current version of MOLSCAT are also available locally.

Other routines which may be faster, especially if optimized LAPACK and BLAS are not available include.

Other MOLSCAT source code:

Time and date routines:
CPU and wall-clock times are provided by platform-specific routines. For the RS/6000 the following FORTRAN and C routines should replace dummy routines in the distributed version.
Special potential routines:
Potentials obtained from ab initio calculations are often expressed as tabulations of radial coefficients for the angular terms. This replacement POTENL routine obtains the radial coefficients from a data set and does spline interpolation as described in S. Green, J. Chem. Phys. 67, 715 (1977).

In addition to the source code for MOLSCAT available here, source codes and documentation are available from Jeremy Hutson via anonymous ftp. These include the related BOUND program and machine-dependent code for other platforms.

Sample Test Input Decks and Output Files

These exercise some of the MOLSCAT options and are available below for some of the supported "collision types". A brief description of the calculations found in the tests is available.

Earlier Versions

Attempts are made in upgrading MOLSCAT to provide compatibility with earlier versions; in general input decks should run without modification and interfaces to user-supplied potential energy routines are changed only if deemed absolutely necessary. Nonetheless, some changes in version 14 may require modification of input decks and/or ancillary codes (for example, nonstandard potential routines, routines for nonstandard collision types, and programs which process the S-matrix save tape). These changes are noted in the documentation. For those requiring compatibility with earlier codes, MOLSCAT, version 12 is still being provided.

Parallel Version

A parallel version of MOLSCAT, based on version 12 of the serial code, was implemented for the Intel iPSC/860 parallel supercomputer at Daresbury Laboratory, UK.

More recently, George McBane at Grand Valley State University has implemented a parallelized version of MOLSCAT v. 14 called PMP MOLSCAT.

S-matrix Save Tape "Post-processor" Codes.

MOLSCAT routinely calculates state-to-state integral cross sections and will optionally calculate impact approximation pressure broadening cross sections. For other phenomenological cross sections it is generally necessary to save the S-matrices (see Section 2.5 of the documentation) for further processing. Several programs are available to compute a variety of cross sections, including state-to-state differential cross sections, pressure broadening cross sections, and Senftleben-Beenakker effect (generalized transport property) cross sections from S-matrices saved by MOLSCAT.

Molecular Collision Rates of Astrophysical Interest

MOLSCAT was developed to obtain collision rates for molecules in the interstellar gas which are needed to understand microwave and infrared astronomical observations.


These pages were originally created by MOLSCAT co-author Dr. Sheldon Green, who passed away in December 1995. They are not currently being updated except for purposes of page formatting and link verification.

If you have any questions about MOLSCAT, please contact Dr. Jeremy M. Hutson at