Answer from Sirware team on the ability to just use Sir97 for structure refinement and ability to import a shelx file

%struct Cstisilicate
%job
%data
 cell   13.386  7.423 15.134  90 107.71  90
 spacegroup  c 2/c
 content     cs 8 o 60  si 24  ti 4
 reflections    csti.hkl
 format  (3i4,2f8.2)
 Fosquare
 wave Mo
%normal
%fourie
 fragment atoms.dat
 recycle 0
%menu
%end

The file "atoms.dat" contains the Dirdif solution:

      Specie X  Y  Z  [B]
i.e.

cs       0.11742   0.24082   0.40874  0.5
ti       0.25000   0.25000   0.00000  0.5
si      -0.00159   0.23204  -0.14521  0.5
si       0.32507   0.43041  -0.17053  0.5
si       0.34397   0.03826  -0.14276  0.5
o        0.38231   0.23392  -0.17581  0.5
o        0.30131   0.05257  -0.05820  0.5
o        0.10809   0.19559  -0.07492  0.5
o        0.25250  -0.03613  -0.22988  0.5
o        0.00000   0.26348  -0.25000  0.5
o        0.23947   0.07478   0.09150  0.5
o        0.42304   0.54841  -0.15646  0.5
o        0.44798  -0.07915  -0.12709  0.5
o        0.34989   0.22658  -0.24012  0.5

"if B is not supplied the Wilson B is used.
The directive "recycle 0" is used to avoid the automatic
L.S procedure. With the %menu command you can use the graphic
interface to refine the structure."

Least Squares Refinement

• Run Sir97 and open the above (or similarly edited) file which will not try and solve the structure and leave you in the Sir97 menu ready to refine the structure.

• 

• It is possible to employ a least squares refinement, setting parameters and options via the menu system.

  Before starting, go into Plot Style and set Ellipsoids for display with atom Labels. More optimal colours for each atom type can be created using the Settings, Drawing Colours (It is also possible to see the entire UNIT CELL/Multiple UNITCELLS and to set Polyhedra; by using Plotstyle, Polyhedra and/or Options, Edges/Symmetry)

• 

  ---

• Go into the Refine option. This leaves a number of options, including refinement; automatic placement of hydrogens; fourier map methods and display of various information relevant to the structure. For example, Th.Fact. checks for any abnormal thermal values.

• 

• To refine the structure, go into L.S.Q.

• 

• Click on Modify, then Isotropic. Crosses indicate atoms that are set to refine isotropically, which in this case are all atom types. These can be fixed or released in a variety of ways, including by species, individually, or all.

• 

• To start refining, click on OK until you get back to the Least Squares Refinement window and select GO to start refining.

• 

• In this example, it should be obvious that one of the Oxygens has too large a thermal and could be a spurious atom. If considered too large, Sir97 will prompt the user to delete this. Delete this atom and refine again, setting the number of cycles to 9 in the "Least Squares Refinement" options. Also, it may be best to set the reflection cutoff to zero and use all the reflections. Though opinions on this may vary from person to person. If you cut off too many reflections, the refinement may become unstable if you have many low intensity reflections, especially with some types of inorganics.

• 

• 

• Within the Refine option, you can use a Fourier button to do an F(o)-F(c) difference map and graphically see if any peaks make sense. Most of the found peaks could be due to "Fourier ripples" from the heavy atoms; or noise so be careful how you interpret these.

• 

• Within the Refine, L.S.Q., Modify option, go into the Aniso, Species dialog box and set the heavy Cs and Ti atoms to refine anisotropically. Then refine. If any of the thermals go to beserk values, Sir97 will set them back to Isotropic. If thermals do start going beserk, this could imply something is misset or not optimally set, such as the reflection cutoff or damping.

• 

• 

• Within the Refine, L.S.Q., Modify option, go into the Aniso, Species dialog box and set the medium Si atoms to refine anisotropically. Then refine. If any of the thermals go to beserk values, Sir97 will set them back to Isotropic. If thermals do start going beserk, this could imply something is misset or not optimally set, such as the reflection cutoff or damping.

• 

• Within the Refine, L.S.Q., Modify option, go into the Aniso, ALL dialog box and set all atoms (including Oxygens) to refine anisotropically. Then refine. If any of the thermals go to beserk values, Sir97 will set them back to Isotropic. If thermals do start going beserk, this could imply something is misset or not optimally set, such as the reflection cutoff or damping.

• 

• As an example of the extensive functionality within Sir97, using the Options, Edges and Symmetry you can define how many unit cells you wish to have drawn. The ViewPlane option can then be used to give a requested view of the structure.

• 

• 

• 

After quitting Sir97, the resulting atom positions will exported to a Shelx *.ins file.