THIS INFORMATION IS OBSOLETE AND IS PROVIDED ONLY FOR ITS HISTORICAL VALUE

QCPE
Main / Catalog / Section09


559. BIGSTRN-3: General-Purpose Empirical Force-Field Program

by Robert B. Nachbar, Jr., Merck Sharp & Dohme Research Laboratories, Rahway, New Jersey 07065 and Kurt Mislow, Chemistry Department, Princeton University, Princeton, New Jersey 08544

Converted by F. Sidhu Department of Chemistry, Indiana University, Bloomington, Indiana 47405

This system is a direct conversion of QCPE 514 for use on the IBM 3090 with vector facility. This system makes extensive use of IBM's ESSL (Engineering and Scientific Subroutine Library).

BIGSTRN-3 is a general-purpose empirical-force-field (EFF) program. Its primary goal is to optimize a given molecular geometry in a given force field, possibly under a user-defined set of constraints, to the nearest stationary structure. Ancillary operations include heat of formation calculation, calculation of thermodynamic functions, normal mode analysis, and tracing of conformational reaction pathways.

Four empirical force fields (Andose-Mislow (AM), Engler-Andose-Schleyer (EAS), Allinger (MM2), and Ermer-Lifson (CFF)) are provided with the program in the form of individual data files. Because most of the force field is defined in the data file, BIGSTRN-3 has maximum independence from any particular force field. Additional parameter data files can be searched at run time to resolve undefined interactions. There are a number of geometry optimization schemes available to the user, in which either the energy or the gradient norm of the energy (the forces) can be minimized. These optimizers use the steepest-descent, conjugate- gradient, variable-metric or Newton-Raphson methods. Analytical first and/or second derivatives of the energy or the gradient norm of the energy with respect to Cartesian coordinates are used throughout the program. The calculation of transition states (single partial maxima) and higher-order partial maxima is easily achieved with BIGSTRN-3; these stationary structures can be calculated directly, and one does not need to resort to complex "driving" techniques. Upon convergence, the matrix of analytical second derivatives of the energy is evaluated and diagonalized. From the number of negative eigenvalues, one can determine the nature of the stationary structure obtained: zero, a minimum (M(0)); one, a single-partial maximum (m(1)); two, a double-partial maximum (M(2)); etc.

The imposition of geometric constraints has been enhanced to include distances, bond angles, torsion angles and out-of-plane angles.

Comprehensive diagnostic messages are provided for user input errors. The format of the printed output has been modified to provide the user the maximum amount of information about the force field used, the given molecule, and the geometry optimization in a form as unambiguous and clear as possible.

FORTRAN (IBM VS2.2 PUT8801) Lines of Code: 21,202



Computational Chemistry List --- QCPE Main --- About This Site