[Molvis-list] Errors in Chem Viz

Roger Sayle roger at eyesopen.com
Sun Sep 17 13:37:37 EDT 2006

Hi Mario,

On Fri, 15 Sep 2006, Mario Valle wrote:
> I just started a page on "errors in chemistry visualization tools and
> chemistry data". It is available on:
> <http://www.cscs.ch/~mvalle/ChemViz/errors/index.html>

Where you write:
> > but the line+dashed line meaning still eludes me.

The line and dash line representation is commonly used in molecular
graphics to indicate a resonant or aromatic bond.  Effectively, a
bond with order greater than one and less than two.

To paraphrase Democritos, "Nothings exists but atoms and space; all
else is opinion".  This particularly true of bonds and bond orders.
Chemists traditionally resort to "valence-bond theory" to rationalize
the interactions between neighboring atoms, assigning them as single,
double, triple and even quadruple bonds.  Unfortunately, like most
models, this view has its limitations.  In quantum mechanics the
number of electrons shared between neighboring atoms (or over a set
of atoms) is not a discrete integer value, but a continuous value.

Hence, whilst assigning single and double covalent bonds to molecules
can be used to rationalize/explain many of the properties and behaviors
of organic molecules, its only part of the story.  The discretization
(or rounding) of electron sharing to bonds with orders of precisely
one or two, can have artifacts.

The classical example of this kind of problem is benzene, where the
reactivity and geometry of benzene indicate that all six bonds appear
equivalent with an effective bond order of 1.5 each.  To extend
classical valence-bond theory to capture these cases, the concept of
a resonant, aromatic or factional bond can be used.  In depictions
and molecular graphics representation, this is shown as a dashed line.
A aromatic bond, with bond order 1.5, is shown as a solid line (1) and
a dashed line (half).   Hence a common image of benzene has single and
dashed lines everywhere.  Likewise, hydrogen bonds (with effective bond
orders less than one) are often shown as dashed lines (with out a solid

An approximate way to think about this is that the dashed line has 0.5
of a bond order.  In your five valence carbon example, you'll notice
that the bond order's effectively sum to four, suitable for a neutral
four valent carbon.   Of course, to think of resonant bonds as being
exactly 0.5 leads to the same artifacts as above.  The aromatic bonds
in furan and pyrrole (or on the oxygen in your example) can make the
atom look like it has a strange valence.  Best to think of solid-dashed
lines a more than bond order one, and less than bond order two.

Another example is the iron co-ordination in heme where the iron
shares two bonds, with four nighboring/co-ordinating nitrogens equally.

To contribute to your interesting set of examples exploring the
boundaries of chemistry models, might I suggest that you add the
perchlorate ligand in 1jfv.  Perchlorate is one chlorine atom
surrounded by four oxygens, the entire molecule having a net charge
of -1.  In reality, the four oxygens contribute/shared electrons
equally, and the molecule is symmetric, but it is difficult to decribe
this in traditional valence bond theory.  Hence most people approximate
it as the asymmetric:


instead of either of

or	[Cl+3]([O-])([O-])([O-])[O-]

{forgive the SMILES notation but I'm sure you can figure out what I mean}

This representational artifact then affects the parameters used in
x-ray crystallographic refinement.  Such that in pdb1jfv.ent we end
up with a perchlorate anion with three short (double) bonds, and one long
(single) bond.  It's an example of the limitations of the models we
use to rationalize the world we see around us affecting the experimental
observations themselves.

I hope this helps,

Roger Sayle,                         E-mail: roger at eyesopen.com
OpenEye Scientific Software,         WWW: http://www.eyesopen.com/
Suite 1107, 3600 Cerrillos Road,     Tel: (+1) 505-473-7385
Santa Fe, New Mexico, 87507.         Fax: (+1) 505-473-0833

More information about the Molvis-list mailing list