Crystals Manual

Chapter 1: Introduction To The System

: 1.1: Layout of this manual; 1.2: MAJOR CHANGES for WINDOWS

1.1: Layout of this manual

This manual is organised as follows

 1      Definitions and Conventions
 2      Initial Data Input
 3      Reflection data
 4      Atoms and Parameters
 5      Structure factors and least squares
 6      Fourier and Patterson maps
 7      Geometry and printing results
 8      Graphics - CAMERON
 9      Immediate commands - tailoring the program
 10     Twinned Data
 11     SCRIPTS

The current version of CRYSTALS , Issue 11, was released in 1999 for machines running WINDOWS95. It continues to be based on a version (Issue 2) written by J.R. Carruthers in collaboration with J.S.Rollett during 1977-78, which was a rewrite of the 1975 CRYSTALS system implemented on the ICL 1900 series of computers. It contains significant contributions by Paul Betteridge, David Kinna, Lisa Pearce, Alen Larsen, and Eric Gabe and many students and visitors to the Chemical Crystallography Laboratory, Oxford.

In CRYSTALS, the course of a calculation is controlled from a master control channel, which may either be a card image file or an on-line terminal. When the data is being presented directly to the program from the terminal the program is being used in the INTERACTIVE and ON-LINE mode. When the job is initiated from a terminal but the data is read from a control file the program is being used in the ON-LINE mode, and when both initiation and control is from a control file the task is a BATCH job. A mixed mode is available in some implementations, in which control is passed from the INTERACTIVE terminal to a pre-prepared control file, and later returned to the terminal.

In all ON-LINE modes, a special subset of the output normally sent to the printer is sent to the terminal. This provides the user with the information that he needs to monitor the calculation in order to initiate the next step of the computation. In all modes of use CRYSTALS produces a well laid out and comprehensive line printer record of the calculation.

1.2: MAJOR CHANGES for WINDOWS

The opportunity has been taken to make substantial changes throughout CRYSTALS to coincide with the release of the 32 bit version for WINDOWS95/NT. The majority of the changes (several thousand edits) are concerned with the screen and file output, and should not distress existing users.

Major changes are:

1      New .DSC file
2      New ADP handling
3      New weighting for Fsq
4      New items stored in LIST 30

New .DSC File
The internal format of some of the lists has been extended to accomodate information needed to meet current publication requirements. There is some degree of compatibility between old format .DSC files and the new CRYSTALS. A facility in /EDIT attempts to do the internal reformatting. There is no backwards compatibility. There is full forward compatibility at the ASCII level - the old LISTS5,6,12 and 16 can be input into the new program.

New Temperature Factor (ADP) handling
In the original version of CRYSTALS, the value of Uiso was used to indicate whether it was a real usable value (e.g. 0.055), or a flag indicating that the atom was anisotropic (e.g. 0.00).

This location in LIST 5 has now been changed to a proper flag, whose value indicates what kind of information is stored after the positional parameters. This has enabled us to have more complex models for the electron distribution.

original keys:

      type serial occ u[iso] x y z u[11] u[22] u[33] u[23] u[13] u[12] spare

revised keys

      type serial occ FLAG x y z u[11]  u[22] u[33] u[23] u[13] u[12] spare
                                 U[ISO]                               spare
                                 U[ISO] SIZE                          spare
                                 U[ISO] SIZE  DECLINAT AZIMUTH        spare

The value of 'FLAG' is used on input of atoms to indicate what kind of patameters will follow, and is used during calculations for the interpretation of the parameters.

FLAG interpretation The following table shows the interpretation of the FLAG parameter.


FLAG  meaning    parameters
'old' types of atoms:

 0    Aniso ADP  u[11]  u[22] u[33] u[23] u[13] u[12] 
 1    Iso ADP    U[ISO]                               

New 'special' shapes:

 2    Sphere     U[ISO] SIZE                          
 3    Line       U[ISO] SIZE  DECLINAT AZIMUTH        
 4    Ring       U[ISO] SIZE  DECLINAT AZIMUTH

The parameters have the following meaning for the new special shapes:

Special U[iso] U[iso] is related to the 'thickness' of the line, annulus or shell.

Special SIZE SIZE is the length of the line, or the radius of the annulus or shell.

Special DECLINAT DECLINAT is the declination angle between the line axis or annulus normal and the z axis of the usual CRYSTALS orthogonal coordinate system, in degrees/100.

Special AZIMUTH AZIMUTH is the azumuthal angle between the projection of the line axis or annulus normal onto the x - y plane and the x axis of the usual CRYSTALS orthogonal coordinate system, in degrees/100.

If either of these angles is input with a value greater than 5.0, it is assumed that the user has forgotten to divide by 100, which is thus done automatically.

Accessing 'Special shapes' The special shape parameters can be accessed just like any traditional parameters. FLAG is not a refinable parameter, and should only be changed in /EDIT if corresponding changes are to be made to the actual parameter values. This is normally automatic if the FLAG is changed to 0 or 1. To facilitate the generation of special shpes from groups of atoms, the following directives are available in /EDIT.

SPHERE newserial atomlist This creates a 'shell' shape from the specified atom list. The centre of the shell is at the centre of gravity, the size is the mean distance of the given atoms from the centre, and the occupancy is equal to the sum of the occupancies of the atoms listed. U[iso] is the mean of the U[iso] or Ueqiv of the listed atoms. The atom TYPE is QS, with the given serial number. The original atoms are not deleted, though they should be or their occupancy set to zero. The atom type, QS, should be changed to something appropriate.

RING newserial atomlist This creates an 'annulus' shape from the specified atom list. The centre of the ring is at the centre of gravity, the size is the mean distance of the given atoms from the centre, and the occupancy is equal to the sum of the occupancies of the atoms listed. U[iso] is the mean of the U[iso] or Ueqiv of the listed atoms. The atom TYPE is QR, with the given serial number. The original atoms are not deleted, though they should be or their occupancy set to zero. The atom type, QS, should be changed to something appropriate. The DECLINATION and AZIMUTH are computed from the constiuent atoms.

LINE newserial atomlist This creates an 'line' shape from the specified atom list. The centre of the line is at the centre of gravity, the size is twice the mean distance of the given atoms from the centre, and the occupancy is equal to the sum of the occupancies of the atoms listed. U[iso] is the mean of the U[iso] or Ueqiv of the listed atoms. The atom TYPE is QL, with the given serial number. The original atoms are not deleted, though they should be or their occupancy set to zero. The atom type, QS, should be changed to something appropriate. The DECLINATION and AZIMUTH are computed from the constiuent atoms.

REFORMAT This instruction converts an old (non-FLAG) version of LIST 5 to the new format.

Currently, no action is taken by CRYSTALS for special shapes lying on positions of special site symmetry. The user must set up the appropriate constraints in LIST 12 (usually preventing the refinement of one or more parameters), and set up the correct occupancy in LIST 5.

Since the special shapes can coexist with normal atoms, it is possible to embed normal atoms in the special shapes to give a 'peakey' electron distribtion. The sums of the occupancies will need to be constrained.

New Weighting for Fsq refinement
Scheme 14 (Chebychev weighting) has been made more robust to ragged distributions of delta squared. The major influence is in Fsq refinement, though there is also a small (improved) effect on F refinement.

New items stored in LIST 30
LIST 30 has benn extended because of changes in the new CIF dictionary.