[Schrödinger]

Hey guys!

I was wondering... How do you draw MO diagrams for entire molecules? I've been trying to do some geometry optimizations and when it comes to specifying whether a molecule's ground state is triplet or singlet, I'm stumped! I tried  using Arguslab to plot the MO surfaces and then arrange them according to the number of nodes. But that didn't really work out well simply because there are too many when a large molecule is taken. I'm still not able to arrange MO's in increasing order of energy for a molecule as such. All I know is how to do the same for homo/hetero-nuclear diatomics like N₂, O₂, NO, CO, etc. I tried googling it, but every result is somehow related to either diatomics, conjugated π-bonded systems or simple molecules like NH₃, ₃, etc. How do I find the MO diagram for any molecule in general, and how do I say whether it is singlet or triplet in its ground state just by looking at the molecule (if that's even possible)

[curiouscat]

I've been trying to do some geometry optimizations and when it comes to specifying whether a molecule's ground state is triplet or singlet, I'm stumped!

Aren't these just initial guesses? If so, try both and let the system converge to its lowest E state.

[Schrödinger]

Yes indeed, there are computational techniques. The software (say Gaussian) will tell me if the singlet state is stable so that in case it is unstable, I can change it to triplet. But I was asking from a more theoretical standpoint than a computational one. How exactly do you draw the MO diagram for any random molecule? Does it involve super-complex math or is there a handy and easy way to do it?

[vex]

I think it becomes very hand-wavey after a certain point. I think the best you can do is to work from electronegativity and electron withdrawing/donating trends to deduce where partially localized electrons will be happiest. In metal centers, symmetry reduction using Walsh correlation diagrams from a high-symmetry point group (octahedron being the favorite) can help a lot. But, in my limited experience, especially when you're juggling multiples of these factors simultaneously, at some point you have to call it a day and do a calculation.

Honestly, I would go with curiouscat's suggestion if I had to solve this problem myself. Even a low-level, quick and dirty calculation should get the stability with respect to spin state right. If you're worried about efficiency, you could start from the geometry given in a crystal structure and just do a single-point energy calculation with each spin state.

[Schrödinger]

Thanks a ton! I've been breaking my head over this problem for the past few days. I didnt want to do it computationally. I guess it can't be avoided once you delve into the world of molecules more complex than the the usual diatomics given is textbooks!

[dayglo]

How complex are the molecules that you are trying to model? Are they simple/symmetric enough that you can use group theory and Tanabe-Sugano diagrams to approximate what your ground state should be? You might not get excited state orders very well but you should be able to predict ground states relatively well.