Zhong-Yuan Ren

The chemical structure of the Hawaiian mantle plume.

The Hawaiian–Emperor volcanic island and seamount chain is usually attributed to a hot mantle plume, located beneath the Pacific lithosphere, that delivers material sourced from deep in the mantle to the surface. The shield volcanoes of the Hawaiian islands are distributed in two curvilinear, parallel trends (termed ‘Kea’ and ‘Loa’), whose rocks are characterized by general geochemical differences. This has led to the proposition that Hawaiian volcanoes sample compositionally distinct, concentrically zoned, regions of the underlying mantle plume. Melt inclusions, or samples of local magma ‘frozen’ in olivine phenocrysts during crystallization, may record complexities of mantle sources, thereby providing better insight into the chemical structure of plumes. Here we report the discovery of both Kea- and Loa-like major and trace element compositions in olivine-hosted melt inclusions in individual, shield-stage Hawaiian volcanoes—even within single rock samples. We infer from these data that one mantle source component may dominate a single lava flow, but that the two mantle source components are consistently represented to some extent in all lavas, regardless of the specific geographic location of the volcano. We therefore suggest that the Hawaiian mantle plume is unlikely to be compositionally concentrically zoned. Instead, the observed chemical variation is probably controlled by the thermal structure of the plume.

Lead isotope analysis of melt inclusions by LA-MC-ICP-MS

Pb isotope compositions of melt inclusions provide unique information about the composition of primary magmas and their source. In this study, we have developed a method for measuring Pb isotopes in small olivine-hosted melt inclusions (>40 μm) from young and old volcanoes by LA-MC-ICP-MS. We used a new interface cone assemblage consisting of a Jet sample cone and X skimmer cone. A small flow of N2 gas was added to the carrier gas and passed through the assemblage to enhance the signal intensity. In addition the energy and repetition rate of the laser conditions were reduced and the signal integration time was shortened in order to lengthen the laser ablation time and to collect enough data. Mass bias and instrument drift were corrected using a standard–sample–standard bracketing method. The analysis routine employed eight ion counters to receive 238U, 235U, 232Th, 208Pb, 207Pb, 206Pb, 204Pb and 202Hg signals simultaneously, which allowed Hg interference to be corrected on 204Pb, and in old samples U–Th decay to be age-corrected. Using the Jet and X cones, under the same laser ablation conditions, the precisions for almost all the measured standard glasses are improved by at least a factor of two compared to using standard cones. At 208Pb signal intensity >200 000 cps, external precisions of ratios involving 204Pb are better than 1.3% (2RSD) and precisions of 208Pb/206Pb and 207Pb/206Pb are better than 0.23% (2RSD). The results of Pb isotopes in olivine-hosted melt inclusions, using 45 μm laser spots, show that the internal precisions of 208Pb/206Pb and 207Pb/206Pb for most analyzed melt inclusions are better than 0.2% (2RSE) and for ratios involving 204Pb are better than 0.8% (2RSE). We are able to present the first ever Pb isotope data from ∼260 Ma Emeishan flood basalt olivine-hosted melt inclusions. They show the importance to do age correction which results in the reduction of the spread of data in old samples. The mean values of age-corrected 208Pb/206Pb and 207Pb/206Pb have 1.2% and 2.8% deviations from the uncorrected mean values, respectively. The method developed here provides a fast, precise and accurate in situ Pb isotopic composition analysis, applicable not only to melt inclusions from young basalts, but also from old samples that require correction for U–Th decay.

Petrogenesis of Early Cretaceous basaltic lavas from the North China Craton: Implications for cratonic destruction

The North China Craton (NCC) is believed to be the best example of cratonic destruction. However, the processes leading to cratonic destruction remain unclear, largely due to a lack of knowledge of the nature of the Mesozoic NCC lithospheric mantle. Here we report new petrological and geochemical data for Early Cretaceous NCC basalts, which provide insights into the nature of the underlying lithospheric mantle. The Early Cretaceous basalts (all tholeiites) show a limited variation in geochemical composition. In contrast, olivine-hosted melt inclusions from these basalts display a wide range in compositional variation, and include both alkalic and tholeiitic basaltic compositions. This result provides the direct evidence of the contribution of silica-undersaturated alkali basaltic melts in the petrogenesis of the Early Cretaceous NCC basalts. In addition, the compositions of olivine phenocrysts and reconstructed primary melts indicate that the Early Cretaceous basalts are derived from a mixed peridotite and refertilized peridotite source. The Pb isotopic compositions of melt inclusions in high-Fo olivines combined with trace element characteristics of these basalts reveal that heterogeneous lithospheric mantle sources for Early Cretaceous basalts were metasomatized by carbonate-bearing eclogite-derived melts. The Pb isotopic variations of the melt inclusions, and clinopyroxene and plagioclase phenocrysts demonstrate that the mantle-derived magmas were variably contaminated by lower continental crust. We propose that multiple subduction events during the Phanerozoic, combined with mantle-plume activity, likely play a vital role in the generation of the Early Cretaceous voluminous magmatism and cratonic destruction.

The influence of the double spike proportion effect on stable isotope (Zn, Mo, Cd, and Sn) measurements by multicollector-inductively coupled plasma-mass spectrometry (MC-ICP-MS)

This study investigates the double spike (DS) proportion effect on measurements of stable Zn, Mo, Cd, and Sn isotopes by multicollector-inductively coupled plasma-mass spectrometry (MC-ICP-MS). The effect of DS proportion values between 0.1 and 0.9 (i.e., 10–90% DS) on the compositions of measured references materials was determined. The δ66/64Zn, δ114/110Cd, and δ120/118Sn values positively correlate with the proportion of the DS within a mixture, whereas δ98/95Mo values negatively correlate with the proportion of DS. Stable Mo, Cd, and Sn isotopes have a range of DS values around the optimum where the resulting delta values are insensitive to DS proportion values and reflect the values of the reference materials with the optimum DS proportion values. However, Zn isotopes do not have a stable DS proportion range, indicating that δ66/64Zn values are much more sensitive to DS proportion values than those of the other three elements. The measured δ98/95Mo values of a mixture of the IAPSO seawater reference with variable amounts of DS solution also show a negative correlation with DS proportions. This indicates that although it is difficult to explain why DS proportion values have a significant influence on isotopic analyses, we suspect that the iterative calculations involved in the three algebraic equations used to resolve DS data generate correlations between delta and DS proportion values. This study highlights the importance of assessing DS proportion effects when establishing analytical protocols for stable isotope measurements using the DS technique. Any correlation between the resulting measured isotopic compositions and the associated DS proportion values means that DS solutions must be precisely added to unknown samples to ensure that appropriate proportions of DS are used.