To make EAM potential as addressed in J Cai's paper, parameters are
different for the case of between pure and alloy system.
In my case, for example, f_e (scaling constant used in electron density
function) is 1.0 for pure Al and Cu, but 0.383372113 and 0.485336561 for
alloy Al and Cu, respectively. Therefore, rho_e (host electron density at
equilibrium) also should be changed.
Did you consider this?