Talk:Hexamethyltungsten

If the groups attached to the tungsten were somehow unique, how many stereoisomers would it have? I'm not terribly experienced on the subject, but this seems non-obvious considering that most stereochemical arrangements have only *four* unique groups around the chiral atom. Deranged bulbasaur 07:04, 13 February 2007 (UTC)

I'm certainly no expert on the stereochemical considerations of trigonal prismatic molecules, but I think we can reason through this. In the case of hexamethyl tungsten, as each substituent contains four atoms itself, the total number of possible stereoisomers would be quite large. For the sake of simplifying things, let's assume the hydrogen atoms on each methyl group are identical and indistinguishable, so that we can consider each methyl group to be analogous to a single atom instead of four. Think of WCl6, which has octahedral geometry but only chlorine atoms to consider. There are two possible configurations for the top three groups in hexamethyl tungsten, if we are to consider them being somehow distinguishable through perhaps isotopic labelling of the carbon atoms as C1, C2, and C3. You can either iterate clockwise or counter-clockwise through all three groups. For each of these two configurations, let's call them R1 and S1, the bottom three groups can take on any of three possible rotations around the principal axis of symmetry (the D3h axis) and still retain the original trigonal prismatic structure. On top of that, there are two possible stereochemical configurations for those bottom three methyl groups, same as the top: R2 and S2. So, we have 6 configurations for R1 and S1 each. Thus I believe you get 12 possible configurations, assuming 6 unique methyl groups and discounting the hydrogen atoms uniqueness. If you start considering the possibility of labelling each hydrogen atom on the methyl groups, the problem becomes vastly more complex. Perhaps someone more versed on trigonal prismatic stereochemistry can weigh in. --Tyler Matthews 06:02, 19 February 2007 (UTC)

Number of isomers
Update: Prof. Gregory S. Girolami at UIUC sent me a link regarding calculating the number of isomers for point group symmetries of complex compounds. The answer is decidedly not 12; there are 60 (!) isomers for a trigonal prismatic structure. I. V. Krivoshei and B. Yu. Vvedenskii. Calculation of the number of isomers for the basic geometrical models of mononuclear complex compounds. Theoretical and Experimental Chemistry 1967, 3(4), 296-300. —The preceding unsigned comment was added by Tyler Matthews (talk • contribs) 07:22, 25 February 2007 (UTC). Tyler Matthews 07:22, 25 February 2007 (UTC)
 * That's interesting. I'd forgotten I asked, but then I come back to see I've gotten a rather surprising answer. Thanks for digging it up. Deranged bulbasaur 14:31, 27 May 2007 (UTC)

Which symmetry group?
In the section Molecular Geometry, the first line says C3v, but near the end of the second paragraph the article says D3h with a reference. Which is correct please? The figures appear D3h to me with a horizontal reflection plane. However in the article Octahedral molecular geometry, W(CH3)6 was given as an example of D3h but has today been changed to C3v. Is this change correct? Dirac66 17:31, 7 November 2007 (UTC)


 * The most recent X-ray results say C3v, a distorted trigonal prism (Seppelt, K.; Pfennig, V.Science1996,271, 626-8. and Kleinhenz, S.; Pfennig, V.; Seppelt, K.;Chem. Eur. J.1998,4, 1687-91.) This structure also agrees with various theoretical studies. I haven't looked at the electron diffraction paper from 1990 that says D3h; I'll look at it tomorrow. --Itub 18:48, 7 November 2007 (UTC)


 * I couldn't wait till tomorrow. ;-) I read the 1990 paper, and it says "We conclude, therefore, that the coordination geometry of hexamethyltungsten is trigonalprismatic D3h or slightly deformed to C3v symmetry." They mostly stick to D3h because it is the simplest model that fits the data, as it requires fewer parameters. Both geometries, however, fit their data equally well. The X-ray papers from 1996 and 1998 are more conclusive, because they can resolve the differences in the W-C bond lengths and C-W-C angles. --Itub 19:24, 7 November 2007 (UTC)

Thanks. I was in the library today and read the 1990 and 1996 papers also. The 1990 seems chiefly concerned to show that it is not Oh. So I now agree that your edit to Octahedral molecular geometry is correct, and I also see that this article (Hexamethyl tungsten) needs a little revision to be consistent with the papers. Dirac66 03:50, 8 November 2007 (UTC)