MOPAC c.i. calculation

```I am a graduate student in Organic Chemistry trying to overcome his math-
phobia and learn Physical Chemistry. Therefore I thank you in advance on
helping me in theis goal.
I have been performing calculations on systems involving Bergman Cyclization,
and I have reached a point of confusion in the study of a simple enediyne
system.
When the enediyne is converted to the 1,4 benzene diradical, I realized that
simple RHF is going to fail miserably in this respect. But on trying to
understand the complexities of the OPEN and CI commands, I ran across the
following problem.
First, I tried a simple CI calculation using the following commands
AM1 SYMMETRY BONDS VECTORS SINGLET OPEN(2,2) C.I.=(2,0) MECI
which gave a Hf of 131.77 for the diradical. Of course, since I was
unsure of the diradical's 'actual' multiplicity of the HOMO and LUMO, I
thought that perhaps a simple mixing between those two might be insufficient.
It was my (naive) assumption that mixing in more configuractions couldn't
hurt, because if the contribution was small, things wouldn't be affected.
So I tried
OPEN(2,3) C.I.=(3,0) to mix another empty orbital. This gave a Hf of
134.70 (!) and then OPEN(2,4) C.I.=(4,0) gave 135.22 (!).
The eigenvectors were very curious as follows
eigenvector     2,2             2,3             2,4
13              -10.03          -9.86           -9.92
14              -5.93           -5.09           -4.39
15              -4.70           -3.80           -3.15
16              -0.15           -1.49           -1.40
17              0.59            0.45            -1.10
18              2.68            2.41            2.31
Now assuming double filling of all the orbitals, 14 would be the HOMO. It was
my (again, naive) assumption that 14 and 15 would be (reasonably) degenerate
in a biradical species such as this, and only the 2,4 calculation seems to
come close to that.
The states reported by MECI are as follows
state   multiplicity    2,2             2,3             2,4
1       singlet         -5.40           -7.81           -9.27
2       triplet         -5.17           -7.60           -9.05
... and others.
This seems reasonable for the biradical: the triplet and singlet states should
be fairly close in energy, no? But if each subsequent state is being
stabilized, why is my total heat of formation going up?
And what EXACTLY am I looking at here?
I would be grateful if anyone could xplain exactly what evil things I have
done, or at least point me in the right direction concerning the literature
on CI and such.
And one other point: would it be a bad thing to do C.I. on say, ground state
molecules? Isn't CI adding to the description of the electronic state of the
molecule, and thus 'improving' the definition?
My adviser has suggested that CI might be a little problem for semiempirical
methods because it has been parameterized without it, and I would like a
little more information from anyone willing to point me in the right direction.
I apologize for my naivete, and thank you in advance,
--
matt
-----------------------------------------------------------------------------
|/ _  /\     Matthew Harbowy  (ikf |-at-| lithium.tn.cornell.edu)
|\ - /__\       "I'm the bear that went over the mountain"
-----------------------------------------------------------------------------
What kind of rule Can overthrow a fool and leave the land with no stain?
-Suzanne Vega
```