Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4703497 | Geochimica et Cosmochimica Acta | 2011 | 20 Pages |
In order to fully assess the role of rutile in fractionation of Nb/Ta during partial melting of hydrous metabasalt, we have measured rutile – felsic melt partition coefficients (D values) for Nb and Ta with tonalitic to trondhjemitic compositions at 1.5–3.5 GPa, 900–1350 °C and ∼5.0–20 wt% H2O. DNb, DTa and DNb/DTa range from 17 ± 1 to 246 ± 13, 34 ± 2 to 232 ± 25 and 0.51 ± 0.04 to 1.06 ± 0.13, respectively. For the compositions investigated, melt composition appears to have no observable effect on the partitioning; the effect of pressure is also slight; whereas temperature and H2O have marked effects. DNb, DTa and DNb/DTa increase with decreasing temperature and H2O content, showing a reversal of DNb/DTa from <1.0 to >1.0. Using the data that approached equilibrium and obeyed Henry’s law, expressions describing the dependences of DNb, DTa and DNb/DTa on temperature, pressure and melt H2O content were obtained: equation(1)ln(DNb)rutile/melt=-2.846(±0.453)+9621(±470)/T+0.207(±0.101)P-0.042(±0.009)H2Oln(DNb)rutile/melt=-2.846(±0.453)+9621(±470)/T+0.207(±0.101)P-0.042(±0.009)H2Oequation(2)ln(DTa)rutile/melt=-0.775(±0.398)+6954(±413)/T+0.140(±0.089)P-0.009(±0.007)H2Oln(DTa)rutile/melt=-0.775(±0.398)+6954(±413)/T+0.140(±0.089)P-0.009(±0.007)H2Oequation(3)ln(DNb/DTa)rutile/melt=-2.075(±0.289)+2657(±301)/T+0.075(±0.065)P-0.033(±0.006)H2Oln(DNb/DTa)rutile/melt=-2.075(±0.289)+2657(±301)/T+0.075(±0.065)P-0.033(±0.006)H2O where T is in K, P is in GPa, and H2O are in wt%. These expressions are applicable to calculate Nb and Ta partitioning coefficients between rutile and felsic (highly siliceous) melts and Nb/Ta fractionation during partial melting of metabasalt.The derived models suggest that fluid-present partial melting of metabasalt always leads to DNb/DTa lower than 1.0, whereas dehydration melting may lead to DNb/DTa higher than 1.0 at temperatures <1000 °C and melt H2O contents <10 wt%. Low-degree melting cannot substantially change the Nb/Ta ratio of rutile-bearing eclogite residues relative to their basalt precursors, whereas high-degree melting leads to lower Nb/Ta ratios in the residues. These results thus confirm that partial melting of hydrous metabasalt cannot produce rutile-bearing eclogites with superchondritic Nb/Ta. Fractionation of Nb/Ta for the melts relative to their basalt precursors can occur during low-degree melting due to the high DNb and DTa and variable DNb/DTa. The fractionation direction depends on melting degree (temperature) and protolith initial H2O content (assuming that hydrous phases are consumed entirely during partial melting), whereas the fractionation extent depends on protolith initial H2O and TiO2 contents. Fluid-present melting at <20 wt% melt fractions will significantly elevate melt Nb/Ta ratio, whereas dehydration melting with low initial H2O (<1.2 wt%) may reduce melt Nb/Ta ratio. In general, rutile in eclogite residues accounts for the highly variable Nb/Ta ratios in the Archean TTG magmas. In particular, the very high Nb/Ta ratios of some TTG magmas may only be explained by low-degree (<20 wt%) melting in the presence of both rutile and a hydrous fluid.