I want to perform the relaxed scan (or relaxed surface scan) for the valence angle energy of Si–O–Ca, Si–O–Mg, Al–O–Ca, and Al–O–Mg. I will use DFT (ORCA), probably B3LYP and triple zeta basis set (Def2-TZVP or 6-311++G(2df,2pd)) I'm making reference data to parameterize an emperical force field.
But the problem is metal ions, Mg²⁺ and Ca²⁺, because their coordination chemistry is very different from Si or Al. More precisely, I'm not sure how can I 'terminate' the Mg or Ca atoms.
I was thinking of making simple molecules, like (OH)₃–Si–O–Mg–OH or (OH)₂–Al–O–Mg–OH, just giving one OH to Mg or Ca. In this case, Mg or Ca have 2 oxygen neighbors. I can perform geometry optimization, but not sure if that geometry is proper one to perform valence angle scan or not.
I'm not considering aqueous environment (because this is for condensed phase material), so I'm not planning to use water molecules at current stage. But maybe I need to if there are no other way...