At 24:40 into this lecture, technology development scientist Laurent Sibille briefly describes an experiment done with a regolith simulant for Lunar Mare. The process of producing oxygen and reusing methane et cetera doesn't seem to be relevant to the actual melting of regolith to metal.
Does the surface of the Moon really turn into solid metal if one simply concentrates sunlight on it?
If that is solid metal, it must be easy to heat it again and use a centrifuge, magnetism or chemical reactions to separate the different pure metals. But for a landing pad or a road or most structures, maybe a random iron-magnesium-titanium alloy works very well?
Thermal energy having turned regolith into apparently solid metal lumps. From a slide in Sibille's lecture at SSERVI.