I'm computationally working on ruthenium azopyridine complexes, with GaussView and Gaussian 03. unfortunately, all calculation end up with metallic bonds broken although I do use Goem=connectivity. Please help me.
Could you add the output file. Think that all that you see in gW it isn't real, it is only a representation. The fact, there aren't any signification about bond in gW and g03. All calculation is about distances, not bonds. This is an "effect" if distance and electron probability. So, in the result you have a "good" distance but without bond, the opt its correct and exist a bond.
On the other hand, if optimized geometry have excessive distances or negative frequencies, you are not in a minimum and you should check the initial geometry.
I have seen the structure and I can see that you have an avarage of distances of 2ams. The viewer program draw a bond from a X distance depending on data. For example, you have an Ru-N (2.04ams) and gview draw "automatically" a bond. The next Ru-N bonds is 2.09 and it don't draw the bond because it is "too far" but it isn't real. You have all bond in a distances of 2.... and really are bonded, only think is that program didn't draw it because it uses a references. (2.04 have bond, 2.041 don't have it, for example). Chemically have no sense.
Only think to do is draw it manually (because for us it is more visual) but the bonds exists in these distances and all calculations you do, uses coordinates, not bonds. (NOTE THAT IT IS DIFFERENT WITH Z-MATRIX NOTATION).
I was watching the structure. The optimized geometry have a really poor symmetry, it is a minimum but I don't see clearly the initial geom. See that for the same Cl, have the two OH bond, dihedral angles are very different from groups. Some atoms should be equivalent, but have a different distances and angles.