for 3D printing.
OK, Victor - in Lunar soil samples:
Plagioclase 30%-35% compositions NaAlSi3O8 to CaAl2Si2O8 , melting point 1100—1550 °C.
Pyroxenes 54%-60% usually for earth that is formula XY(Si,Al)2O6 (X = calcium, sodium, iron+2 , magnesium, rarely zinc, lithium; Y = in smaller size chromium, aluminium, manganese, scandium, titanium, vanadium, iron+2), with melting point 1300°C.
Olivine 3%-8% formula (Mg,Fe)2SiO4 melting point Mg2SiO4 = 1897°C.
Ilmenite (absent in Lunar highland areas) 18%-2% formula (FeTiO3). Melting point 1400°C.
I understand that to get Ti, Fe, or Al from oxides needs to do all chemical reaction. And standard way does not work, no CO, no H2, no C, no H2O. To get Iron from Fe2O3 it is possible to use Aluminum: Fe+32O3 +2Al -> 2Fe0 + Al2O3, and not sure about Ti. Looks like way to start will be Aluminum, with molten oxide electrolysis (it is 1200C) - will requare a lot of electrisity - but side effect – it will be 40% of the oxygen ! http://en.wikipedia.org/wiki/File:Moon_vs_earth_composition.svg . I am not talking about reaction like Fe+32O3 +2Al it surplus heat at night time!
With 3D printing on the moon first step will be to separate fractions of a lunar dust.