CCP14 Homepage - Tutorials and Examples - Louis Farrugia's WinGX Single Crystal Suite - Example Structure refinement of an Inorganic Structure Using Shelx

[CCP14 Home: (Frames \| No Frames)] CCP14 Mirrors: [UK] \| [CA] \| [US] \| [AU]	What's New	Introduction	Site Map
Search the CCP14	Download Programs What do you want to do? (lists of software by crystallographic method)	Tutorials	Solutions

(This Webpage Page in No Frames Mode)

CCP14

Tutorials and Examples

Louis Farrugia's WinGX Single Crystal Suite

Example Structure refinement of an Inorganic Structure Using Shelx

The CCP14 Homepage is at http://www.ccp14.ac.uk

[Tutorials page] | [WinGX Tutorial Index]

This example uses the raw hkl data used to determine the crystal structure of Cs2TiSi6O15 as published in I.E. Grey, R.S. Roth, M.L. Balmer, Journal of Solid State Chemistry, 131, 38-42 (1997). While the published structure was solved by manual interpretation of a Patterson map, the following example is based around an automatic Patterson based structure solution by Dirdif for Windows.

Click here to obtain the csti.hkl Analytical Absorption Corrected FoSQ (HKLF 4) HKL data.

Assuming you have graphically sussed out the solution and checked the bond-lengths and angles, a quick way to sort out if some of the atoms may be spurious and misassigned is do a few cycles of refinement; and note if some of the thermal parameters go very large. Open the Results file in WinGX by the menu option, Refine, RES FileOpen then save it via the editor command File, Save as INS.

results file as generated by dirdif

To do some cycles of structure refinement, select Refine, SHELXL-97, in this case, using the Analytically Absorption Corrected data.

Shelx refining via the WinGX interface

(Note: This run-through glosses over graphically looking at the structure and evaluating the bond-length and angles list which would normally be occuring.) Looking at the results, the thermal on O14 is very large, implying it could be a bogus atom. Delete this and extend the number of Least Squares cycles to 9 by changing L.S. 4 to L.S. 9

Gui WinOrtep view showing very large thermal for O14

CS1   1    0.117374    0.241669    0.409920    11.00000    0.03160
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01444
SI3   3   -0.002245    0.231140   -0.146855    11.00000    0.01286
SI4   3    0.324271    0.429524   -0.166738    11.00000    0.01928
SI5   3    0.345265    0.040898   -0.144329    11.00000    0.02120
O6    2    0.381402    0.240320   -0.169454    11.00000    0.03022
O7    2    0.301623    0.051163   -0.057210    11.00000    0.02355
O8    2    0.107629    0.189508   -0.074593    11.00000    0.02792
O9    2    0.253355   -0.037705   -0.230301    11.00000    0.04742
O10   2    0.000000    0.250040   -0.250000    10.50000    0.07091
O11   2    0.236178    0.075979    0.090143    11.00000    0.02842
O12   2    0.418745    0.564177   -0.151408    11.00000    0.06600
O13   2    0.451799   -0.079446   -0.128166    11.00000    0.04612
O14   2    0.367496    0.229404   -0.245969    11.00000    0.19651

Select Refine, SHELXL-97, and refine to convergence.

TITL SHELXL INS file, from DIRDIF output for EG1SCR
CELL   0.71073   13.38600  7.42300 15.13400  90.0000 107.7100  90.0000
ZERR   4          0.00200  0.00200  0.00100   0.0010   0.0500   0.0500
LATT   7
SYMM     -X   ,     Y   , 1/2-Z
SFAC  CS  O   SI  TI
UNIT      8   60   24    4
L.S.  9
BOND
FMAP  2
PLAN  10
WGHT    0.100000
FVAR       0.18918
CS1   1    0.117308    0.241523    0.409980    11.00000    0.03158
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01398
SI3   3   -0.002364    0.231166   -0.146692    11.00000    0.01230
SI4   3    0.324126    0.429424   -0.166513    11.00000    0.01926
SI5   3    0.345325    0.040957   -0.144281    11.00000    0.02107
O6    2    0.381421    0.237640   -0.170296    11.00000    0.02928
O7    2    0.302075    0.050575   -0.056359    11.00000    0.02317
O8    2    0.107916    0.188514   -0.074217    11.00000    0.02764
O9    2    0.253222   -0.039078   -0.230187    11.00000    0.04666
O10   2    0.000000    0.244150   -0.250000    10.50000    0.06607
O11   2    0.236870    0.076324    0.090848    11.00000    0.02940
O12   2    0.415746    0.574306   -0.143080    11.00000    0.05927
O13   2    0.450497   -0.081256   -0.125431    11.00000    0.04875
HKLF    4
 
REM  SHELXL INS file, from DIRDIF output for EG1SCR
REM R1 =  0.1508 for   1783 Fo > 4sig(Fo)  and  0.1544 for all   1874 data
REM     48 parameters refined using      0 restraints
 
END

After checking that the result is reasonable, set the heavy Cs and Ti atoms refining anisotropically using the ANIS command. (e.g., ANIS CS1 TI1)

TITL SHELXL INS file, from DIRDIF output for EG1SCR
CELL   0.71073   13.38600  7.42300 15.13400  90.0000 107.7100  90.0000
ZERR   4          0.00200  0.00200  0.00100   0.0010   0.0500   0.0500
LATT   7
SYMM     -X   ,     Y   , 1/2-Z
SFAC  CS  O   SI  TI
UNIT      8   60   24    4
L.S.  9
BOND
ANIS CS1 TI1
FMAP  2
PLAN  10
WGHT    0.100000
FVAR       0.18918
CS1   1    0.117308    0.241523    0.409980    11.00000    0.03158
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01398
SI3   3   -0.002364    0.231166   -0.146692    11.00000    0.01230
SI4   3    0.324126    0.429424   -0.166513    11.00000    0.01926
SI5   3    0.345325    0.040957   -0.144281    11.00000    0.02107
O6    2    0.381421    0.237640   -0.170296    11.00000    0.02928
O7    2    0.302075    0.050575   -0.056359    11.00000    0.02317
O8    2    0.107916    0.188514   -0.074217    11.00000    0.02764
O9    2    0.253222   -0.039078   -0.230187    11.00000    0.04666
O10   2    0.000000    0.244150   -0.250000    10.50000    0.06607
O11   2    0.236870    0.076324    0.090848    11.00000    0.02940
O12   2    0.415746    0.574306   -0.143080    11.00000    0.05927
O13   2    0.450497   -0.081256   -0.125431    11.00000    0.04875
HKLF    4
 
REM  SHELXL INS file, from DIRDIF output for EG1SCR
REM R1 =  0.1508 for   1783 Fo > 4sig(Fo)  and  0.1544 for all   1874 data
REM     48 parameters refined using      0 restraints
 
END

After checking that the result is reasonable, set the heavy Si atoms refining anisotropically using the ANIS command. (e.g., ANIS SI3 SI4 SI5)

TITL SHELXL INS file, from DIRDIF output for EG1SCR
CELL   0.71073   13.38600  7.42300 15.13400  90.0000 107.7100  90.0000
ZERR   4          0.00200  0.00200  0.00100   0.0010   0.0500   0.0500
LATT   7
SYMM     -X   ,     Y   , 1/2-Z
SFAC  CS  O   SI  TI
UNIT      8   60   24    4
L.S.  9
BOND
ANIS SI3 SI4 SI5
FMAP  2
PLAN  10
WGHT    0.100000
FVAR       0.19086
CS1   1    0.117273    0.241538    0.409945    11.00000    0.01793    0.03219 =
         0.04174   -0.00994    0.00096    0.00008
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01523    0.01518 =
         0.01076    0.00395    0.00350   -0.00115
SI3   3   -0.002755    0.231644   -0.146764    11.00000    0.01281
SI4   3    0.324445    0.428725   -0.166448    11.00000    0.02078
SI5   3    0.345283    0.040753   -0.144349    11.00000    0.02063
O6    2    0.380806    0.237015   -0.171688    11.00000    0.03363
O7    2    0.302621    0.051421   -0.056591    11.00000    0.02191
O8    2    0.108595    0.190086   -0.075372    11.00000    0.02977
O9    2    0.253002   -0.038443   -0.231263    11.00000    0.04516
O10   2    0.000000    0.246029   -0.250000    10.50000    0.07364
O11   2    0.236263    0.076421    0.091812    11.00000    0.03069
O12   2    0.415762    0.573198   -0.143304    11.00000    0.06339
O13   2    0.449321   -0.083765   -0.124798    11.00000    0.05232
HKLF    4
 
REM  SHELXL INS file, from DIRDIF output for EG1SCR
REM R1 =  0.0850 for   1783 Fo > 4sig(Fo)  and  0.0879 for all   1874 data
REM     58 parameters refined using      0 restraints
 
END

After checking that the result is reasonable, set the remaining atoms (the Oxygens) refining anisotropically using the ANIS command. (e.g., ANIS)

TITL SHELXL INS file, from DIRDIF output for EG1SCR
CELL   0.71073   13.38600  7.42300 15.13400  90.0000 107.7100  90.0000
ZERR   4          0.00200  0.00200  0.00100   0.0010   0.0500   0.0500
LATT   7
SYMM     -X   ,     Y   , 1/2-Z
SFAC  CS  O   SI  TI
UNIT      8   60   24    4
L.S.  9
BOND
ANIS
FMAP  2
PLAN  10
WGHT    0.100000
FVAR       0.19064
CS1   1    0.117251    0.241524    0.409953    11.00000    0.01729    0.03302 =
         0.04096   -0.00996    0.00060    0.00014
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01381    0.01640 =
         0.01041    0.00368    0.00285   -0.00108
SI3   3   -0.002884    0.231826   -0.146687    11.00000    0.01089    0.01552 =
         0.01232    0.00152    0.00319   -0.00389
SI4   3    0.324718    0.428033   -0.166340    11.00000    0.01289    0.03511 =
         0.01554    0.01156   -0.00048   -0.00885
SI5   3    0.344785    0.040247   -0.144441    11.00000    0.01201    0.03711 =
         0.01397   -0.00645    0.00018    0.00289
O6    2    0.380780    0.235843   -0.172454    11.00000    0.03249
O7    2    0.302476    0.052071   -0.056874    11.00000    0.02128
O8    2    0.108585    0.190819   -0.075395    11.00000    0.03035
O9    2    0.253319   -0.038318   -0.231468    11.00000    0.04662
O10   2    0.000000    0.246251   -0.250000    10.50000    0.07335
O11   2    0.236073    0.077522    0.091696    11.00000    0.02842
O12   2    0.416150    0.576791   -0.142692    11.00000    0.06225
O13   2    0.449002   -0.085954   -0.124223    11.00000    0.04851
HKLF    4
 
REM  SHELXL INS file, from DIRDIF output for EG1SCR
REM R1 =  0.0655 for   1783 Fo > 4sig(Fo)  and  0.0677 for all   1874 data
REM     73 parameters refined using      0 restraints
 
END

After checking that the result is reasonable, this may be a good time to refine on the Extinction (using the EXTI command) and use the suggested Shelx Weighting values that are usually located at the bottom of the results file.

TITL SHELXL INS file, from DIRDIF output for EG1SCR
CELL   0.71073   13.38600  7.42300 15.13400  90.0000 107.7100  90.0000
ZERR   4          0.00200  0.00200  0.00100   0.0010   0.0500   0.0500
LATT   7
SYMM     -X   ,     Y   , 1/2-Z
SFAC  CS  O   SI  TI
UNIT      8   60   24    4
L.S.  9
BOND
EXTI
FMAP  2
PLAN  10
WGHT      0.0542     19.0467
FVAR       0.19225
CS1   1    0.117275    0.241605    0.409937    11.00000    0.01732    0.03338 =
         0.04108   -0.01012    0.00093    0.00014
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01347    0.01722 =
         0.01017    0.00380    0.00321   -0.00076
SI3   3   -0.002806    0.231736   -0.146612    11.00000    0.01004    0.01733 =
         0.01294    0.00144    0.00316   -0.00394
SI4   3    0.324831    0.428613   -0.166328    11.00000    0.01297    0.03539 =
         0.01563    0.01249    0.00015   -0.00756
SI5   3    0.344677    0.039975   -0.144490    11.00000    0.01218    0.03822 =
         0.01440   -0.00624    0.00074    0.00330
O6    2    0.381252    0.235101   -0.172214    11.00000    0.02110    0.05526 =
         0.02970    0.00915    0.01571    0.00416
O7    2    0.302815    0.052230   -0.056678    11.00000    0.03182    0.02102 =
         0.01566    0.00227    0.00786    0.00375
O8    2    0.108563    0.190839   -0.075009    11.00000    0.01600    0.04174 =
         0.03020   -0.00613   -0.00232   -0.00079
O9    2    0.253216   -0.040035   -0.231202    11.00000    0.02683    0.09931 =
         0.01841   -0.01588   -0.00174   -0.01244
O10   2    0.000000    0.246165   -0.250000    10.50000    0.04954    0.15928 =
         0.02659    0.00000    0.01398    0.00000
O11   2    0.236012    0.076956    0.091479    11.00000    0.04389    0.02471 =
         0.02532    0.00789    0.01677    0.00148
O12   2    0.416550    0.573642   -0.142456    11.00000    0.03549    0.07254 =
         0.10190    0.05914   -0.02439   -0.03412
O13   2    0.448334   -0.084071   -0.123133    11.00000    0.02201    0.06592 =
         0.07065   -0.03911   -0.00896    0.01638
HKLF    4
 
REM  SHELXL INS file, from DIRDIF output for EG1SCR
REM R1 =  0.0417 for   1783 Fo > 4sig(Fo)  and  0.0439 for all   1874 data
REM    111 parameters refined using      0 restraints
 
END
 
WGHT      0.0542     19.0467

In theory, the refinement is now finished, ready for publication. However, now is the time to be checking the resulting structure through various structure checking programs such as Platon. Platon will warn you (in this case) of two oxygen thermals worth checking. Looking at these with Fourier Difference maps will give an indication of atom-splitting via the peanut shapes of the electron density contours of the two Oxygens. Platon and related software may also be able to tell you of extra symmetry you missed, and other properties or possible problems with the structure. The shelx Q peaks may tell you of missing atoms. This was not touched on as in this "contrived" example, DIRDIF found all the atoms and there are no hydrogens in this structure.

TITL SHELXL INS file, from DIRDIF output for EG1SCR
CELL   0.71073   13.38600  7.42300 15.13400  90.0000 107.7100  90.0000
ZERR   4          0.00200  0.00200  0.00100   0.0010   0.0500   0.0500
LATT   7
SYMM     -X   ,     Y   , 1/2-Z
SFAC  CS  O   SI  TI
UNIT      8   60   24    4
L.S.  9
BOND
FMAP  2
PLAN  10
WGHT    0.054200   19.046701
EXTI    0.002180
FVAR       0.19625
CS1   1    0.117273    0.241622    0.409959    11.00000    0.01792    0.03420 =
         0.04130   -0.01004    0.00113    0.00019
TI1   4    0.250000    0.250000    0.000000    10.50000    0.01445    0.01780 =
         0.01083    0.00377    0.00358   -0.00086
SI3   3   -0.002850    0.231760   -0.146670    11.00000    0.01073    0.01790 =
         0.01328    0.00098    0.00304   -0.00397
SI4   3    0.324826    0.428730   -0.166334    11.00000    0.01315    0.03672 =
         0.01582    0.01255    0.00045   -0.00745
SI5   3    0.344540    0.039870   -0.144653    11.00000    0.01299    0.03883 =
         0.01489   -0.00628    0.00128    0.00302
O6    2    0.380975    0.235421   -0.172287    11.00000    0.02190    0.05659 =
         0.02766    0.00830    0.01555    0.00362
O7    2    0.302698    0.052618   -0.056965    11.00000    0.02980    0.02189 =
         0.01655    0.00234    0.00660    0.00331
O8    2    0.108405    0.191355   -0.075251    11.00000    0.01676    0.04391 =
         0.02971   -0.00586   -0.00292   -0.00126
O9    2    0.253049   -0.039058   -0.231302    11.00000    0.02877    0.09761 =
         0.01937   -0.01713   -0.00156   -0.01240
O10   2    0.000000    0.245797   -0.250000    10.50000    0.04571    0.16386 =
         0.02423    0.00000    0.01356    0.00000
O11   2    0.235750    0.076728    0.091600    11.00000    0.04878    0.02347 =
         0.02638    0.00816    0.01979    0.00300
O12   2    0.417099    0.573272   -0.142384    11.00000    0.03519    0.07143 =
         0.10127    0.05715   -0.02008   -0.03221
O13   2    0.448287   -0.083566   -0.122844    11.00000    0.02139    0.06463 =
         0.07344   -0.04017   -0.00726    0.01420
HKLF    4
 
REM  SHELXL INS file, from DIRDIF output for EG1SCR
REM R1 =  0.0367 for   1783 Fo > 4sig(Fo)  and  0.0389 for all   1874 data
REM    112 parameters refined using      0 restraints
 
END
 
WGHT      0.0394     19.4756
REM Highest difference peak  1.409,  deepest hole -1.589,  1-sigma level  0.199
Q1    1   0.1040  0.3201  0.3864  11.00000  0.05    1.41
Q2    1   0.0888  0.2196  0.4354  11.00000  0.05    1.19
Q3    1   0.4235  0.5463 -0.1719  11.00000  0.05    0.99
Q4    1   0.0000  0.1776 -0.2500  10.50000  0.05    0.98
Q5    1   0.4180  0.6135 -0.1222  11.00000  0.05    0.96
Q6    1   0.0000  0.3109 -0.2500  10.50000  0.05    0.89
Q7    1   0.1476  0.2388  0.4223  11.00000  0.05    0.88
Q8    1   0.4269 -0.0688 -0.1695  11.00000  0.05    0.87
Q9    1   0.4361 -0.1328 -0.1099  11.00000  0.05    0.79
Q10   1   0.0295  0.2385  0.3808  11.00000  0.05    0.71

Structure Viewed in Cameron

Output Window from Platon

[Tutorials page] | [WinGX Tutorial Index]

[CCP14 Home: (Frames \| No Frames)] CCP14 Mirrors: [UK] \| [CA] \| [US] \| [AU]	What's New	Introduction	Site Map
Search the CCP14	Download Programs What do you want to do? (lists of software by crystallographic method)	Tutorials	Solutions

(This Webpage Page in No Frames Mode)

If you have any queries or comments, please feel free to contact the CCP14