> Avogadro doesn't seem to correctly predict the shape of the PCl5
> molecule. I just drew the molecule and started the autoopt tool... and
> got an odd shape which doesn't seem to comply with VSEPR
The UFF implementation in Avogadro (and Open Babel) has exceptions to
help with UFF. I've tried very hard to handle 5-coordinate molecules,
and it doesn't work well. What you probably see is that the equatorial
atoms have a 90-degree angle.
What I try to do is "guess" which atoms will become axial, and set the
target angle to 90 degrees. The problem is that it cannot ensure there
will always be three 120 degree angles.
I'm open to suggestions. Consider that you start with a rough
geometry, which is probably very wrong.
Can we codify some rules to help with 5-coordinates?
P.S. I *think* these are the only problems with UFF vs. traditional
VSEPR. Please try other coordination sites and let me know. It does
handle 7- 8- and 9- coordinate molecules nicely. :-)
> I'm still studying basic chemistry.. so I wouldn't be able to give
> you any advice on making new rules to correct this.
> Are tere any problems with 2 and 3 atom molecules?
No. As far as I know, 5-coordinate is the only problem. The MMFF94
force field is recommended for general organic molecules and does have
problems with unusual coordination or elements. UFF is designed for
the entire periodic table (it stands for Universal Force Field).
P.S. Rules? In practice, van der Waals repulsion should force things
into VSEPR shapes. There are technical (programming) reasons this
doesn't work so well for PCl5.