MOPAC Question: Summary
- From: "Hr. Dr. S. Shapiro" <toukie.,at,.zui.unizh.ch>
- Subject: MOPAC Question: Summary
- Date: Wed, 10 Sep 1997 08:40:21 +0200
Some time ago I submitted a question regarding how to do a MOPAC 6.0 run
on p-tolylacetylene, which has four atoms [C(ar)-C#C-H] all in a straight
line. I am grateful for the plethora of responses that I received.
Very briefly, four suggestions were made:
(a) use the normal Z-matrix, just add "xyz" to the list of keywords.
didn't work. Maybe it's a quirk in the Win95 port of MOPAC 6.0 that I use,
but when I tried this strategy the calculation aborted with an error msg.
to the effect that the programme was very unhappy with C(ar)-C#C-H = 180
(b) alter the Z-matrix so that C(ar)-C#C-H =/= 180 degrees at the start of
the calculation. Sorry, but my port of MOPAC wasn't fooled by this. As
soon as the geometry found itself adjusted to C(ar)-C#C-H = 180 degrees, it
(c) alter the geometry of the C(ar)-C#C-H unit using dummy atoms. Indeed,
this works, but I was hoping for a slightly simpler solution that playing
around with dummy atoms if I could possibly help it.
(d) use Cartesian coordinates as the input. This was the solution I
ultimately adopted. I drew the desired structure using PCModel and saved
it as a normal MOPAC Z-matrix. I converted the input to Cartesian
coordinates using Babel 1.6 for DOS and very slightly edited it to make it
compatible with my port of MOPAC. It ran it with my "normal" roster
keywords; I did _not_ need to include "xyz" for the programme to
that the input was in Cartesian coordinates or that the calculation was to
proceed using Cartesian coordinates. When I did this, the calculation
quickly came to the desired conclusion.
Again, many thanks to the respondents who were generous enough to share
their experience and expertise with me.
P.-S.: A summary re my LUMO question is pending.