dearomatization of coordinated benzene

User 870ab5b546

08-05-2013 21:37:10

I have a compound that has a benzene ring coordinated to a metal through a multicenter group involving two of the six C atoms.  When JChem dearomatizes the compound, the two C atoms that are coordinated to the metal are connected by a single bond.  Really, they should be connected by a double bond.  Could you modify JChem's dearomatization algorithm to take account of coordinate bonds and to place the double bonds where the multicenter groups are located, if possible?


before processing: *[Zr](*)*.c1cccc1.c1cccc1.c1ccccc1 |m:0:4.5.6.7.8,2:9.10.11.12.13,3:14.15,C:0.0,2.1,3.2|
after dearomatizing: *[Zr](*)*.c1cccc1.c1cccc1.C1=CC=CC=C1 |c:13,15,17,m:0:4.5.6.7.8,2:9.10.11.12.13,3:14.19,C:0.0,2.1,3.2|

User 870ab5b546

10-05-2013 15:36:13

While we're at it, when aromatizing a compound, if a subset of a potential aromatic group's atoms are coordinated to a metal, but not all the atoms are, I think JChem should *not* aromatize the group.

ChemAxon a202a732bf

14-05-2013 15:22:11

Hi Bob,


regarding the dearomatization issue: I have tried it with several versions, but could not reproduce the problem, after dearomatization I always got the coordinate bond connected to the center of a double bond, the SMILES format of the result is:


*[Zr](*)*.c1cccc1.c1cccc1.C1=CC=CC=C1 |c:13,15,17,m:0:4.5.6.7.8,2:9.10.11.12.13,3:14.15,C:0.0,2.1,3.2|


What version of Marvin or JChem are you using?


Regarding the aromatization issue: thank you for reporting, we will discuss it with our chemists.


Regards,


Zsuzsa


	
	

User 870ab5b546

14-05-2013 15:37:07

We're using JChem 5.11.5.  I think the atom numbering makes a difference as to where the double bonds appear in the dearomatization.  The conversion to CXSMILES may have reordered the atom numbers.  Try dearomatizing this:


<?xml version="1.0"?>
<cml version="ChemAxon file format v5.10.0, generated by v5.10.2">
<MDocument>
<MChemicalStruct>
<molecule molID="m1">
<atomArray atomID="a1 a2 a3 a4 a5 a6 a7 a8 a9 a10 a11 a12 a13 a14 a15 a16 a17 a18 a19 a20" elementType="X Zr X C C C C C C C C C C C C C C C C X" sgroupRef="0 0 0 sg2 sg2 sg2 sg2 sg2 sg1 sg1 sg1 sg1 sg1 0 sg3 sg3 0 0 0 0" x2="3.593333333333333 2.053333333333334 3.593333333333335 3.5933333333333324 4.839219504670751 4.363333333333333 2.8233333333333333 2.3474471619959134 3.593333333333335 4.839219504670754 4.363333333333336 2.823333333333336 2.347447161995916 -1.693506227580685 -0.3598271057526472 -0.3598271057526503 -1.6935062275806871 -3.0271853494087217 -3.0271853494087204 -0.35982710575264876" y2="2.6673582436560705 0.0 -2.66735824365607 3.9773604885182126 3.072171199987805 1.6075441648932678 1.6075441648932667 3.072171199987803 -1.3573559987939285 -2.2625452873243366 -3.7271723224188733 -3.7271723224188746 -2.262545287324338 0.385000000000002 -0.385000000000002 -1.925000000000002 -2.695 -1.9249999999999978 -0.384999999999998 -1.155000000000002"/>
<bondArray>
<bond atomRefs2="a1 a2" convention="cxn:coord"/>
<bond atomRefs2="a3 a2" convention="cxn:coord"/>
<bond atomRefs2="a4 a5" order="A"/>
<bond atomRefs2="a5 a6" order="A"/>
<bond atomRefs2="a6 a7" order="A"/>
<bond atomRefs2="a7 a8" order="A"/>
<bond atomRefs2="a4 a8" order="A"/>
<bond atomRefs2="a9 a10" order="A"/>
<bond atomRefs2="a10 a11" order="A"/>
<bond atomRefs2="a11 a12" order="A"/>
<bond atomRefs2="a12 a13" order="A"/>
<bond atomRefs2="a9 a13" order="A"/>
<bond atomRefs2="a14 a15" order="A"/>
<bond atomRefs2="a15 a16" order="A"/>
<bond atomRefs2="a16 a17" order="A"/>
<bond atomRefs2="a17 a18" order="A"/>
<bond atomRefs2="a18 a19" order="A"/>
<bond atomRefs2="a14 a19" order="A"/>
<bond atomRefs2="a20 a2" convention="cxn:coord"/>
</bondArray>
<molecule id="sg1" role="MulticenterSgroup" molID="m2" atomRefs="a9 a10 a11 a12 a13" center="a3"/>
<molecule id="sg2" role="MulticenterSgroup" molID="m3" atomRefs="a4 a5 a6 a7 a8" center="a1"/>
<molecule id="sg3" role="MulticenterSgroup" molID="m4" atomRefs="a15 a16" center="a20"/>
</molecule>
</MChemicalStruct>
</MDocument>
</cml>

ChemAxon a202a732bf

14-05-2013 16:22:54

Hi Bob,


yes, that does the trick, I could reproduce it, we'll check it out and get back to you.


Regards,


Zsuzsa

ChemAxon a202a732bf

15-05-2013 13:49:49

Hi Bob,


I have talked to chemists about the aromatization/dearomatization problem with coordinate bonds. They revealed that connecting a coordinate bond to an aromatic part of the molecule has meaning only if it is connected to the whole aromatic part, connecting it to one bond is not meaningful, thus the aromatization and dearomatization process does not support it.


It may be considered that applying a Structure Checker-Fixer for these kind of invalid structures, which would reconnect the coordinate bond to the whole aromatic part instead of the bond.



Best regards,


Zsuzsa


User 870ab5b546

15-05-2013 14:44:56

I would direct your chemists to Fei Ding and W. Dean Harman, J. Am. Chem. Soc.2004126 (42), pp 13752–13756, for one example of an η2-benzene complex.  


I agree with your chemists that aromatization of an η2-benzene is not meaningful; that's why I suggested you change your algorithm so that an η2-benzene would not be aromatized.  On the other hand, if a user draws a benzene in the aromatic form with an η2 linkage to a metal, and then dearomatization is done, would it be so terrible to choose to put one of the double bonds between the two atoms coordinated to the metal?  The same for η4-benzene, BTW.

ChemAxon 25dcd765a3

16-05-2013 07:13:36

Dear Bob, 


thank you for the suggested article.I will deifinitely check it and ask some question.


"would it be so terrible to choose to put one of the double bonds between the two atoms coordinated to the metal"


Definitely not, if it can be done with no slowdown in dearomatization. If it slows down the dearomatization algorithm with more than 3% than it should go to a structure checker/fixer action. 


On the other hand would this structure checker/fixer solve your problem?

User 870ab5b546

28-08-2013 17:41:10

Sorry, this conversation fell off my radar.  


It seems silly to write a StructureChecker for the very rare instance of an aromatic η2-benzene that has been dearomatized and the double bonds put in the wrong place, and then to subject every single molecule to this StructureChecker after dearomatizing.  Plus, it's not a trivial operation to alternate the double bonds in a benzene group, is it?


Better just to have the aromatizer put the double bonds in the right place in the first place.  It just needs to check if any of the aromatic atoms are part of a multicenter group.

ChemAxon 25dcd765a3

03-09-2013 14:28:16

"Better just to have the aromatizer put the double bonds in the right place in the first place.  It just needs to check if any of the aromatic atoms are part of a multicenter group."


Sounds feasible, let's see the answer of the developer of the dearomatizer code.