Two stage reaction would be ridiculously inefficient, it would make more sense - if the chemistry was suitable - to use aluminium directly as a battery electrode.
The problem with our use of aluminium is the immense amount of electricity required to refine it in the first place.
The reason the balloon will not float is not so much the purity of the hydrogen, but that there is too much weight of rubber. And of course, rubber balloons leak hydrogen in a few hours. A hydrogen generator and (two, decent sized) balloons was part of the "Gibson Girl" EPIRB system back around WWII.