We present results of a secondary ion mass spectrometry study of the rare earth elements (REEs) in the minerals of two samples of lunar ferroan anorthosite, and the results are applicable to studies of REEs in all igneous rocks, no matter what their planet of origin. Our pyroxene analyses are used to determine solid-solid REE distribution coefficients (D = C_REE in low-Ca pyroxene/C_REE in augite) in orthopyroxene-augite pairs derived by inversion of pigeonite. Our data and predictions from crystal-chemical considerations indicate that as primary pigeonite inverts to orthopyroxene plus augite and subsolidus reequilibration proceeds, the solid-solid Ds for orthopyroxene-augite pairs progressively decrease for all REEs; the decrease is greatest for the LREEs. The REE pattern of solid-solid Ds for inversion-derived pyroxene pairs is close to a straight line for Sm-Lu and turns upward for REEs lighter than Sm; the shape of this pattern is predicted by the shapes of the REE patterns for the individual minerals.

Equilibrium liquids calculated for one sample from the compositions of primary phases, using measured or experimentally determined solid-liquid Ds, have chondrite-normalized REE patterns that are very slightly enriched in LREEs. The plagioclase equilibrium liquid is overall less rich in REEs than pyroxene equilibrium liquids, and the discrepancy probably arises because the calculated plagioclase equilibrium liquid represents a liquid earlier in the fractionationsequence than the pyroxene equilibrium liquids. “Equilibrium” liquids calculated from the compositions of inversion-derived pyroxenes or orthopyroxene derived by reaction of olivine are LREE depleted (in some cases substantially) in comparison with equilibrium liquids calculated from the compositions of primary phases. These discrepancies arise because the inversion-derived and reaction-derived pyroxenes did not crystallize directly from liquid, and the use of solid-liquid Ds is inappropriate. The LREE depletion of the calculated liquids is a relic of formation of these phases from primary LREE-depleted minerals. Thus, if one attempts to calculate the compositions of equilibrium liquids from pyroxene compositions, it is important to establish that the pyroxenes are primary. In addition, our data suggest that experimental studies have underestimated solid-liquid Ds for REEs in pigeonite and that REE contents of liquids calculated using these Ds are overestimates.

Our results have implications for Sm-Nd age studies. Our work shows that if pigeonite inversion and/or subsolidus reequilibration between augite and orthopyroxene occurred significantly after crystallization, and if pyroxene separates isolated for Sm-Nd studies do not have the bulk composition of the primary pyroxenes, then the Sm-Nd isochron age and ε_Nd will be in error.