ShelxtlRestraints教程手册 Using Free Variables and Restraints.pdf

Using Free Variables and Restraints Copyright © by Gregory S. Girolami, 1999 Crystal structures can be divided into two classes: those in which the refinement is well behaved from start to finish, and those in which the refinement is impeded by one or more kinds of crystallographic problems. The latter variety of crystal structure is the subject of the present document. Among the most common kinds of crystallographic problems are disorder and poor convergence of the least-squares refinement. A definition of disorder will be given below, but the presence of disorder is easy to establish: disorder is signaled by the appearance of electron density peaks (either as Q peaks in the difference maps or as unacceptably elongated thermal ellipsoids) whose locations correspond to chemically reasonable alternative locations for certain atoms in the structure. In order to obtain the lowest possible value for wR2 (the least-squares residual), you must determine the chemical identities of these electron density peaks and include these disordered atoms in the refinement. Poor convergence of the least-squares refinement is a different problem, and is signaled by the fact that the maximum shift/error (a value which is calculated and reported on-screen by XL) stays large (e.g. over 1) even after many cycles of refinement. Herein we describe the use of two tools which are invaluable for managing these problems: free variables and restraints. These tools allow the user to exercise precise control over the least-squares refinement using the SHELXTL software package. A good working knowledge of the use of free variables and restraints is crucial for the successful refinement of many crystal structures. It should be emphasized, however, that these crystallographic problems can only be solved by thinking like a chemist ; the software package will not do your