I mean, what are you trying to do? There isn’t a standard methodology yet for reactive molecular dynamics near an electrified electrode.
What technique you decide to use depends on what you are trying to model (and therefore what errors you can tolerate – remembering that “all models are wrong”).
It also means you have to “get your hands dirty” with method development and validation, and that is far harder than simply running simulations.
I’m trying to understand effect of external electric field applied at the solution-metal interface on the the oxidation of Fe-based alloys.
Well that is a Very Hard Problem™. But you need to look through the literature and work out for yourself (1) what has previously been done with similar interfacial phenomena, especially without electric fields; (2) whether there are methodologies designed to then impose an electric field; (3) whether molecular dynamics is even the correct methodology to use. (Remember that transition metal chemistry revolves heavily around d-orbitals and their funky shapes, and there are no orbitals in molecular dynamics.)
Seconded.
It is even worse. ReaxFF parameters are usually generated from DFT calculations, but to represent the different oxidation states of iron with DFT is already a tricky problem due to spin contamination and extended DFT schemes like DFT+U or non-DFT methods like multi-configuration SCF (e.g. CASSCF) are required to get accurate ab initio results. So if you cannot trust the input data for the parameterization, how can the simpler ReaxFF model with the resulting parameters produce reliable results?