Least-squares-fit overlays

6/7: Nine tin complexes

For this example we'll superimpose nine closely related tin complexes from a recent paper in the Journal of Molecular Structure.

Here's what the structures look like:

As with the previous example, these all have the same ligand attached to their central Sn atom, with two other groups attached. These are: two methyl (1), two butyl (2), two octyl (3), two benzyl (4), butyl and chloro (5), phenyl and chloro (6), butyl and azido (7), butyl and thiocyanato (8), and two chloro (9). For clarity, we'll pare these groups down to the first atom. One additional complication is that the structure of (9) straddles the mirror plane of Pnma, so we'll have to generate the full model first.

The CIFs and other files for this example are available here.

For structures 1-8, over-rip.py does a good job. Structures 2, 5, 8 have disorder so we'll take the major component only. Solvent molecules present in 5 and 9 will also need to be manually edited out. For structure 9, 'GROW' in Shelxtl XP will generate the whole molecule, which may then be written to a file using 'ORTH' or 'FILE'd and converted to CIF using Mercury. The ORTH-generated file is easily edited to .xyz format. Other methods would also work. Over-rip.py is run in the usual way, viz:

./over-rip.py  or  python over-rip.py 

The result is nine .xyz files, copies of which are here.

The nine Sn complexes each have 33 atoms (after editing the ligands). Since there could be torsion of the phenyl rings, only their ipso carbons make sense to include in the fit. Similarly, since not all of the pendant groups bond to Sn via carbon, it would be a mistake to include them in the fit. Thus, we'll feed over-rot.py  an atom list using an over-rot-atoms.txt file with the following atoms:

Sn1 O1 O2 N1 N2 N3 N4 N5 C1 C2 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 

With the nine .xyz files in the same directory as over-rot.py, a successful run using:

./over-rot.py or python over-rot.py 

Leads to the following output, which is written to a log file 'over-rot.log'.

Using reference structure: 1.xyz

Using the following atoms for fitting: Sn1, O1, O2, N1, N2, N3, N4, N5, C1, C2, C8, C9, C10, C11, C12, C13, C14, C15, C16, C17, C18
All specified atoms found in reference structure

=== Pre-alignment Analysis ===

Targeted Atoms RMSD Matrix (Å):
-----------------------------------------------------------------------------------
             1       2       3       4       5       6       7       8       9   
-----------------------------------------------------------------------------------
 1 1.xyz   0.000   8.553   6.361   7.726  12.708  12.701  10.485  12.654  10.578
 2 2.xyz   8.553   0.000   6.417  12.636   9.511   9.197   7.434   9.340   7.950
 3 3.xyz   6.361   6.417   0.000  10.101  11.222  11.620   8.256   9.727  10.125
 4 4.xyz   7.726  12.636  10.101   0.000  18.638  18.340  13.969  17.633  15.778
 5 5.xyz  12.708   9.511  11.222  18.638   0.000   4.415   9.506   5.772   7.262
 6 6.xyz  12.701   9.197  11.620  18.340   4.415   0.000   9.139   6.086   6.318
 7 7.xyz  10.485   7.434   8.256  13.969   9.506   9.139   0.000   7.293   9.838
 8 8.xyz  12.654   9.340   9.727  17.633   5.772   6.086   7.293   0.000  10.023
 9 9.xyz  10.578   7.950  10.125  15.778   7.262   6.318   9.838  10.023   0.000
-----------------------------------------------------------------------------------

Global RMSD (pre-fitted atoms): 4.1243 Å

=== Alignment ===
Note: Inverted coordinates give better RMSD (0.1144 Å vs 0.1760 Å)
Use inverted coordinates for this structure? (y/n): Aligned 2.xyz to reference | Fit RMSD: 0.1144 Å | Rotation: 93.97° (inverted)
Note: Inverted coordinates give better RMSD (0.0969 Å vs 0.1081 Å)
Use inverted coordinates for this structure? (y/n): Aligned 3.xyz to reference | Fit RMSD: 0.0969 Å | Rotation: 120.18° (inverted)
Aligned 4.xyz to reference | Fit RMSD: 0.1642 Å | Rotation: 142.31° 
Note: Inverted coordinates give better RMSD (0.1437 Å vs 0.2101 Å)
Use inverted coordinates for this structure? (y/n): Aligned 5.xyz to reference | Fit RMSD: 0.1437 Å | Rotation: 171.53° (inverted)
Note: Inverted coordinates give better RMSD (0.2199 Å vs 0.2567 Å)
Use inverted coordinates for this structure? (y/n): Aligned 6.xyz to reference | Fit RMSD: 0.2199 Å | Rotation: 99.51° (inverted)
Note: Inverted coordinates give better RMSD (0.1710 Å vs 0.1992 Å)
Use inverted coordinates for this structure? (y/n): Aligned 7.xyz to reference | Fit RMSD: 0.1710 Å | Rotation: 100.99° (inverted)
Note: Inverted coordinates give better RMSD (0.1734 Å vs 0.2092 Å)
Use inverted coordinates for this structure? (y/n): Aligned 8.xyz to reference | Fit RMSD: 0.1734 Å | Rotation: 110.20° (inverted)
Aligned 9.xyz to reference | Fit RMSD: 0.2096 Å | Rotation: 167.68° 

Overall Rotation Angles:
----------------------------------------------------
Index  Filename                 Angle (°)   Inverted
----------------------------------------------------
 1      1.xyz                       0.00         No
 2      2.xyz                      93.97        Yes
 3      3.xyz                     120.18        Yes
 4      4.xyz                     142.31         No
 5      5.xyz                     171.53        Yes
 6      6.xyz                      99.51        Yes
 7      7.xyz                     100.99        Yes
 8      8.xyz                     110.20        Yes
 9      9.xyz                     167.68         No
----------------------------------------------------

All aligned structures written to over-rot.rot

=== Post-alignment Analysis ===

Rotated Atoms RMSD Matrix (Å):
-----------------------------------------------------------------------------------
             1       2       3       4       5       6       7       8       9   
-----------------------------------------------------------------------------------
 1 1.xyz   0.000   0.114   0.097   0.164   0.144   0.220   0.171   0.173   0.210
 2 2.xyz   0.114   0.000   0.135   0.095   0.108   0.181   0.117   0.141   0.154
 3 3.xyz   0.097   0.135   0.000   0.198   0.165   0.229   0.180   0.179   0.219
 4 4.xyz   0.164   0.095   0.198   0.000   0.154   0.208   0.140   0.182   0.185
 5 5.xyz   0.144   0.108   0.165   0.154   0.000   0.127   0.068   0.084   0.092
 6 6.xyz   0.220   0.181   0.229   0.208   0.127   0.000   0.105   0.079   0.070
 7 7.xyz   0.171   0.117   0.180   0.140   0.068   0.105   0.000   0.060   0.081
 8 8.xyz   0.173   0.141   0.179   0.182   0.084   0.079   0.060   0.000   0.082
 9 9.xyz   0.210   0.154   0.219   0.185   0.092   0.070   0.081   0.082   0.000
-----------------------------------------------------------------------------------

Global RMSD (rotated atoms): 0.0578 Å
Improvement: 4.0666 Å 

In this example, six of the nine were inverted. The space-group symmetry for 6 was P 2₁2₁2₁, so one might expect inversion to be invalid. In actuality, the crystals of 6 were twinned by inversion, so the inverted form was present. Either way, the molecule of 6 is achiral and (chemically, not crystallographically) mirror symmetric, so it makes no difference. In fact, if 6 had been numbered the other way round it wouldn't have suggested inversion (likewise for 2, 3, 5, 7, 8).

Ok, on to the 3D graphics. If we feed the new .rot file to over-lay.py :

./over-rot.py or python over-rot.py 

The resulting html/JavaScript should look like this:

This example introduced yet more complexity. One of the structures had to be grown to get its full complement of atoms due to it being on a crystallographic mirror plane. Nevertheless, given proper attention to detail, the procedure works well.

---