Charge consistency in fragment qm/mm methods

I'm calculating the multipole moments on several fragments of a molecule. What are some procedures to ensure that the charges etc are representative of the complete molecule? How do I ensure I'm not getting an overall charge in a neutral molecule?

• Draw a Lewis Dot Structure. This may help figure out the position of $e-$ and thus help you find the charge of a fragment. Apr 26, 2015 at 0:20
• That might help if we were dealing with integer charge, but any sub part of a molecule most of the time doesn't have that. For instance in water the O might have a charge of -0.35 and H1 might be 0.15 and H2 might be 0.2 . I'm looking for a somewhat mathematical rigorous procedure to do this. For instance if the water described above is the true result and my fragments are each atom If my O is -0.4 H1 0.1 and H2 0.15, how do I approach the true result? Apr 26, 2015 at 0:42
• I really don't understand what you are asking. Do you think, that you will lose or gain an electron during a calculation? If you set up a calculation, you specifically set the parameters to a fixed number of electrons and nuclei. The program just finds the best orbitals where to put the electrons around the nuclei. So the second case in your comment is just impossible. It would be helpful, if you could describe your methodology more precisely, maybe I am interpreting it wrong. Apr 26, 2015 at 7:54
• I'm performing a Distributed Multipole Analysis on a protein. Apr 26, 2015 at 20:25

If you're doing any sort of computational run (either quantum chemistry, solid state, or molecular dynamics), you have to set the charge on the overall system.

You said in your comments that you're doing a distributed multipole analysis across multiple fragments of a protein.

In this case, one of the constraints on the calculation is exactly that the overall charge must match that specified. The program won't arbitrarily change the charge to be an anion or a cation if you specified a neutral system.

Now, that's not to say that the individual fragments won't have fractional or integer charges. Depending on the method, it's entirely reasonable that there's some amount of partial charge transfer between fragments.

It's just that there's absolutely no way that the sum of the charges will be different than the constraint.