| In order to better understand the magma processes, we need to figure out two basic questions:how materials transported and the time it took. Therefore, two independent scientific projects have been done based on these two basic questions. We revealed the diffusion properties of Li isotopes in the rhyolite under temperature gradient with high temperature high pressure experiments. On the other hand, we explained the observed U-series disequilibrium with numerical models and further to place constrains on the time scales of magma processes.Thermal diffusion is a possible factor driving the fractionations of elements and isotopes during varies geological processes. A lot of works have been done on the thermal diffusion of major elements and isotopes (Ca, Mg, Fe, Si, O, Sr, Hf, U,Li and K). However, the mechanism behind it is still debated. In this paper, we investigated the Li isotopes fractionation in the rhyolite with the piston cylinder apparatus, considering that previous works focused on the basalt and andesite systems. Li isotopes diffuse fast and the viscosity of rhyolite is high so that it is likely to generate noticeable Li isotope fractionation but not change in major element contents in the melts in a short period. Therefore we can use the CAMECA1280to in-situ measure Li isotope composition of the melt without considering matrix effect. Moreover, we conducted a series of temporal experiments (3h,12h,1d,2d,4d) to provides insight into the dynamic properties of Li isotopes induced by thermal diffusion.Based on in-situ analyses using SEM, EMPA and CAMECA1280for a time-series experiment, I observed that there is no detectable major element variation but>8%o Li isotope fractionation with256℃/5.8mm after96hours. The extent of Li isotope fractionation is a function of experimental durations.Although most arc lavas have gone through high degree of significant magma differentiation, the effect of magma differentiation on U-series disequilibria is yet clarified. Here we establish a mathematic model to simulate the effect of time-dependent magma differentiation processes on U-series disequilibria in arc lavas. In a closed system with fractional crystallization, the radioactive decay can decrease U-series disequilibria; in an open system, radioactive decay and assimilation of old crustal material can also decrease the initial U-series disequilibria in young lava. With recharge of young magma from greater depths, the magma chamber can maintain substantial226Ra excess even if the residence time is longer than8000years.We use this model to simulate the oceanic island arc has been well studied (Tonga) and continental subduction zone (Chile) lava sequence, as the representatives of the closed and open systems, respectively. In the Tonga case, the positive correlations between (226Ra/230Th) and Sr/Th or Ba/Th (to a lesser extent) can be reproduced by fractional crystallization of plagioclase and amphibole as well as the decay of226Ra in the same period. In this case, U/Th and (238U/230Th) do not change dramatically, In the Chile case, assimilation of old crustal materials can be used to explain the positive correlations between (226Ra/230Th) and10Be/Be or (238U/230Th) because the old crustal assimilant can decrease the primary U-series disequilibria and high10Be/Be.The observed correlations between (226Ra/230Th) and geochemical features (e.g., Sr/Th, Ba/Th, and10Be/Be) often result from fluid addition from the slab. It also can reflect time-dependent magma differentiation processes in the crust. Because the correlations between these trace element ratios and U-series data cannot be used to support the fluid addition model, the previous conclusions about recent fluid addition and ultra-fast ascent rates of arc magmas is strongly questioned. Instead, our model can reconcile the time scales of arc magmatism displayed by U-Th-Pa-Ra and10Be/Be systems and multi-stage fluid addition might not be necessary. |
PDF Full Text Request