| Summary: | 碩士 === 國立成功大學 === 地球科學系碩博士班 === 96 === Mantle chemical dynamics is an important link to understand element cycling in solid Earth. Profound variations in the REE patterns of clinopyroxene from the Oahu spinel lherzolite xenoliths provide an opportunity for investigating chemical dynamics in the intra-plate lithospheric mantle. Core-to-rim variations in REE abundances of clinopyroxene grains were determined by in-situ ion probe analysis. REE variations in clinopyroxene from the Pali spinel lherzolite xenoliths were classified into three types: (1) LREE-depleted patterns typical of melting residues, (2) U-shaped REE patterns, and (3) nearly-flat REE and slight LREE-enriched patterns. The U-shaped and flat REE patterns were typically attributed to interaction between LREE-depleted melting residues and LREE-enriched melts (Sen et al., 1993; Yang et al., 1998). The Sr-Nd-Os-Hf-Pb isotope data of these xenoliths indicate involvement of Honolulu Volcanics (Vance et al., 1989; Okano and Tatsumoto, 1996; Lassiter et al., 2000; Ducea et al., 2002; Bizimis et al., 2003; 2007). Three major types of melt-peridotite interactions have been proposed, including: (1) simple element diffusion from melt into peridotite (Klugel, 1998; Van Orman et al., 2001), (2) mixing between melt and peridotite (Song and Frey, 1989), and (3) melt percolating through peridotite (e.g., Takazawa et al., 1992; Ionov et al., 2002; 2006). However, simple element diffusion during transportation of peridotitic xenoliths to surface by magmas results in core-to-rim REE concentration gradients higher than that in our samples, while mixing between peridotite and alkalic siliceous melts does not show intensive concave down patterns of the samples. Therefore, REE patterns of the samples must be dominated by processes prior to incorporation of the xenoliths into their host lavas. The “melt percolation model” that simulates exchange of trace elements between melt and matrix solid as a function of time and distance from melt entrance provides satisfactory explanation for the observed U-shaped patterns. Modeling results indicate that both large-scale and local-scale melt percolation processes can cause the U-shaped REE patterns of clinopyroxene in the Pali lherzolitic xenoliths. Large-scale porous flow percolation (~10 km) with rapid melt velocity (~20 cm/yr) and a porosity of ~0.5 % may produce U-shaped REE patterns satisfactorily fitting to that of analyzed clinopyroxene (77PAII-10-cpx1) with percolation time of 50000 years. The reaction time can be reduced to several thousand years, if the melt velocity increases to several meters per year. In such a case, the channel flow should once occur in the upper mantle beneath Oahu, Hawaii. An alternative model is that the U-shaped REE patterns were resulted from local-scale porous flow prevalently occurring in the upper mantle. With a melt velocity of 1 cm/yr and a porosity of ~0.5 %, strong zonation REE pattern in 77PAII-10-cpx1 could be produced over relatively short distances of ~100 m by local-scale porous flow coupled with short time interval (~12000 yr). This inference must be confirmed by more analyses on clinopyroxene for LREE abundances and robust constraints on the equilibrium depths of the xenoliths.
|
|---|
