Understanding Nd model ages of granite rocks: The effects of the 147Sm/144Nd variability during partial melting and crystallization.

Nd model ages are valuable tools for studying the structure and evolution of the continental crust. However, their calculation is severely affected by the 147Sm/144Nd lack of constancy during crustal evolution because it results in spurious Nd model ages that may induce wrong geodynamic and geochemical interpretations. An extensive database of granite analyses shows that 147Sm/144Nd may vary significantly in the high-SiO 2 range of granite rocks and that these variations are far more common in peraluminous than in metaluminous varieties. Only a tiny fraction of the latter has anomalously low ratios, likely resulting from apatite, perhaps titanite, fractionation. In contrast, many high-SiO 2 peraluminous granites also show abnormally high, even superchondritic 147Sm/144Nd ratios that likely resulted from monazite fractionation. A Sm and Nd residence study in peraluminous sources revealed that Sm/Nd ratios are lower in monazite and the feldspars than in the bulk rock but higher in apatite and garnet. Monazite carries 90% of Nd and 80% of Sm, apatite carries about 3–5% of Nd and up to 16% of Sm, and garnet may concentrate up to 27% of Sm. Accordingly, the interplay among monazite, apatite and garnet determines the 147Sm/144Nd of resulting magmas, obscuring the role of the other phases. Garnet is expected to be primarily a residual or early crystallized phase, although some can be entrained in melts. Apatite may also behave as a residual phase in high-SiO 2 metaluminous systems. However, it is highly soluble in peraluminous systems. Monazite may dissolve partially, remaining in the residua of peraluminous melts and crystallizing early from these, as seen in leucosome-melanosome pairs from the Ivrea-Verbano Zone, NW Italy, and Peña Negra, Central Iberia, and the upwards differentiated Pedrobernardo granite sill, also in Central Iberia. Garnet or apatite retention in residua, or segregated from a crystallizing magma, produces melts with low 147Sm/144Nd that may cause anomalously young Nd model ages, as exemplified by the Roccapietra granite of the Ivrea-Verbano zone, NW Italy. Monazite retention or crystallization, in contrast, produces high 147Sm/144Nd melts that yield anomalously old Nd model ages. The effects of intracrustal 147Sm/144Nd variations may be softened or eliminated by employing the two-stage Nd model age calculation method (DePaolo et al., 1991) with a Sm/Nd evolution model adapted to each specific crustal segment. • Realistic Nd model ages require the constancy of Sm/Nd during crustal evolution. • The constancy does not hold during the generation and evolution of peraluminous magmas. • Monazite fractionation causes abnormally high Nd model ages. • Garnet and, likely, apatite fractionation cause abnormally low Nd model ages. • These effects can be corrected using the two-step Nd model age calculation. [ABSTRACT FROM AUTHOR]