# QM modelling of charge separation reactions

I have a general question regarding the modelling of reactions in which neutral reactants form charged species/ions (in solution). How can one get the energetics approximately right on reactions like thes? Examples would be very simple acid/base reactions like $$\ce{NH3 + HCl -> NH4+ + Cl-}$$ but also more complex reactions like $$\ce{Ru(CO)H2(PPh3)3 + HCl -> Ru(CO)H(PPh3)3+ + H2 + Cl-}$$

Important contributions that I can think of, but do not seem to be sufficient in many cases, are:

1. Solvation. Some correction must be applied to account for solvation effects of the ions. I usually try to obtain corrections to $\Delta$G for this from COSMO-RS, and the solvation energies are a lot higher than for neutral species, but still seem to low. It is hard to judge whether these corrections are good, since for most of the interesting cases no data on solvation enthalpies in certain solvents etc. exists.
2. Diffuse functions. For better description of anions, diffuse functions in the basis set can make a difference. However that change also seems to be small when the basis set is already large (say def2-QZVPP to def2-QZVPPD), and for very large basis sets, the computational cost is also considerable.
3. Coordination sphere of ions. Many (smaller) ions probably have a special coordination sphere, which also depends on the solvent, and may form aggregates or oligomers. For new species it's quite likely that solution structure is not known, but even for seemingly simple cases (for example $\ce{NaOMe}$ in $\ce{MeOH}$) it is hard to find data.

Are there other things that should be considered? What is a good "standard" protocol to deal with these kinds of reactions?

• It is not clear how much is approximately right for you, especially if you are asking a rather wide range of reactions, systems and solvent. – Greg Mar 21 '18 at 14:44
• What about explicit solvation? – diogom Mar 24 '18 at 18:33