I did my PhD in Mn-salen and oxo-Mn-salen complexes and found Mn to be quite a challange, even among other 1st-row Transition Metals. On a work by Jacobsen and Cavallo (J. Phys. Chem. A, 2003, 107, 5466) they studied the impact of two different funtionals on the epoxidation mechanism. Indeed, Cavallo and Jacobsen made a series of studies on these systems in the 2000's which are a must-read to anyone doing research in the field, not only due to their results, but also to the detailed explanations they provide in those papers. More recently, I did a short review on the matter and compiled a table with the ordering of the singlet, triplet and quintuplet states on mn-salen and oxo-mn-salen species at different levels of theory. Concerning structure and properties (not reaction mechanisms) any DFT at or above GGA should provide concordant results (that is not the same as accurate results) (see Tables 2 and 3 of Catalysts 2017, 7, 2). But overall, my experiente tells me to use the best basis set possible, at least a triple-zeta basis with a generous amount of polarization, like Def2-TZVP. As for the functional, I recommend always using two different ones (for example B3LYP and TPSS, or B3LYP and one from the M06 family) and check for concordance among them.
I have read both the papers (J. Phys. Chem. A, 2003, 107, 5466 and Catalysts 2017, 7, 2). You are right. To get the accurate results, we have to check the different functional with higher basis set like triple-zeta basis with a generous amount of polarization.
But my question is not about the functionals or basis sets. I want to know how to calculate the electrochemisty (reduction potentional) of high-valent metal-oxo complexes theoritically because this property, we can easily calculate experimentally using cyclic voltammetry and relate this to theory.
There are a few methods for that, although I'm not an expert in electrochemistry. I've seen some papers where the authors relate the chemical hardness with the reduction potential, but it was a few years ago and I haven't kept the reference. I'd go for the easiest way, which is to calculate the Delta G for the redox transformations involved using a standard thermochesmistry (optimization + vibrational analysis and thermochemistry). That would follow the same framework as the one on the paper suggested by Dr. Ivanova. The problem with such an approach is that DFT is not accurate enough (specially considering the exponential/logaritmic relationships between Delta G, K and the redox potential), but if you have some experimental data on (at least some of) the same reactions, perhaps you might be able to make an empirical adjustment (similar to what is done for the vibrational frequencies).