2
$\begingroup$

How can we "draw" a coordination bond (dashed bond) in MOLFILE?

I cannot find a bond property or option that would allow me to choose such a thing.

So many drawings of organo-metallic compounds (e.g., metal ion and chelator) contain that.

How come the MOLFILE format doesn't support such a thing?

Improve this question
$\endgroup$
2
  • $\begingroup$ Generally, Lewis structures do not support dative bonds (dashed or arrows and the like); V2000 simply might be too old to have this implemented. IUPAC recommends something different for schemes for organometallic compounds, where the lines lose their strict character (and are definitively not Lewis(-like) structures). $\endgroup$
    – Martin - マーチン ♦
    Oct 18, 2021 at 23:17
  • $\begingroup$ The BIOVIA documentation about CTfiles, August 2020 edition, pdf file documents both v2000 and v3000 format, with preference for the v3000 format (p. 1). Chapter 2, p. 11 includes a table «Meaning of the values in the bond block [of v3000 format]» which just follows the atoms and their coordinates, this mentions type 9 with two sub-types about coordination, either for metal complexes, or for Lewis acids/bases. $\endgroup$
    – Buttonwood
    Oct 19, 2021 at 7:27

1 Answer 1

Reset to default
3
$\begingroup$

While currently not functional for me, RDKit's cookbook includes a relevant entry to this, though. Instead of dashes, the dative bond would be represented by an arrow pointing toward the transition metal, e.g.

enter image description here

(from RDKit's Cookbook)

in a section called Organometallics with Dative Bonds.

If you use .sdf in the v3000 format,* RDKit may be adjusted to visualize dative bonds as dotted lines, too:

enter image description here

(Esben Jannik Bjerrum, cheminformania.com)

*) The current definition of .sdf files by Biovia (August 2020, link to .pdf) describes in detail the types of bonds available in the v3000 format to the bond block (cf. p. 11). Type 9 is set for coordination with dipolar bonds of either sub-type COORD for coordination bonds in metal complexes, or DATIVE for bond between Lewis acids and bases. (There might be a concptual overlap here.) Bond type 10, by the way, offers two sub-types to handle hydrogen bonds, too.

Improve this answer
$\endgroup$
8
  • $\begingroup$ So that's not really anything explicit in the MOLFILE, but I suppose we could use the same logic they used: "Process organometallic SMILES by detecting single bonds between metals and replacing with dative bonds." So whenever in the bond table in a MOLFILE I find one of the atoms a metal, I would draw a dashed line instead of a solid line. I have my own MOLFILE renderer so I can do that. $\endgroup$
    – Gunther Schadow
    Sep 16, 2021 at 22:12
  • $\begingroup$ I noticed that PubChem also uses that convention, and it was easy enough for me to implement. Single bond with any ligand being a metal --> dashed line. $\endgroup$
    – Gunther Schadow
    Sep 17, 2021 at 1:14
  • $\begingroup$ Alex Clark published a proposal for zero order bonds as an extension to SD files. It’s supported by some software, but I don’t think there’s a list anywhere. $\endgroup$
    – Geoff Hutchison
    Sep 17, 2021 at 3:00
  • $\begingroup$ @GuntherSchadow Just out of curiosity about «I have my own MOLFILE renderer so I can do that», is this program somewhere presented in a publication, mirrored on a repository like GitHub/commercialized? An answer in lines of «No, it is an application working fine for me, not intended for publication» were acceptable, too. $\endgroup$
    – Buttonwood
    Sep 17, 2021 at 6:14
  • $\begingroup$ @GeoffHutchison I just passed a fairly recent documentation of the ctfiles (BIOVIA, August 2020) here. Page 11 (or, counting the pages in the pdf, p. 17) describes 10 different bond types (single, double, triple, etc.) without even touching the bond's configuration as either «up», «down», «either», or «none». Possibly «no bond at all» does not fit level 8, any bond order. $\endgroup$
    – Buttonwood
    Oct 18, 2021 at 17:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.