Lunar volatiles can be mobilized by impact processes, and pass from rock/regolith into the moon’s atmosphere and eventually adsorbed on the surface or lost to space. Much of the volatiles mobilized by impact must pass through lunar regolith or rock, and can react there to form volatile-rich metasomatic deposits.
One such metasomite is Apollo sample 79215: a polymict impact melt breccia, annealed and chemically equilibrated to a granulitic texture [1,2]. It is unlike other lunar granulites containing abundant apatite (1%), but little of other KREEP elements . Its apatite occurs in millimeter-sized euhedra and subhedra to micron-sized anhedra (like the surrounding grains). Apatite is concentrated in curvilinear bands. It is chemically homogeneous with F/Cl≈10 and ~35 ppm H2O. Its hydrogen and chlorine are isotopically heavy at dD=+1000±280‰ (2s) and d37Cl=+31±3‰ (2s), like many lunar materials including KREEP [4-6]. Thus, the source of the apatite was likely to have been KREEP, the only lunar component rich in P, F and Cl. However, 79215 is minimally enriched in other KREEP elements (e.g., REE, U, Th, Ta, Zr, K and Ba), so derivation of the 79215 apatite from KREEP required significant elemental fractionations. These fractionations could arise naturally if the source of the metasomatic fluid retained merrillite, zircon, rutile, and K-feldspar.
P, F, and Cl in 79215 were likely deposited from vapor, akin to apatite deposits known in lunar regoliths [7,8]. The placement of apatite grains in curving traces suggests deposition from a fluid. Aqueous liquid is unlikely given the low H2O content of the apatite; silicate (and ionic) melts provide inadequate fractionation of REE and excessively high solubility of zircon, rutile and K-feldspar. This leaves only vapor, which must carry halogens, P, and probably oxygen; likely there was little S, as 79215 still contains Fe metal. The vapor could have been halogen-dominated , but fugacities of C-bearing species are unconstrained and could have been large. Because the apatite was annealed with the silicate minerals, it is likely that the apatite formed during (or before) the annealing, which occurred as the protolith was emplaced as an impact ejecta blanket . Thus, we infer that the impact event that produced the 79215 granulite also mobilized volatile constituents, likely from underlying KREEPy rock or soil. 79215 became enriched in P, F, and Cl, but not in more volatile elements like Zn or Cd. Perhaps these passed through 79215, eventually to form deposits like those in ‘rusty rock’ 66095 .
References  Bickel C.E. et al. (1976) Proc. 7th LPSC, 1393.  McGee J.J. (1978) Proc. 9th LSC, 743.  Treiman A.H. et al. (2013) Lunar Planet. Sci. 44th, Abstr. 1567.  Sharp Z. et al. (2010) Science 329, 1050.  Greenwood J.P. et al. (2012) Lunar Planet. Sci. 43rd, Abstr. 2089.  Boyce J. et al. (2013) Lunar Planet Sci. 44th, Abstr. 2851.  McKay D.S. et al. (1972) Proc. 3rd LSC 1, 739.  Davis A.M. et al. (2001) Meteoritics 36, Abstr. 5284.  Hunter R.H. & Taylor L.A. (1981) Proc. 12th LPSC, 261-280.