@cm-beilstein
Since the notebook does not appear to allow comments I have moved them here:
|
" elif num_neighbors == 1:\n", |
|
" return \"1-bond\", metal_atom\n", |
Instead of 1-bond the correct term for this coordination mode is terminal
|
" if np.isclose(angle, 180, atol=10): \n", |
|
" return \"linear\", metal_atom\n", |
|
" else:\n", |
A molecule is linear only when the angle is exactly 180°. One needs to discuss with our colleagues from CCDC about what the average inaccurancy of bond angles in the CSD is - but a tolerance of atol=10 is clearly too high
|
" if all(np.isclose(a, 120, atol=15) for a in angles): \n", |
|
" return \"trigonal planar\", metal_atom\n", |
The requirement for a trigonal-planar structure is NOT that the three angles are all 120° but that the three ligand atoms and the metal are all in one plane (+/- a bit of tolerance). For example, all atoms planar and angles of 110°, 130° and 120° would also be trigonal-planar (and even more extreme distorted cases as angles of 90°, 100° and 170°).
The only special case is where all four atoms of MX3 are in one plane that is not trigonal-planar is when two angles are exactly 90° and the third one is extactly 180° which is "T-shaped"
All other 3-coordinated structures which are not T-shaped and not trigonal-planar are trigonal-pyramidal (and not unknown), therefore
|
" elif (np.isclose(avg_angle, 107, atol=5) or np.isclose(avg_angle, 109.5, atol=5)) and not is_planar:\n", |
|
" return \"trigonal pyramid\", metal_atom # Highly likely given the angle range\n", |
is too narrow a definition as trigonal-planar cannot be identified from the L-M-L' angle but only from the fact that the metal atom is not in the same plane as the three ligands and
|
" # print(f\"Angles: {[f'{a:.2f}' for a in angles]} degrees\")\n", |
|
" return \"3-unknown\", metal_atom\n", |
is simply incorrect.
@cm-beilstein
Since the notebook does not appear to allow comments I have moved them here:
InChI/INCHI-1-DOC/Notebooks/Molecular_inorganics/Geometries/geometry_identify_v1.ipynb
Lines 320 to 321 in 5b5bd93
Instead of
1-bondthe correct term for this coordination mode isterminalInChI/INCHI-1-DOC/Notebooks/Molecular_inorganics/Geometries/geometry_identify_v1.ipynb
Lines 326 to 328 in 5b5bd93
A molecule is linear only when the angle is exactly 180°. One needs to discuss with our colleagues from CCDC about what the average inaccurancy of bond angles in the CSD is - but a tolerance of
atol=10is clearly too highInChI/INCHI-1-DOC/Notebooks/Molecular_inorganics/Geometries/geometry_identify_v1.ipynb
Lines 349 to 350 in 5b5bd93
The requirement for a trigonal-planar structure is NOT that the three angles are all 120° but that the three ligand atoms and the metal are all in one plane (+/- a bit of tolerance). For example, all atoms planar and angles of 110°, 130° and 120° would also be trigonal-planar (and even more extreme distorted cases as angles of 90°, 100° and 170°).
The only special case is where all four atoms of MX3 are in one plane that is not trigonal-planar is when two angles are exactly 90° and the third one is extactly 180° which is "T-shaped"
All other 3-coordinated structures which are not T-shaped and not trigonal-planar are trigonal-pyramidal (and not unknown), therefore
InChI/INCHI-1-DOC/Notebooks/Molecular_inorganics/Geometries/geometry_identify_v1.ipynb
Lines 352 to 353 in 5b5bd93
is too narrow a definition as trigonal-planar cannot be identified from the L-M-L' angle but only from the fact that the metal atom is not in the same plane as the three ligands and
InChI/INCHI-1-DOC/Notebooks/Molecular_inorganics/Geometries/geometry_identify_v1.ipynb
Lines 360 to 361 in 5b5bd93
is simply incorrect.