One has to use some external basis set with ECP (e.g. LANL2DZ).
Is there a way to put basis sets and ECPs to .lib file and then to read ECP from such file (like we do with basis sets using EXTFIL=.t. in $BASIS group? Or giving $ECP group with ECP for each atom is the only way to use such ECP?
Another problem. Let there be a molecular cluster containing, say, 10-20 atoms like Al or Si (let it be support for some kind of catalyst), and also some O and H atoms and also containing several heavy atoms those definitely require relativistic basis set (e.g. Bi, Pt, Re, W) and besides we have some organic reactant or reactants (e.g. ethanol). It seems that using LANL2DZ (or something like it) for the whole system will not give good results, so one should like to use LANL2DZ with corresponding ECP for heavy atoms, 6-31G (maybe with several polarization functions) for Al, Si, lattice O and some bigger basis set, e.g. 6-31G with some polarization and diffuse functions for O atoms nearest to active center and the most important reactants' atoms. How to do all this with the least time consumed (taking into account that there can be a hundred and more .inp files and "copypasting" basis sets and ECPs each time will take a lot of time)?