[MMTK] Triple bonds?

Christopher Drost chris.drostie at gmail.com
Wed Jul 6 13:13:56 UTC 2011

Hello. I want to find vibrational modes for a couple smallish
molecules which I've been studying as part of my Master's thesis.
Perfection is not required, and it's not a huge issue for me, but I
just want to know the rough energy scale of certain phonon modes.

(1) The first question I have regards sp-hybridization. My molecule
has both C-C and C-N triple bonds, which can be something like 50-100%
stiffer than double bonds, and prefer a different bending geometry. So
I have to specify the sp-hybridization somehow. I've found
documentation for specifying these in Allinger's MM models, but I'm
still uncertain how to specify a triple bond for the Amber99 force
field as implemented in MMTK. Can anyone help with how this might be
specified? I have trouble believing that Amber99 would be ~15 years
behind the MM models.

(2) To get the rest of the program working, I built a molecular model
for benzene and made sure that I could calculate its vibrational
modes. At first it segfaulted. (!) I found out that this was because I
forgot to put in a "guess" at the structure, and the energy minimizer
crashed when it was asked to minimize a structure which doesn't start
with an initial "configuration". You might want to fix this to be a
more verbose error message. (The benzene model works up to at most a
10% error, sometimes as good as 1% for several vibrational modes.)

(3) One limitation: my molecules have a net dipole moment. (Probably
the benzene C-H bonds do as well?) I don't yet know how to calculate
the proper 'amber_charge' values in such a case. There is some
documentation about the ChargeFit class but I can't really divine how
it works, exactly. So, does anyone have an example for calculating
these amber_charge parameters, something like the way that Amber's
energy minimization algorithm can be used to calculate their

-- Chris Drost

More information about the mmtk mailing list