Im trying to optimize ochratoxine a, but when tell hyperchem to show hydrogens, and put any semi-empirical method , and leave the semi-empirical
options>Charge and Spin>total charge: 0, spin multiplicity 2, and click in geometry optimization and ok it says:
Quote:
Warning message received from node=0:
The optimized geometry may not correspond to the geometry with the minimum energy because of the half electron approximation.
So the molecule isn't in a singlet state, then i should change the total charge from 0 to something, then the spin multiplicity has to be 1, 3 or 5.
BUT, any of those numbers give me a geometry with some bonds too big, like 4.6Å, 2.17Å,etc.
Is there a way to make it show me the right calc?arsphenamine - 7-1-2012 at 12:34
Use neutral charge and singlet state.
The optimization step can be very sensitive to the initial structure.
Try generating your hyperchem input file from Avogadro after pre-optimizing the structure with MMFF94.
Check the initial bond lengths by exporting to a z-matrix file format first with Avogadro or openBabel.
With a poor starting point, ab initio methods like B3LYP/6-311G** may not converge in reasonable time or, if allowed to run long enough, can rearrange
your molecule undesirably. They take long enough even when correct and convergent.
Semi-Empirical PM6 optimization under MOPAC is very rapid, but yields very similar results.
Compared to the PM6 method, ab initio optimization of Ochratoxin had a slightly (and visibly) different handling of the geometry about the peptide
linkage and the nearby intramolecular hydrogen bond. Published structures do NOT show the lower energy hydrogen-bonded conformer.
Ab initio methods put more ring pucker in the lactone moiety, too.
