Is science sometimes in danger of getting tunnel vision? Recently published ebook author, Ian Miller, looks at other possible theories arising from data that we think we understand. Can looking problems in a different light give scientists a different perspective?

Do We Learn From Our Mistakes?

Polywater might have been an obvious error for chemistry, but I still question, what did we learn from it? My guess is, not much. What we eventually realized is that while fused silica does not dissolve in water at any appreciable rate, it does if it is on the surface of a very small capillary. Why? Is it due to the curvature of the surface, or is a micro-column of water somehow more active? A general theory here could be of great help to medicine, or to much of research into nanotechnology, but such was the scorn thrown at polywater that a potential advance of great significance was dealt with like the baby discarded with the bath water.
In previous posts I mentioned the problem of whether cyclopropane could delocalize is ring electrons into adjacent unsaturation. The textbooks say it can, and this is justified because MO theory says it can. Do you believe that? Are you still convinced when you are told that the computational programs that "settled" this issue were the same ones that asserted that polywater had very significant enhanced stability? The original MO treatment of cyclopropane was due to Walsh. His concept was that the methylene units were trigonal sp2 centres, with the third orbital of each carbon forming three-orbital overlap at the centre of the ring system. This left a p orbital on each methylene to overlap with the two p orbitals from the other methylene carbon atoms in partial side-on overlap. Since only two electrons were in the three-centre bond, there were four electrons for the three p-electron bonds, which led to two pairs for three bonds, one such bond being a "non-bond". These were obviously delocalized (assuming the model was correct in the first place) but the p orbitals were also properly aligned to overlap with adjacent p orbitals on unsaturated centres, so conjugation should follow. This was a perfectly good theory because it made predictions, however it is also imperative that such predictions were tested by observation.
There is an obvious consequence to this theory. Perhaps the biggest reason cited for cyclopropane conjugation is that a cyclopropane ring adjacent to a carbenium ion centre has an additional stabilization of about 100 kJ/mol over other comparable carbenium ions. Of course electron delocalization might be the reason for this, but if it is, then the p electrons of the cyclopropane ring must become localized in the orbitals that can overlap with the carbenium centre, at least to some extent, therefore the “non-bond” must become localized, to the same extent, in the distal bond. With less electron density in the distal bond, it should lengthen. There have been alternative MO computations, which drastically shorten the distal bond, e.g. to 143.6 pm, but significantly lengthen the vicinal bonds e.g. to 159 pm (J. Am. Chem. Soc. 1982, 104, 2605-2612) although it is far from clear why this change of bond length happens. The predicted lengthening of the vicinal bonds presumably occurs because charge in them is delocalized towards the carbenium ion, but it is unclear to me why the "non-bond" shortens.  As it happens, it is not important.  A structural study has been carried out on such a carbenium ion, and the distal bond is so considerably shortened but the vicinal bonds are not so lengthened (J. Am. Chem.Soc. 1990, 112, 8912-8920). Accordingly, the computations are wrong. The polarization theory I mentioned in previous posts is in accord with this observation: the vicinal bonds remain unchanged because nothing much changes while the distal bond shortens because the positive field allows the electrons in the bond to align better with the internuclear axis.
Now, the interesting point about this is that when the measurement was made, nobody questioned whether the Walsh MO theory might be wrong. Such is the power of established theory that even when observation brings in a result opposite to that predicted, and even when there is clear evidence (from polywater) that the computational methodology that led to this result is just plain wrong, we do not want to revisit it. Why is this? A general lack of interest in why things happen? Simple sloth? Who knows? More to the point, who cares?
Posted by Ian Miller on Apr 22, 2013 5:49 AM Europe/London

Share this |

Share to Facebook Share to Twitter Share to Linked More...

Leave a comment?

You must be signed in to leave a comment on MyRSC blogs.

Register free for an account at